EqualLogic Configuration Guide

Document technical information

Format pdf
Size 3.9 MB
First found May 22, 2018

Document content analysis

Category Also themed
Language
English
Type
not defined
Concepts
no text concepts found

Persons

Organizations

Places

Transcript

Dell EqualLogic
Configuration Guide
Dell Storage Engineering
•
Configure unified block and file storage solutions based on EqualLogic PS Series
arrays and the FS Series Family of NAS Appliances.
•
Recommendations and best practices for iSCSI SAN and scale-out NAS network fabric
design.
•
Updated capacity guidelines, capabilities, limitations, and feature sets for the
EqualLogic product family.
Version 14.1 March 2013
Abstract
This configuration guide provides technical guidance for designing and implementing Dell EqualLogic
PS Series storage solutions in iSCSI SAN environments. The goal of this guide is to provide a single
reference for product technical information and recommended SAN and NAS design methods:
•
•
•
•
Details on the capacity and capabilities of different EqualLogic PS Series family and FS Series
family of NAS Appliances
An overview of the peer storage feature set provided by the EqualLogic storage controllers and
the NAS feature set provided by the FS Series NAS Service
iSCSI SAN design guidelines for optimizing redundancy, availability and performance of
EqualLogic storage arrays.
Host configuration guidelines and recommendations for Ethernet switch configuration
This document should be used strictly as a guide in planning an EqualLogic SAN solution. It should not
be considered a statement of support for any specific configuration. Actual viability of any
configuration will depend on the capabilities of the individual components (switches, network
interface controllers, initiators, etc.) that make up the SAN infrastructure.
This configuration guide is intended for storage administrators, SAN designers, storage consultants, or
anyone who is considering purchasing or has purchased EqualLogic PS Series Arrays for use in a
production storage area network.
THIS WHITE PAPER IS FOR INFORMATIONAL PURPOSES ONLY, AND MAY CONTAIN TYPOGRAPHICAL ERRORS AND
TECHNICAL INACCURACIES. THE CONTENT IS PROVIDED AS IS, WITHOUT EXPRESS OR IMPLIED WARRANTIES OF ANY KIND.
© 2013 Dell Inc. All rights reserved. Reproduction of this material in any manner whatsoever without the express written
permission of Dell Inc. is strictly forbidden. For more information, contact Dell.
Dell, the DELL logo, and the DELL badge, PowerConnect™, Dell™ EqualLogic™, PowerEdge™ and PowerVault™ are trademarks
of Dell Inc. Broadcom™ is a registered trademark of Broadcom Corporation. Cisco® is a registered trademarks of Cisco
Systems, Inc. and/or its affiliates in the U.S. and other countries. Intel™ is a registered trademark of Intel Corporation in the U.S.
and other countries. Microsoft™, Windows™, Windows Server™, and Active Directory™ are either trademarks or registered
trademarks of Microsoft Corporation in the United States and/or other countries.
March 2013
Dell EqualLogic Configuration Guide v14.1
i
Revision history
Revision
Date
Changes
14.1
March 2013
Added section 14 Data Center Bridging
Added volume limits for Synchronous Replication in Table 3
Modified section 7.1.2 and section 7.2
Changed all references in tables in section 4 to read PS41x0
and PS61x0
Added note in section 5.1
Updated Appendix D – EqualLogic Upgrade Paths
Updated the Related Publications
13.4
December
2012
Updates:
New section numbering
Updated Preparing for Firmware Upgrade and Controller
Failover
Added note box to reflect 10Gb support only on PS4110 and
PS6110 ports
Updated for PS-M4110XS and PS65x0 Hybrid Arrays
Added note box to reflect no support for Direct Attach
Storage (DAS)
New Sections:
Appendix D: Upgrade Paths for EqualLogic PS Series Arrays
13.3
September
2012
Updates:
Updated for PS-M4110 Array.
Updated Blade Integration section to include Dell Force10
MXL.
Updated for FS7600/FS7610 and Fluid FS 2.0.
Updated Controller Firmware Upgrade Path to Firmware 6.x.
Updated Section 2.4 RAID Policies
Corrected table 25 in Section 4.7.1.
New sections:
Section 4 Capacity Planning.
Section 6.2 Synchronous Replicaton (SyncRep).
Section 6.3 Internet Protocol Security (IPsec).
March 2013
13.2
June 2012
Added PS Series Firmware Compatibility with EqualLogic
Tools in Table 7.
13.1
March 2012
Updated for PS6110/4110. Updated Blade Integration in
section 4.7. Updated the capacity tables from raw storage
sizes to usable storage sizes. Added replication partner
compatibility information.
12.4
November
2011
New sections and other changes throughout the document
to include coverage of FS Series NAS Appliance (FS7500)
12.3
October 2011 Updates for PS4100/PS6100 family arrays: controller details;
vertical port failover; SAN configuration guidelines; RAID level
capacity tables
Dell EqualLogic Configuration Guide v14.1
ii
12.2.1
July 2011
Corrected Spanning Tree portfast guidance in Appendix D.
12.2
May 2011
Updated figures and content additions to Replication section
12.1
March 2011
Section 3.3.7: Replication
Section 4.6: Integrating 1GbE and 10GbE SANs
Section 2.1: Supported Configuration Limits (maintaining
consistency with current firmware release notes)
Appendix B and Appendix C: Content removed and now
published as a separate Validated Components List
document
All significant changes included in the most recent version of this document are
highlighted using blue text.
March 2013
Dell EqualLogic Configuration Guide v14.1
iii
Table of contents
Abstract ......................................................................................................................................................................... i
Revision history ........................................................................................................................................................... ii
1
PS Series storage arrays ................................................................................................................................1-2
1.1
Array models ...........................................................................................................................................1-2
1.2
PS Series supported configuration limits ...........................................................................................1-3
1.3
Array models prior to PS4100/PS6100 ............................................................................................... 1-5
1.3.1
Controller types in all models prior to PS4100/PS6100 ......................................................... 1-6
1.3.2
Controller redundancy in all models prior to PS4100/PS6100 .............................................. 1-7
1.3.3
Controller failover behavior in all models prior to PS4100/PS6100 ..................................... 1-7
1.4
1.4.1
Controller types in PS4100/PS6100 models .......................................................................... 1-10
1.4.2
Controller redundancy in PS4100/PS6100 controllers ......................................................... 1-11
1.4.3
Controller failover behavior: PS41x0/PS61x0 ......................................................................... 1-11
1.4.4
Vertical port failover behavior in PS4100/PS6100 controllers ............................................. 1-12
1.4.5
Vertical port failover behavior in PS4110/PS6110 controllers .............................................. 1-15
1.5
2
Array models PS4100/PS6100 .......................................................................................................... 1-10
Array model PS-M4110 ........................................................................................................................ 1-17
1.5.1
Controller type in PS-M4110 model ......................................................................................... 1-17
1.5.2
Configuration options .................................................................................................................1-18
1.5.3
Failure behavior in the PS-M4110 ..............................................................................................1-18
1.5.4
Networking considerations and guidelines .............................................................................1-18
Controller firmware .................................................................................................................................... 2-20
2.1
About member firmware.................................................................................................................... 2-20
2.2
Firmware upgrade considerations .................................................................................................... 2-20
2.2.1
2.3
3
PS Series Firmware Compatibility with EqualLogic Tools .................................................... 2-22
Optimizing for High Availability and preparing for Array Firmware updates ............................ 2-23
2.3.1
Verify hosts are designed for high availability ........................................................................ 2-23
2.3.2
Microsoft Windows Servers ....................................................................................................... 2-23
2.3.3
VMWare ESX Servers ................................................................................................................... 2-23
2.3.4
Storage Heartbeat on vSphere 5.0, 4.1, and 4.0 .................................................................... 2-24
2.3.5
Linux servers ................................................................................................................................. 2-24
2.3.6
Network topology health check ............................................................................................... 2-24
2.3.7
Environmental health.................................................................................................................. 2-24
RAID policies ................................................................................................................................................ 3-25
March 2013
Dell EqualLogic Configuration Guide v14.1
iv
3.1
Setting the RAID Policy for a member ............................................................................................. 3-25
3.2
Guidelines for choosing a RAID policy ............................................................................................ 3-25
3.2.1
RAID level ...................................................................................................................................... 3-26
3.2.2
Drive configuration ..................................................................................................................... 3-26
3.2.3
Spare drive policy ........................................................................................................................ 3-26
3.3
RAID level characteristics ................................................................................................................... 3-26
3.3.1
3.4
4
5
Supported RAID policy conversions ................................................................................................ 3-28
Capacity planning........................................................................................................................................4-30
4.1
RAID 6 drive layouts and total reported usable storage ...............................................................4-30
4.2
RAID 10 drive layouts and total reported usable storage ............................................................. 4-32
4.3
RAID 50 drive layouts and total reported usable storage ............................................................ 4-33
4.4
RAID 5 drive layouts and total reported usable storage ............................................................... 4-34
4.5
Array RAID configurations and associated RAID sets ................................................................... 4-35
PS Series array concepts ............................................................................................................................ 5-36
5.1
Groups and pools ................................................................................................................................ 5-36
5.1.1
5.2
Pools .............................................................................................................................................. 5-36
Volumes ................................................................................................................................................ 5-38
5.2.1
Volume attributes ........................................................................................................................ 5-38
5.2.2
Volume features .......................................................................................................................... 5-39
5.3
Snapshots and clones ......................................................................................................................... 5-39
5.3.1
5.4
Clones ........................................................................................................................................... 5-40
Thin provisioning ................................................................................................................................. 5-40
5.4.1
6
Performance and availability characteristics of the supported RAID levels ...................... 3-27
Template volumes and thin clones .......................................................................................... 5-41
Array firmware features .............................................................................................................................. 6-42
6.1
Replication ............................................................................................................................................ 6-42
6.1.1
Replication limits ......................................................................................................................... 6-42
6.1.2
Replication paths ......................................................................................................................... 6-43
6.1.3
Replication process .....................................................................................................................6-44
6.1.4
Fast failback ..................................................................................................................................6-46
6.1.5
Sizing Replica Reserve and Delegated Space .........................................................................6-46
6.1.6
Effect of TCP Window Size on Latency across WAN links ...................................................6-48
6.1.7
Replication partner compatibility .............................................................................................6-48
6.1.8
Clustering ......................................................................................................................................6-49
March 2013
Dell EqualLogic Configuration Guide v14.1
v
6.2
6.2.1
About Synchronous replication ................................................................................................6-49
6.2.2
How Synchronous replication works.......................................................................................6-49
6.2.3
Synchronous states ..................................................................................................................... 6-50
6.2.4
Caveats: ......................................................................................................................................... 6-50
6.2.5
Requirements for using SyncRep ............................................................................................. 6-50
6.2.6
How SyncRep protects volume availability............................................................................. 6-51
6.2.7
SyncAlternate volume unavailable ........................................................................................... 6-51
6.2.8
Tracked changes written to the SyncAlternate volume ....................................................... 6-52
6.2.9
SyncActive volume unavailable................................................................................................. 6-53
6.3
7
Synchronous replication ....................................................................................................................6-49
Protecting your EqualLogic group with Internet Protocol Security........................................... 6-56
6.3.1
Types of protected traffic .......................................................................................................... 6-56
6.3.2
Protected Intra-Group Traffic ................................................................................................... 6-57
6.3.3
IPsec and Replication ................................................................................................................. 6-57
6.3.4
About IPsec Security Parameters.............................................................................................. 6-57
6.3.5
About IPsec Certificates ............................................................................................................. 6-57
6.3.6
About IPsec Pre-Shared Keys .................................................................................................... 6-58
6.3.7
About IPsec policies .................................................................................................................... 6-58
6.3.8
IPsec considerations and limitations ....................................................................................... 6-58
6.3.9
Performance considerations ..................................................................................................... 6-59
6.3.10
Host Connectivity Considerations ........................................................................................... 6-59
EqualLogic SAN design............................................................................................................................... 7-60
7.1
General requirements ......................................................................................................................... 7-60
7.1.1
Implementation of standards .................................................................................................... 7-60
7.1.2
General requirements and recommendations ....................................................................... 7-61
7.1.3
Quality of service (qos) ............................................................................................................... 7-62
7.2
Ethernet switches and infrastructure ............................................................................................... 7-62
7.2.1
Connecting SAN switches in a Layer 2 network .................................................................... 7-64
7.2.2
Sizing inter-switch connections ............................................................................................... 7-67
7.2.3
Comparing inter-switch connection types ............................................................................ 7-67
7.3
Building a high-availability SAN ........................................................................................................ 7-68
7.3.1
Design guidelines for host connectivity in a redundant SAN .............................................. 7-69
7.3.2
Multi-path I/O .............................................................................................................................. 7-70
7.3.3
Equallogic iSCI SAN Design ........................................................................................................ 7-71
March 2013
Dell EqualLogic Configuration Guide v14.1
vi
8
7.3.4
Redundant SAN configuration .................................................................................................. 7-72
7.3.5
Partially redundant SAN configurations .................................................................................. 7-74
7.3.6
Minimum cabling scenarios: PS4100 and PS6100 ................................................................ 7-75
Mixed speed environments - Integrating 1GbE and 10GbE SANs ...................................................... 8-78
8.1
Design considerations ........................................................................................................................ 8-78
8.1.1
8.2
9
Optimizing Rapid Spanning Tree Protocol behavior ............................................................ 8-79
Mixed speed SAN best practices ...................................................................................................... 8-80
Blade server chassis integration ............................................................................................................... 9-82
9.1
Designing a SAN using blade chassis I/O modules with arrays directly attached ...................9-84
9.1.1
9.2
SAN design for multiple M1000e enclosure .......................................................................... 9-90
Designing a SAN using blade Pass-through module ....................................................................9-94
9.3
Designing a SAN using blade chassis I/O modules as host access to external switches for array
connection .......................................................................................................................................................9-96
10
Fluid File system................................................................................................................................... 10-103
10.1
FS Series architecture ..................................................................................................................... 10-104
10.1.1
10.2
11
FS Series solution for file only storage................................................................................. 10-104
Dell FluidFS ................................................................................................................................... 10-105
FS Series NAS Appliances ....................................................................................................................11-107
11.1
Equallogic NAS appliance supported configuration limits........................................................11-107
11.2
Initial NAS cluster valid configurations ........................................................................................ 11-109
11.3
FS7500 system components .......................................................................................................... 11-110
11.4
FS7500 file system operation on controller failover .................................................................. 11-111
11.5
FS7600 components ....................................................................................................................... 11-111
11.6
FS7610 components ........................................................................................................................ 11-112
12
FS Series file level operations ............................................................................................................. 12-113
12.1
NAS cluster ........................................................................................................................................ 12-113
12.2
NAS reserve ....................................................................................................................................... 12-113
12.2.1
12.3
Relationship between PS Series groups, pools and NAS reserve..................................... 12-113
NAS Container ..................................................................................................................................12-114
12.3.1
NAS Container security styles ................................................................................................ 12-115
12.4
NAS Container snapshots ...............................................................................................................12-116
12.5
NAS Snapshots and Replication ..................................................................................................... 12-117
13
12.5.1
Snapshots................................................................................................................................... 12-117
12.5.2
Replication ................................................................................................................................. 12-117
FS Series NAS Configuration ............................................................................................................. 13-120
March 2013
Dell EqualLogic Configuration Guide v14.1
vii
13.1
FS7500 connection paths .............................................................................................................. 13-120
13.2
FS7600/7610 connection paths ................................................................................................... 13-122
14
Data Center Bridging (DCB) .............................................................................................................. 14-126
14.1
DCB Overview .................................................................................................................................. 14-126
14.2
DCB requirements for EqualLogic ............................................................................................... 14-129
14.3
Methods for configuring DCB ....................................................................................................... 14-130
14.4
Basic Deployment Topology Example .....................................................................................14-131
14.5
Tested SAN designs ..........................................................................................................................14-131
14.5.1
Blade IOM switch only ............................................................................................................ 14-132
14.5.2
ToR switch only ....................................................................................................................... 14-132
14.5.3
Blade IOM switch with ToR switch ...................................................................................... 14-133
14.6
Data Center In A Chassis DCB design ..................................................................................... 14-134
14.7
VLANs for iSCSI ................................................................................................................................ 14-135
Appendix A
Network ports and protocols ............................................................................................ 14-137
A.1
Required ports and protocols ....................................................................................................... 14-137
A.2
Optional ports and protocols ........................................................................................................ 14-137
Appendix B
Recommended switches ................................................................................................... 14-139
Appendix C
Supported iSCSI initiators................................................................................................... 14-140
Appendix D
Upgrade paths for EqualLogic PS SeriesArrays .............................................................. 14-141
Related publications ....................................................................................................................................... 14-142
March 2013
Dell EqualLogic Configuration Guide v14.1
viii
Introduction
With the Dell™ EqualLogic™ PS Series of storage arrays Dell provides a storage solution that delivers
the benefits of consolidated networked storage in a self-managing, iSCSI storage area network (SAN)
that is affordable and easy to use, regardless of scale. By eliminating complex tasks and enabling fast
and flexible storage provisioning, these solutions dramatically reduce the costs of storage acquisition
and ongoing operations.
To be able to leverage the advanced features provided by an EqualLogic array, a robust, standardscompliant iSCSI storage area network (SAN) infrastructure must be created. While almost any industry
standard, enterprise class Ethernet switch can be used to connect hosts to the EqualLogic-based
virtual storage volumes, some switches, network interface controllers(NICs), iSCSI Host Bus Adapters
(HBAs), and iSCSI initiators can directly impact the overall performance of the SAN storage solution.
The Dell EqualLogic Configuration Guide is provided as an aid to help storage administrators
determine how best to build an iSCSI infrastructure for use within an EqualLogic SAN solution. This
document focuses on network configuration, host integration, and other topics that help to ensure a
smooth deployment with optimum PS Series SAN and FS Series NAS appliance performance.
Note: This version of the EqualLogic Configuration Guide (v13.3, September 2012)
introduces coverage of the EqualLogic PS-M4110 Blade Array. Unless otherwise stated,
the recommendations involving specific 10GbE solutions apply to the first generation
PS6010, the second generation PS4110, PS-M4110, and the PS6110.
Unless otherwise stated, recommendations in this document are applicable to both file and block
environments.
This document should be used strictly as a guide in planning an EqualLogic storage solution. It should
not be considered a statement of support for any specific configuration. Actual viability of any
configuration will depend on the capabilities of the individual components (switches, network
interface controllers, initiators, etc.) that make up the SAN infrastructure.
Audience
This configuration guide is intended for storage administrators, SAN/NAS system designers, storage
consultants, or anyone who is considering purchasing or has purchased EqualLogic PS Series storage
or FS Series Array appliances for use in a production storage area network. It is assumed that all
readers have experience in designing and/or administering a shared storage solution. Also, there are
some assumptions made in terms of familiarity with all current and possibly future Ethernet standards
as defined by the Institute of Electrical and Electronic Engineers (IEEE) as well as all TCP/IP and iSCSI
standards as defined by the Internet Engineering Task Force (IETF).
March 2013
Dell EqualLogic Configuration Guide v14.1
1-1
1
PS Series storage arrays
PS Series Storage SANs provide a peer storage architecture comprised of one or more independent
arrays. Each array contains its own controllers, cache, storage, and interface ports. Grouped together
they can create one or more single instance storage pools that are based on the IETF iSCSI standard.
The capabilities of the PS Series solution are provided through a unique set of software and hardware
capabilities that, when used cooperatively, can provide a full featured SAN solution. The following
sections provide additional insight into specifications and functions of PS Series arrays.
1.1
Array models
The array models available prior to introduction of the PS4100/PS6100 family are shown in Table 1
below.
Table 1
Array Models prior to PS4100/PS6100
Array Model
Drive Type
Number of Drives
PS50E – PS2400E
SATA
14, 16
PS3000X, PS3x00XV
SAS
16
PS4000E
SATA
16
PS4000X, PS4000XV
SAS
16
PS5000E
SATA
16
PS5000X, PS5000XV
SAS
16
PS5500E
SATA
48
PS6000E
SATA
16
PS6000X, PS6000XV
SAS
16
PS6000S, PS6010S
SSD
16
PS6010E
SATA
16
PS6010X, PS6010XV
SAS
16
PS6000XVS, PS6010XVS
SAS / SSD
8/8
PS6500E
SATA
48
PS6510E
SATA
48
PS6510X
SAS
48
PS6500ES
SAS / SSD
41 SAS + 7 SSD
PS6510ES
SAS / SSD
41 SAS + 7 SSD
(a)
(a)
PS2400E comprised of four drive enclosures with 14 drives each
Starting with the introduction of the PS4100 and PS6100 family of arrays, configurations using 2.5” and
3.5” disks are available.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-2
Table 2
PS4100/PS6100 array models
1.2
Array model
Drive type
Number of drives
PS4100E
3.5” SAS 7.2K RPM
12
PS4100X
2.5” SAS 10K RPM
24
PS4100XV
2.5” SAS 15K RPM
24
PS4100XV
3.5” SAS 15K RPM
12
PS6100E
3.5” SAS 7.2K RPM
24
PS6100X
2.5” SAS 10K RPM
24
PS6100XV
2.5” SAS 15K RPM
24
PS6100XV
3.5” SAS 15K RPM
24
PS6100S
SSD
12 or 24
PS6100XS
SSD + SAS 10K RPM
7 SSD + 17 SAS
PS4110E
3.5” SAS 7.2K RPM
12
PS4110X
2.5” SAS 10K RPM
24
PS4110XV
2.5” SAS 15K RPM
24
PS4110XV
3.5” SAS 15K RPM
12
PS6110E
3.5” SAS 7.2K RPM
24
PS6110X
2.5” SAS 10K RPM
24
PS6110XV
2.5” SAS 15K RPM
24
PS6110XV
3.5” SAS 15K RPM
24
PS6110S
SSD
12 or 24
PS6110XS
SSD + SAS 10K RPM
7 SSD + 17 SAS
PS-M4110E
7.2K RPM NL-SAS 2.5”
14
PS-M4110X
10K RPM SAS 2.5”
14
PS-M4110XV
15K RPM SAS 2.5”
14
PS-M4110XS
SSD + 10K RPM 2.5” SAS
5 SSD + 9 SAS
PS Series supported configuration limits
The Dell EqualLogic PS6xxx Series provides the full range of features and capabilities available with the
EqualLogic PS Series storage products. The Dell EqualLogic PS4xxx Series provides a subset of features
and capabilities, targeted at remote office and small to medium business storage deployments. The
supported configuration limits for a PS Series group provided in Table 3.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-3
Table 3
Supported configuration limits
PS4000/PS4100 and
a
PS-M4110 groups only
Configuration
Volumes and replica sets per group
Volume size
512
c
h
Volumes enabled for replication (outbound)
15 TB
32
256
2048
10,000
Snapshots per volume
128
512
Replicas per volume
128
512
Volumes that have Synchronous Replication
Enabled
4
32
Schedules (snapshot or replication) per volume
or volume collection
16
16
Persistent Reservation registrants per volume
96
96
Replication partners per group
16
16
Replication partners per volume
1
1
Members per group
2
16
Members per pool
2
8
Pools per group
2
4
Volumes per collection
8
8
Snapshots and replicas per group
Collections per group (snapshot and replication)
b
1024
15 TB
d
All other groups
a
100
100
g
h
Volume connections (each time an iSCSI
initiator connects to a volume counts as a
e,f
connection)
512 per pool
1024 per group with
2 pools
1024 per pool
4096 per group with
4 pools
Access control records per volume and its
snapshots
16
16
Simultaneous management sessions (any
combination of GUI, telnet, or scripting sessions)
7
7
10% of volume size
10% of volume size
100
100
i
Thin provisioning limits
(minimum allocation)
Administrator accounts per group
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
A group can contain a maximum of two PS4000, PS4100, and/or PS-M4110 arrays.
Includes groups of mixed array types and all other group types except groups containing only PS40x0, PS41x0, and
PS-M4110 arrays. When a group contains a mix of PS40x0, PS41x0, and PS-M4110 arrays and other array models,
the higher limits prevail.
Practical maximum volume size is operating system-specific. A PS Series group can create and present volumes up
to 15 TB.
Up to 16 simultaneous replication operations can take place.
To avoid service outages or possible loss of connectivity during failovers, Dell recommends increasing initiator
timeouts for groups with more than 512 connections. See the iSCSI Initiator and Operating Systems Considerations
document for more information.
Inbound replication connections count towards the total number of connections.
With firmware version 5.1, maximum number of volumes per group and maximum number of connections per pool
increased from 256 to 512
Up to 1024 connections per pool are supported for all groups except groups containing only PS40x0, PS41x0, and
PS-M4110 arrays. This provides increased scalability in virtualized environments. These connections may be
distributed across a maximum of 512 volumes.
Volumes automatically created and managed by the NAS Service inside the NAS Reserve pool used by an FS Series
appliance are fully allocated at creation and cannot be thin provisioned. Once those volumes are created (when
the NAS Reserve is created they cannot be shrunk.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-4
1.3
Array models prior to PS4100/PS6100
Since the EqualLogic PS Series was introduced, there have been several different array models
released with new features, better performance and greater storage capacity. The storage array
controllers were also improved to take advantage of advances in the underlying networking and
storage technologies.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-5
1.3.1
Controller types in all models prior to PS4100/PS6100
Array controllers can be identified and differentiated by the controller "type" designation. Each
controller type will have a different colored label to help quickly identify the controller type. Table 4
lists each Dell EqualLogic controller along with some characteristics about each.
Table 4
Array controller types – All models prior to PS4100/PS6100
Controller
type
March 2013
Network
interfaces
Storage
type
Type 1
3 x 1GbaseT
3 x 1Gb SFP
(combo)
SATA
Original Controller
Design
PS50 – PS2400
1GB Cache
Type 2
3 x 1GbaseT
3 x 1Gb SFP
(combo)
SATA
PS50 – PS2400
1GB Cache
Type 3 SAS
Type 3
SATA
3 x 1GbaseT
SAS
SATA
PS3000 – PS5000
1GB Cache
Cannot mix Type 3
SAS with Type 3 SATA
Type 4
3 x 1GbaseT
SAS
PS3000 – PS5000
1GB Cache
Cannot mix Type 3
SAS with Type 4
controller
Type 5
3 x 1GbaseT
SATA
PS3000 – PS5000
1GB Cache
Cannot mix Type 3
SAS with Type 5
controller
Type 6
3 x 1GbaseT
SATA
PS5500 only
2GB Cache
Type 7
4 x 1GbaseT
SAS
SATA
SSD
PS6000 – PS6500
2GB Cache
Cannot mix SAS and
SATA drives in one
array
Type 8
2 x1GbaseT
1 x 10/100Mb
mgt
SAS
SATA
PS4000 only
2GB Cache
Cannot mix SAS and
SATA drives in same
array
Faceplate
Dell EqualLogic Configuration Guide v14.1
Notes
1-6
Controller
type
1.3.2
Network
interfaces
Storage
type
Type 9
2 x 1GbaseT
1 x 10/100Mb
mgt
SAS
SATA
2 generation
PS4000
2GB Cache
Cannot mix SAS and
SATA drives in same
array
Type 10
2 x 10GB
SFP+
1 x 10/100Mb
mgt
SAS
SATA
SSD
10Gb Ethernet
PS6010 – PS6510
2GB Cache
Faceplate
Notes
nd
Controller redundancy in all models prior to PS4100/PS6100
Each array can be configured with either a single controller, or dual redundant controllers. The single
controller configuration will provide the same level of I/O performance as a dual controller
configuration. The dual controller configuration provides for redundancy. Redundant controllers will
prevent volume connections between hosts and SAN from being dropped in the event of an active
controller failure.
The Active Controller is the controller which is processing all disk and network I/O operations for the
array. A second controller in dual controller configurations will always be in a “passive” operating
mode. In this mode, the secondary controller will exhibit the following characteristics:
•
•
Each of its Ethernet ports are electrically inactive (active lights are off)
The passive controller cache mirrors the cache of the active controller.
1.3.3
Controller failover behavior in all models prior to
PS4100/PS6100
To support redundant controller fail over, each Ethernet port on the active controller that is connected
to the SAN must have its corresponding port on the passive controller also connected to the same
SAN network. In the event of a controller failure, the passive controller will immediately activate and
continue to process all data requests to the array. The following changes occur during fail over:
•
•
The backup controller automatically enables each of the corresponding Ethernet ports that
were enabled on the failed primary controller
The IP addresses that were assigned to each of the failed controller Ethernet ports are
reassigned to the corresponding ports on the second controller.
A link failure on one or more of the Ethernet ports on the active controller does not constitute a
controller failure. For this reason, it is important to ensure that ports from each controller are
connected to at least two different switches. This will prevent a switch failure from also disabling all
paths between a host and its connected volumes.
It is also critical that port connections for both controllers are configured so that the corresponding
ports on each controller are connected to the SAN. If port 0 on the active controller is the only port
March 2013
Dell EqualLogic Configuration Guide v14.1
1-7
connected to the SAN, then you must use port 0 on the passive controller for the other connection to
the SAN. This is illustrated in the partially connected scenario shown in Figure 1.
Figure 1 Partially connected controller failover
Note how IP addresses are reassigned to the ports during the failover processes shown in Figure 1 and
Figure 2.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-8
Figure 2 Fully connected controller failover
March 2013
Dell EqualLogic Configuration Guide v14.1
1-9
1.4
Array models PS4100/PS6100
1.4.1
Controller types in PS4100/PS6100 models
The new controller types available in the PS4100 and PS6100 model arrays became available starting
in August 2011. Table 5 lists each Dell EqualLogic controller along with some characteristics.
Table 5
PS4100/PS6100 controller types
Controller
Type
March 2013
Network
Interfaces
Storage
Notes
Type
Type 11
4 x 1GbaseT
1x
10/100Mb
mgt
SAS
NL-SAS
SSD
PS6100 Only
4GB Cache
Supports vertical
port failover
Cache to Flash
memory
destaging
Type 12
2 x 1GbaseT
1x
10/100Mb
mgt
SAS
NL-SAS
PS4100 Only
4GB Cache
Supports vertical
port failover
Cache to Flash
memory
destaging
Type 14
1 x 10GbE
SFP+
1x
10GbaseT
10/100Mb
mgt
SAS
NL-SAS
SSD
PS6110 Only
4GB Cache
Support for
vertical port
failover
Cache to Flash
memory
destaging
SFP+ OR
10GBaseT used
at any given time
Type 17
1 x 10GbE
SFP+
1x
10GbaseT 1
x 10/100Mb
mgt
SAS
NL-SAS
PS4110 Only
4GB Cache
Supports vertical
port failover
Cache to Flash
memory
destaging
SFP+ OR
10GBaseT used
at any given time
Faceplate
Dell EqualLogic Configuration Guide v14.1
1-10
1.4.2
Controller redundancy in PS4100/PS6100 controllers
Each array can be configured with either a single controller, or dual redundant controllers. The single
controller configuration will provide the same level of I/O performance as a dual controller
configuration. The dual controller configuration provides for redundancy. Redundant controllers will
prevent volume connections between hosts and SAN from being dropped in the event of an active
controller failure.
The Active Controller is the controller which is processing all disk and network I/O operations for the
array. A second controller in dual controller configurations will always be in a “passive” operating
mode. In this mode, the secondary controller will exhibit the following characteristics:
•
•
Each of its Ethernet ports are electrically inactive (active lights are off), unless a vertical port
failover has occurred.
The passive controller cache mirrors the cache of the active controller.
1.4.3
Controller failover behavior: PS41x0/PS61x0
In the event of a controller failure the following changes occur:
•
•
The passive controller immediately activates and continues to process all data requests to the
array.
Vertical port pairing insures that IP addresses assigned to each of the failed controller Ethernet
ports apply to the corresponding ports on the second controller.
As stated in Section 1.3.3 above, controller models prior to the PS4100/PS6100 required having cabled
connections present on corresponding controller ports for controller failover to work correctly. The
vertical port failover feature in PS41x0/PS61x0 controllers eliminates this requirement, enabling
partially cabled configurations to support controller failover too. See Section 7.3.6, Minimum cabling
scenarios: PS4100 and PS6100, for details.
It is important to ensure that ports from each controller are connected to at least two different
switches. This will prevent a switch failure from also disabling all paths between a host and its
connected volumes. We illustrate controller failover behavior for the PS4100 family controller in Figure
3. Controller failover behavior for the PS6100 (4 port) controller family is identical.
IMPORTANT: To prevent a switch failure from also disabling all paths between a host and
its connected volumes, you should make sure that ports from each controller are
connected to at least two different switches. You should also split vertical port pair
connections between two switches to ensure 100% bandwidth capability is maintained
in the event of a vertical port failover event. Both of these guidelines are illustrated in
Figure 3.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-11
Figure 3 Controller failover process and optimal connection paths
1.4.4
Vertical port failover behavior in PS4100/PS6100 controllers
In PS Series controllers prior to PS4100/6100 families, a link failure or a switch failure was not
recognized as a failure mode by the controller. Thus a failure of a link or an entire switch would
reduce bandwidth available from the array. Referring to Figure 4 or Figure 5, assume that CM0 is the
active controller. In vertical port failover, if CM0 senses a link drop on the local ETH0 port connection
path, it will automatically begin using the ETH0 port on the backup controller (CM1). Vertical port
failover is bi-directional. If CM1 is the active controller then vertical port failover will occur from CM1
ports to CM0 ports if necessary.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-12
Figure 4 PS4100 vertical port failover
Figure 5 PS6100 vertical port failover
With PS4100/PS6100 family controllers, vertical port failover can ensure continuous full bandwidth is
available from the array even if you have a link or switch failure. This is accomplished by combining
corresponding physical ports in each controller (vertical pairs) into a single logical port from the point
of view of the active controller. In a fully redundant SAN configuration, you must configure the
connections as shown in Figure 24 and Figure 25 in section 7.3.4.
In a redundant switch SAN configuration, to optimize the system response in the event you have a
vertical port failover you must split the vertical port pair connections between both SAN switches.
The connection paths illustrated in Figure 6 and Figure 7 show how to alternate the port connection
paths between the two controllers. Also note how IP addresses are assigned to vertical port pairs.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-13
Figure 6 PS4100 vertical port failover and optimal connection paths
IMPORTANT: By alternating switch connection paths between ports in a vertical port pair,
port failover allows the array to maintain 100% bandwidth capability in the event of a
switch failure.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-14
Figure 7 PS6100 vertical port failover process and optimal connection paths
1.4.5
Vertical port failover behavior in PS4110/PS6110 controllers
In PS Series controllers prior to PS4110/6110 families, a link failure or a switch failure was not
recognized as a failure mode by the controller. This caused a failure of a link or an entire switch to
reduce bandwidth available from the array. Referring to Figure 4 or Figure 5, assume that CM0 is the
active controller. In vertical port failover, if CM0 senses a link drop on the local ETH0 port connection
path, it will automatically begin using the ETH0 port on the backup controller (CM1). Vertical port
failover is bi-directional. If CM1 is the active controller then vertical port failover will occur from CM1
ports to CM0 ports if necessary.
Note: The PS4110 and PS6110 PS Series Arrays do not support dual speed connectivity.
They are not designed to support the auto-negotiate feature. The connectivity must be
10Gb ONLY - This is the only supported configuration.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-15
Figure 8 4110/6110 vertical port failover
With the PS4110/PS6110 family of controllers, vertical port failover can ensure continuous full
bandwidth is available from the array even if you have a link or switch failure. This is accomplished by
combining 10GbE “eth0” ports in each controller into a single logical port from the point of view of the
active controller. In a fully redundant SAN configuration, you must configure the connection as shown
in Figure 9.
In a redundant switch SAN configuration, to optimize the system response in the event you have a
vertical port failover you must connect either the SFP+ cable or the 10GbaseT cable from each
controller to a different switch in the SAN network. While it is a supported configuration, it is not
recommended to connect both the SFP+ and 10GbaseT ports on each controller at the same time. In
this scenario, the SFP+ port will always be the preferred active port and this preference cannot be
changed by the user. This preference is not guaranteed in future array firmware releases.
The connection paths illustrated in Figure 9 show how to connect the port connection paths between
the two controllers. Also note how IP addresses are assigned to the vertical port pair.
March 2013
Dell EqualLogic Configuration Guide v14.1
1-16
Figure 9 4110/6110 Vertical port failover scenario
1.5
Array model PS-M4110
1.5.1
Controller type in PS-M4110 model
The PS-M4110 controller is designed based on a modified version of the PS4100 Controller.
Host and SAS cards combined to form a single unit, fixed I/O module, connecting to the M1000e
chassis infrastructure.
Controller Storage blade
Type
image
Type 13
March 2013
Network Interfaces
2 x 10Gb Ports, (One Per
Controller), Connected
through the Backplane
(No Label)
Each of the two ports
has an active link and a
standby link to the fabric
switches in the
backplane of the
M1000e chassis.
Dell EqualLogic Configuration Guide v14.1
Storage
type
SAS
NL-SAS
Notes
Dual, hotpluggable 10GbE
controllers
4GB of memory
per controller,
Cache to Flash
Design
1 x dedicated
management port
– accessible via
CMC
14x 2.5” 6Gb/s
SAS disk drives
1-17
1.5.2
Configuration options
The PS-M4110 has four basic configuration options. It can be configured on Fabric A or Fabric B, and
each Fabric Configuration can use a 10Gb KR switch or a 10Gb KR Pass-Thru Module (PTM). Figure 10
depicts a basic configuration using MXL switch, however any of the supported switches can be used in
this configuration.
S
A
N
S
A
N
LAN
LAN
SAN Stack
LAN/Client Network
SAN Stack
LAN-to-Agg Uplinks
Figure 10 Basic PS-M4110 configuration for data center-in-a-box
1.5.3
Failure behavior in the PS-M4110
Each PS-M4110 array is configured with dual controllers, active and passive connections, and
redundant fabrics to provide redundancy and prevent volume connections between hosts and SAN
from being dropped in the event of a failure.
There are three failure scenarios that are protected by the array:
•
•
•
In the event of a link failure, the active link of the active port goes down, the standby
link for the same port will take over.
In the event of a Switch failure, both active and passive ports will automatically link to the
redundant fabric switch.
If the active controller fails, the passive controller will takeover.
1.5.4
Networking considerations and guidelines
Supported M-Series I/O modules
March 2013
Dell EqualLogic Configuration Guide v14.1
1-18
•
•
•
10G KR is the only supported IOM
Switches: Force10 MXL, PowerConnect M8024-k, M8428-k
Pass-through: 10Gb-K pass-thru modules only with external switches
The following are basic networking recommendations for implementing the PS-M4110 Storage Blade.
•
•
•
•
•
•
•
IOMs must be Interconnected
Datacenter in a Box: IOMs directly interconnected
External Switches can be used to provide interconnection if rack mounted arrays are needed.
Must use “Single Fabric” I/O module Placement; Supported Fabrics are “A” and “B” only
Fabric A not supported on older M1000e systems (Mid-plane v1.0); Mid-plane upgrade kit available
Fabric B supported on all M1000e system
It is recommended that PS-M4110 be placed into their own pool
March 2013
Dell EqualLogic Configuration Guide v14.1
1-19
2
Controller firmware
2.1
About member firmware
Each control module in a group member must be running the same version of the PS Series firmware.
Firmware is stored on a compact flash card or a microSD card on each control module.
Dell recommends the following:
•
•
•
•
•
Always run the latest firmware to take advantage of new features and fixes.
All group members must run the same firmware version. If you are adding a new array to a group,
update the group to the latest firmware before adding the new member.
Dell does not support groups running with mixed-firmware versions, except when you are in the
process of upgrading the firmware of the entire group, and then only for as long as it takes to do
the upgrade.
You can upgrade member firmware to a higher version or downgrade member firmware to a
lower version. In some firmware releases, firmware downgrades are not allowed. See the Dell
EqualLogic PS Series Storage Arrays Release Notes for version-specific information about the
firmware running on your group member(s).
For instructions on how to update the firmware, see the Updating Firmware for Dell EqualLogic PS
Series Storage Arrays, available at the EqualLogic customer support site
(http://www.equallogic.com/support).
Each EqualLogic PS Series array runs a core operating system in firmware that provides all of the PS
Series features and functionality. The firmware version is defined using a version number and will be
updated from time to time as new features are added or for general maintenance improvements.
The firmware version number takes the following form: "X.Y.Z":
•
•
•
"X" is used to identify the "major" release number. This changes when there are very significant
changes in the features and functionality.
“Y" is the "minor" release number. This changes as new minor features or improvements in
existing features are made for a given major release.
"Z" denotes maintenance releases. This changes whenever a maintenance release level is
released to resolve any identified technical or usability issues.
2.2
Firmware upgrade considerations
Before beginning a firmware upgrade process, you must review the Release Notes for the firmware
update revision. The release notes will be included with the firmware update kit. All firmware update
kits are available at http://www.equallogic.com/support/ (Support ID required for login access). The
release notes contain the latest information and guidelines for:
•
•
•
Completing the upgrade with minimal potential for service disruption
Operating system and iSCSI initiator configuration settings that will minimize potential for I/O
disruption during the upgrade process
Latest updates to the supported firmware upgrade paths
March 2013
Dell EqualLogic Configuration Guide v14.1
2-20
In addition to the Release Notes, the process for updating controller firmware is described in detail in
the following document (Support ID required for login access):
•
•
•
Table 6
PS Series Storage Arrays: Updating Storage Array Firmware, available at:
https://www.equallogic.com/support/download_file.aspx?id=594
Supported firmware upgrade paths (up to version 6.0.x) are shown in Table 6 below.
If you are starting with v4.2.* or v4.3.* then you can update straight to v5.0.4. If the array is
already running v5.0 or v5.0.1 then you must first update to v5.0.2 before updating to v5.0.4.
Controller firmware upgrade paths
March 2013
Starting firmware
version
Ending firmware
version
5.2.x
Later 5.2.x releases
6.0.x
5.1.x
Later 5.1.x releases
5.2.x
5.0.x
(excluding 5.0.0 and
5.0.1)
Later 5.0.x releases
5.1.x, 5.2.x
5.0.0, 5.0.1
5.0.2
4.3.x
Later 4.3.x releases
5.0.x (excluding 5.0.0 & 5.0.1)
4.2.x
Later 4.2.x releases
4.3.x
5.0.x (excluding 5.0.0 & 5.0.1)
4.1.x
Later 4.1.x releases
4.2.x, 4.3.x
4.0.x
Later 4.0.x releases
4.1.x
3.3.x
Later 3.3.x releases
4.0.x
3.2.x
Later 3.2.x releases
3.3.x
4.0.x
3.1.x
Later 3.1.x releases
3.2.x, 3.3.x
3.0.x
Later 3.0.x releases
3.1.x
Dell EqualLogic Configuration Guide v14.1
2-21
2.2.1
PS Series Firmware Compatibility with EqualLogic Tools
The following table provides a quick reference of EqualLogic product version compatibility for the
recent major firmware releases.
Table 7
PS Series firmware compatibility
Firmware
V6.0.x
Firmware
V5.2.x
Firmware
V5.1.x
Firmware
V5.0.x
Firmware
V4.3.x
SAN HeadQuarters
2.2.x
2.2.x
2.1.x
2.2.x
2.1.x
2.2.x
2.1.x
2.2.x
2.1.x
Host Integration Tools
for Microsoft
4.0.x
4.0.x
3.5.x
4.0.x
3.5.x
4.0.x
3.5.x
3.4.x
3.3.x
4.0.x
3.5.x
3.4.x
3.3.x
3.5.1
Host Integration Tools
for VMware
3.1.2
3.1.x
3.0.x
3.1.x
3.0.x
2.0.x
3.1.x
3.0.x
2.0.x
2.0.x
EqualLogic Storage
Replication Adapter for
VMware Site Recovery
Manager
2.1.x
2.1.x
2.0.x
1.0.x
2.1.x
2.0.x
1.0.x
2.1.x
2.0.x
1.0.x
1.0.x
EqualLogic
Multipathing Extension
Module for VMware
vSphere
1.1.1
1.1.x
1.0.x
1.1.x
1.0.x
1.1.x
1.0.x
1.1.x
1.0.x
Host Integration Tools
for Linux
1.1.x
1.1.x
1.0.x
1.1.x
1.0.x
1.1.x
1.0.x
1.0.x
Manual Transfer Utility
1.2.3
1.2.1
1.2.1
1.2.1
1.1.2
(Windows
only)
1.0.2 (Linux
only)
1.2.1
1.1.2
(Windows
only)
1.0.2 (Linux
only)
1.0.6
1.1.0
1.2.1
March 2013
Dell EqualLogic Configuration Guide v14.1
2-22
2.3
Optimizing for High Availability and preparing for Array
Firmware updates
Business critical data centers must be designed to sustain various types of service interruptions. In an
EqualLogic SAN environment, software, servers, NICs, switches, and storage are all interdependent
variables that need tuning and configuration with best practices in order to ensure high availability (HA)
and non-disruptive operations.
Plan a maintenance window for upgrades if all configuration is not done for proper failover.
Causes of service disruptions:
•
•
•
Hardware failure – Server, NIC, Cable, Switch, Storage, Controller, Disk, and others
Software failure – Bugs or Defects, Application Crashes, and others
Firmware upgrades – Software Patches, Driver Updates, Firmware Updates
A redundant design will survive most of the possible types of service disruptions, but in some cases
redundancy alone is not enough. The iSCSI data path is the most crucial point of failure in an
EqualLogic SAN environment, and is the focus of this section.
Configuring and tuning your environment with best practice recommendations to design for High
Availability and Non-Disruptive Operations can prevent:
•
•
•
connections to volumes dropping
application timeout
group and/or array failure
2.3.1
Verify hosts are designed for high availability
Designing and configuring hosts for high availability helps minimize the chance of disruptions.
Recommended settings and configuration options on hosts running in the EqualLogic
environment are detailed below:
•
•
Install EqualLogic Host Integration Tools (HIT) on all supported operating systems.
The host software modules change recommended settings for iSCSI Disk Timeouts.
2.3.2
Microsoft Windows Servers
The Windows Disk I/O timeout parameter is set automatically when EqualLogic HIT for Windows is
installed. The TimeOutValue Parameter can be increased to avoid disk I/O timeouts. Verify that the
Windows Disk I/O timeout parameter in the Registry Editor is set to no less than 60 seconds.
2.3.3
VMWare ESX Servers
In vSphere 5.x, VMware allocates five seconds for an iSCSI session to login to an iSCSI target. In a
normal environment this period of time is sufficient. However, it has been observed that this five
second timeout is not long enough to process the large number of iSCSI login requests that occur
after a network failure, switch reboot, or controller failover.
It is recommend that the iSCSI Login Timeout value be increased to 60 seconds.
March 2013
Dell EqualLogic Configuration Guide v14.1
2-23
2.3.4
Storage Heartbeat on vSphere 5.0, 4.1, and 4.0
Note: This recommendation for using Storage Heartbeat applies only vSphere 4.1 and
5.0. It is not necessary with vSphere 5.1.
In the VMware virtual networking model, certain types of vmkernel network traffic is sent out on a
default vmkernel port for each subnet. If the physical NIC that is being used as the uplink for the
default vmkernel port goes down, network traffic that is using the default vmkernel port will fail,
including vMotion traffic, SSH access, and ICMP ping replies.
Even though iSCSI traffic is not directly affected by this condition, a side effect of the suppressed ping
replies is that the EqualLogic PS Series Group will not be able to accurately determine connectivity
during the login process, and therefore a suboptimal placement of iSCSI sessions can occur. This
could cause logins to not be completed in a timely manner.
It is recommended to have a high availability vmkernel port on the iSCSI subnet serving as the
default vmkernel port for such outgoing traffic.
2.3.5
Linux servers
The HIT Kit for Linux automatically implements recommended changes on the Linux hosts,
either at the time of installation or through the built-in “eqltune” utility
2.3.6
Network topology health check
Design your network for redundancy according to recommended best practices:
•
•
•
•
For all members (arrays) in a given SAN Group all ports should be connected to the same subnet.
Use at least two iSCSI SAN ports per host connected, so that at least two of the host ports connect
to two different switches in the SAN
At a minimum Eth0 from Controller0 and Controller1 on each array in the SAN group must be
connected to two different switches in the same SAN subnet.
All switches within the SAN must be interconnected such that there is always a path from any
Ethernet port on one array to all other Ethernet ports on all other arrays in the group.
Proper settings configured on switches:
•
•
•
All switches and host network controllers within the infrastructure must have flow control enabled
for optimal performance. For switches, RX flow control is required. For iSCSI hosts, both TX and
RX is required.
Rapid Spanning Tree must be enabled on all switch ports not directly connected to host or array
ports.
Enable PortFast on all switch ports connected to the host or array ports.
See section 7 for complete iSCSI SAN design recommendations.
2.3.7
Environmental health
Verify there are no current health and/or performance problems with the EqualLogic environment.
March 2013
Dell EqualLogic Configuration Guide v14.1
2-24
3
RAID policies
Each array in an EqualLogic array group is configured with a single RAID policy. Arrays (or group
members) within the same storage pool that have the same RAID policy will cooperatively work to
host volumes by distributing those volumes over multiple arrays. Two things that are defined by the
RAID policy are:
•
•
RAID level
hot-spare configuration
Each array implements a default RAID policy that includes a hot-spare. To configure a RAID policy that
does not include a hot-spare, you must manually initialize the array using the Command Line
Interface.
3.1
Setting the RAID Policy for a member
PS Series arrays protect data by using RAID technology and spare drives. After you add a member to a
PS Series group, you must set the RAID policy for the member. In multi-member groups, you can
assign different RAID levels to individual group members.
Once a RAID policy has been specified, the member automatically configures the disk drives according
to the designated RAID level and assigns the appropriate number of spare drives. The storage in the
member is available after you set the RAID policy, although array performance is not optimal until the
system has completed verification of the new RAID configuration.
The Remote Setup Wizard specifies the RAID policy for members when it creates a group and adds the
first member to it. See the Installation Guide for your array for information.
If you use the setup utility, you must manually configure the RAID policy as a separate step, using
either the Group Manager GUI or the CLI. See the online help or the Dell EqualLogic Group Manager
CLI Reference Guide for more information.
Note: Dell recommends against using RAID 5 for any business-critical data, although it
may be required for certain applications, depending on performance and data availability
requirements. RAID 5 can only be configured using the CLI.
Note: Configuring a RAID policy that does not include a hot-spare will increase the risk of
data loss in the event of multiple drive failures.
3.2
Guidelines for choosing a RAID policy
When specifying the RAID policy for a member, the following factors must be considered
March 2013
Dell EqualLogic Configuration Guide v14.1
3-25
3.2.1
RAID level
Each RAID level offers varying combinations of performance, data protection and capacity utilization.
Choose your RAID preference carefully based on reliability, capacity, and performance requirements.
PS Series arrays support the following RAID types:
• RAID 5 (not recommended for business-critical data)
• RAID 6
• RAID 6 Accelerated
• RAID 10
• RAID 50
See RAID Level Characteristics in Table 8 below for detailed information about the RAID types and
their recommended uses.
3.2.2
Drive configuration
The type of drive used in your system factors into choosing an optimal RAID configuration.
The types of drives used in EqualLogic PS Series arrays are listed below.
Class 0 drives are Solid State Drives (SSDs) that offer the highest level of performance. Typically, SSD
have smaller capacities than Class 1 or 2 drives.
Class 1 drives are 10,000 and 15,000 RPM SAS hard drives that offer higher performance and greater
reliability than Class 2 drives.
Class 2 drives are slower 7200 RPM SATA or Near-Line SAS hard drives that offer greater capacities
than Class 1.
If you do not know what types of drives are used in your array, see the Dell EqualLogic PS Series
Storage Arrays Release Notes, which lists the types of drives used with each PS Series model. In
addition, there is labeling on each drive to indicate its capacity and type.
3.2.3
Spare drive policy
Spare drives increase availability by providing redundancy in the event of a hard disk drive failure or a
solid-state disk drive failure. The number of spare drives depends on the array model and the number
of installed drives. For RAID 6 Accelerated, only one hard disk drive is configured as a spare. Spare
HDD drives can be used as spares for either HDD or SSD drives.
Dell does not recommend using any configuration that does not utilize spare drives, although these
configurations can be implemented using the CLI. For more information about using the CLI, refer to
the Dell EqualLogic Group Manager CLI Reference Guide.
3.3
RAID level characteristics
Table 8 summarizes best practices for RAID policy selection. When selecting a RAID policy, you should
consider the level of data protection you require, as well as I/O performance, capacity needs, and the
types of drives contained in your PS Series array.
March 2013
Dell EqualLogic Configuration Guide v14.1
3-26
Table 8
RAID level characteristics
3.3.1
Recommended drive
configurations
Raid policy
Recommended usage scenarios
RAID 10
Applications and workloads requiring the
highest levels of I/O performance for
random writes.
Systems containing 10K and
15K RPM drives.
RAID 6
Situations in which I/O performance for
random writes is not a key factor.
Applications requiring the highest levels
of data protection and reliability.
Systems containing 24 or
more drives.
Systems containing 7200
RPM SATA or Nearline SAS
(NL-SAS) drives.
RAID 6
Accelerated
Optimizes the use of solid-state drives
for critical data.
Supported only on arrays
that contain both solid state
(SSD) and hard disk (HDD)
drives. For these systems, it
is the only available RAID
policy. It is not a userselectable option.
RAID 50
Applications requiring an optimal
combination of performance, availability,
and capacity.
Systems containing fewer
than 24 drives.
Systems containing 10K and
15K RPM drives.
RAID 5
Dell recommends against using RAID 5
for any business-critical data, although it
may be required for certain applications,
depending on performance and data
availability requirements. RAID 5 carries
higher risks of encountering an
uncorrectable drive error during a
rebuild, and therefore does not offer
optimal data protection. RAID 5 can only
be configured using the CLI.
N/A
Performance and availability characteristics of the supported
RAID levels
Table 9 compares the performance and availability characteristics of the supported RAID levels. The
first column lists workload requirements, with the other columns respectively listing each RAID level's
characteristics with respect to each requirement.
March 2013
Dell EqualLogic Configuration Guide v14.1
3-27
Table 9
RAID Level Characteristic Comparison Workload Requirement
Workload
Requirement
RAID 5
RAID 6
RAID 10
RAID 50
Capacity
Excellent
Good
Average
Good
Availability
Poor
Excellent
Good
Fair
Sequential reads
Excellent
Excellent
Excellent
Excellent
Sequential writes
Good
Good
Good
Good
Random reads
Excellent
Excellent
Excellent
Excellent
Random writes
Good
Fair
Excellent
Good
Longest
Shortest
Medium
Performance impact
Longest
of drive failure or RAID
reconstruction
3.4
Supported RAID policy conversions
While a member remains online, you can convert it from one RAID policy to another only if the new
RAID policy provides the same or more space than the current policy.
The table below shows the supported RAID policy conversions. The first column lists RAID
configurations, and the second column lists supported RAID policy conversions for them.
Table 10 Supported RAID policy conversions
Current RAID policy
Supported conversion
RAID 10
All
RAID 50
RAID 5, RAID 6
RAID 5
None
RAID 6
None
RAID 6 Accelerated
None
If a RAID policy conversion is not supported, you can remove the member from the group and then
add it to the group again. You can then set the RAID policy.
Before changing a member's RAID policy, note the following:
•
•
Single member groups: If you change the RAID policy in a single member group, all data on the
member will be lost.
Multi-member groups: Before changing a group member's RAID policy, you must first vacate its
data to another member in the group, or all data on the member will be lost. When you vacate a
member, there must be sufficient available free space on another member to store the evacuated
data.
March 2013
Dell EqualLogic Configuration Guide v14.1
3-28
To convert from a no-spares RAID policy to a policy that uses spare drives, you must use the CLI. Refer
to the Dell EqualLogic Group Manager CLI Reference Guide. You can also use the CLI to convert to a
RAID policy that does not use spare drives, but Dell recommends against doing this.
March 2013
Dell EqualLogic Configuration Guide v14.1
3-29
4
Capacity planning
4.1
RAID 6 drive layouts and total reported usable storage
RAID6 (striped set with dual distributed parity) combines N disks in an arrangement where each stripe
consists of N-2 disks capacity for data blocks and two disks capacity for parity blocks. Each parity
block generates parity using a different view of the data blocks depending on the RAID 6
implementation. RAID 6 can tolerate up to two drive failures per RAID stripe set at the same time
without data loss. RAID 6 is not recommended for workloads consisting mainly of random writes.
Using a RAID 6 policy, Table 11 shows the drive layouts that are enforced based on the number of
drives in each array/hot spare configuration, and the total usable storage available for each model.
March 2013
Dell EqualLogic Configuration Guide v14.1
4-30
Table 11 RAID 6 drive layouts and total storage available with hot spares (in GB)
Disk drives
Hot spare
No hot spare
6
5 Data/Parity + 1 Hot-spare
6 Data/Parity
7
6 Data/Parity + 1 Hot-spare
7 Data/Parity
7 Data/Parity + 1 Hot-spare
8 Data/Parity
12
11 Data/Parity + 1 Hot-spare
12 Data/Parity
14
13 Data/ Parity + 1 Hot-spare
14 Data/Parity
15 Data/ Parity + 1 Hot-spare
16 Data/Parity
15 Data/Parity(d) + 1 Hot-spare
16 Data/Parity
23 Data/Parity + 1 Hot-spare
24 Data/Parity
23 Data/Parity + 1 Hot-spare
24 Data/Parity
47 Data/Parity + 1 Hot-spare
N/A
8
(f)
16
8+8
24
(c)
(a)(f)
(e)
7 + 17
48
(a)
(a)
(b)
(c)
(d)
(e)
(f)
(b)
24 and 48 drive arrays implement multiple RAID 6 sets within a single chassis.
48 drive arrays cannot implement a no hot-spare RAID policy.
PS60x0 XVS with 8x 100GB SSD and 8x 450GB SAS.
One 450GB hot spare, 2x 450 GB parity and 2x 100GB SSD parity. In the event of a drive failure (SSD or SAS), the
RAID set is reconstructed using the 450GB SAS hot spare.
One 600GB hot spare, 2x 600 GB parity and 2x 400 GB SSD parity. In the event of a drive failure (SSD or SAS), the
RAID set is reconstructed using the 600 GB SAS hot spare.
PS4100/PS6100 models only
Total reported usable storage when using hot spares: All models prior to PS4100/PS6100
Drive
Qty /
Size
(a)
50
(a)
100
74
146 250
300 400 450
500
600
750 1000
2000
3000
(b)
-
-
-
-
888
-
-
-
-
-
-
-
-
-
(c)
-
-
-
-
1111
-
-
-
-
-
-
-
-
-
-
-
660
-
2224
-
3566
-
4460
-
6695
-
-
-
577
1155
-
1696
2895
3475
4638
5220
5800
6964
8708
11613
23633
35717
-
-
-
-
-
-
-
2539
-
-
-
-
-
-
-
-
-
-
-
-
-
-
18062
21677
-
36136
63637
107151
7
8
14
(d)
16
8+8
(e)
48
Total reported usable storage when using hot spares: PS41x0/PS61x0
Drive
Qty /
Size
146
200
300
400
-
-
-
-
1331
1597
-
2672
5457
8253
12
1177
1658
2406
3328
4014
4812
7229
8038
16363
24719
14
1434
-
2938
-
4906
5890,
4286(h)
8841
9824
-
-
24
2478
3512(f)
5084
7034(f)
8478
10178
15278
16947
34550
52202
-
-
-
-
-
-
63637
-
6
500
7 + 17
-
-
-
-
-
48
-
-
-
-
-
(a)
(b)
(c)
(d)
Solid State Disk
Legacy PS50E
PS4000E and PS6000E only
Legacy PS70E, PS100E, PS100X, PS200E, PS300E
and PS400E
March 2013
600
900
(g)
9318
(e)
(f)
(g)
(h)
1000
2000
3000
PS60x0 XVS with 8x100GB SSD and 8x450GB SAS
PS6100S (SSD) only
PS6100XS only
M4110XS only
Dell EqualLogic Configuration Guide v14.1
4-31
4.2
RAID 10 drive layouts and total reported usable storage
Using a RAID 10 policy, Table 12 shows the drive layouts that are enforced based on the number of
drives in each array/hot spare configuration, and the total usable storage available for each model.
RAID 10 (mirrored sets in a striped set) combines two high performance RAID types: RAID 0 and RAID
1. A RAID 10 is created by first building a series of two disk RAID 1 mirrored sets, and then distributing
data over those mirrors. RAID 10 implementations can tolerate one drive failure per mirror pair.
Table 12 RAID 10 drive layouts and total reported usable storage available with hot spares (in GB)
Disk drives
Hot spare
No hot spare
6
4 Data (2 mirrors) + 2 Hot-spares
6 Data (3 mirrors) + 0 Hot-spare
7
6 Data (3 mirrors) + 1 Hot-spare
N/A
6 Data (3 mirrors) + 2 Hot-spares
8 data (4 mirrors)
12
10 Data (5 mirrors) + 2 Hot-spares
12 data (7 mirrors)
14
12 Data (6 mirrors) + 2 Hot-spares
14 data (7 mirrors)
14 Data (7 mirrors) + 2 Hot-spares
16 data (8 mirrors)
22 Data (11 mirrors) + 2 Hot-spares
24 data (12 mirrors)
46 Data (23 mirrors) + 2 Hot-spares
N/A
8
(c)
16
24
(c)
48
(a)
(b)
(c)
(a)
(b)
Not supported. An even number of drives is required by RAID 10
48 drive arrays cannot implement a no hot-spare policy.
PS4100/PS6100 models only
Total reported usable storage when using hot spares: All models prior to PS4100/PS6100
Disk
drive
qty /
Size
(b)
7
(a)
(a)
50
100
74
146
250
300
400
450
500
600
750
1000 2000 3000
-
-
-
-
666
-
-
-
-
-
-
-
-
-
-
-
-
-
666
-
-
-
-
-
-
-
-
-
-
-
395
-
1333
-
2135
-
2671
-
4013
-
-
-
16
309
621
-
913
1556
1867
2492
2805
3117
3744
4683
6247
12749
19230
48
-
-
-
-
-
-
-
-
10647
12779
-
21306
40840
63191
8
(c)
14
(d)
Total reported usable storage when using hot spares: PS41x0/PS61x0
Disk
drive
qty /
Size
146
6
-
-
-
12
650
920
14
781
24
1433
(a)
(b)
(c)
(d)
(e)
(e)
200
300
(e)
400
500
600
900
1000
2000
3000
-
887
1064
-
1781
3635
5498
1331
1843
2222
2672
4014
4454
9103
13762
-
1599
-
2670
3207
4816
5353
-
-
2027
2938
4065
4904
5888
8841
9820
19998
30221
Solid State Disk
Legacy PS50E
PS4000E and PS6000E only
Legacy PS70E,PS100E, PS100X, PS200E and PS300E
PS6100S (SSD) only
March 2013
Dell EqualLogic Configuration Guide v14.1
4-32
4.3
RAID 50 drive layouts and total reported usable storage
Table 13 shows the drive layouts that are enforced when using a RAID 50 policy based on the number
of drives in each array/hot spare configuration and the total usable storage available for each model.
RAID 50 (RAID 5 sets in a striped set) is created by first creating two or more RAID 5 sets and then
striping data over those RAID5 sets. RAID 50 implementations can tolerate a single drive failure per
RAID5 set.
Table 13 RAID 50 drive layouts and total reported usable storage available with hot spares (in GB)
Disk drives
Hot spare
No hot spare
6
5 Data/Parity + 1 Hot-spare
N/A
7
6 Data/Parity + 1 Hot-spare
N/A
6 Data/Parity + 2 Hot-spares
8 data/parity
12
10 Data/Parity + 2 Hot-spares
12 data/parity
14
12 Data/Parity + 2 Hot-spares
14 data/parity
8
(d)
16
24
(a)
(a)
14 Data/Parity + 2 Hot-spares
16 data/parity
(d)
22 Data/Parity + 2 Hot-spares
24 data/parity
(b)
46 Data/Parity + 2 Hot-spares
N/A
48
(a)
(b)
(c)
(d)
(c)
RAID 50 requires an even number of disk drives. A 7 drive configuration without hot-spare would result in odd
number of disk drives.
48 drive arrays implement stripes across multiple RAID 5 sets within a single chassis.
48 drive arrays cannot implement a no hot-spare policy.
PS4100/PS6100 models only
Total reported usable storage when using hot spares: All models prior to PS4100/PS6100
Disk
(a)
(a)
Drive
50
100 74
Qty / Size
146 250
300
400
450
500
600
750 1000 2000 3000
(b)
-
-
-
-
888
-
-
-
-
-
-
-
-
-
(c)
-
-
-
-
888
-
-
-
-
-
-
-
-
-
-
-
660
-
2224
-
3566
-
4460
-
6695
-
-
-
16
532
1066
-
1566
2671
3207
4280
4817
5353
6427
8037
10719
21819
32972
48
-
-
-
-
-
-
-
-
18062
21677
-
36136
69260
107151
7
8
14
(d)
Total reported usable storage when using hot spares: PS41x0/PS61x0
Disk
Drive
146
Qty / Size
6
(e)
200
(e)
300
400
500
600
900
1000
2000
3000
-
-
-
-
1781
2129
-
3563
7280
11008
12
1044
-
2129
2949
3563
4280
6430
7137
14571
21975
14
1304
-
2670
-
4459
5353
8036
8930
-
-
24
2355
3328
4815
6666
8038
9646
14474
16087
32727
49455
(a)
(b)
(c)
(d)
(e)
Solid State Disk
Legacy PS50E
PS4000E and PS6000E only
Legacy PS70E,PS100E, PS100X, PS200E and PS300E
PS6100S (SSD) only
March 2013
Dell EqualLogic Configuration Guide v14.1
4-33
4.4
RAID 5 drive layouts and total reported usable storage
RAID 5 (striped disks with distributed parity) will combine N disks in an arrangement where each stripe
consists of N–1 disks that contain data blocks plus 1 disk that contains a parity block. For each stripe,
the parity block will be placed on a different disk ensuring that the parity blocks are not located on a
single disk in the RAID set. RAID 5 implementations can tolerate a single drive failure without data loss.
Table 14 shows the drive layouts that are enforced when using a RAID 5 policy based on the number
of drives in each array/hot spare configuration, and the total usable storage available for each model.
Table 14 RAID 5 drive layouts and reported usable storage available with hot spares (in GB)
Disk drives
Hot spare
No hot spare
6
5 Data/Parity + 1 Hot Spare
7
6 Data/Parity + 1 Hot-spare
8
7 Data/Parity + 1 Hot-spare
(a)
12
11 Data/Parity + 1 Hot-spare
14
13 Data/Parity + 1 Hot-spare
16
24
RAID Policy for RAID 5 without a hot spare
is not supported
15 Data/Parity + 1 Hot-spare
(a)(b)
23 Data/Parity + 1 Hot-spare
(b)
46 Data/Parity + 2 Hot-spares
48
(a)
(b)
PS4100/PS6100 models only
24 and 48 drive arrays implement multiple RAID 5 sets within a chassis
Total reported usable storage when using hot spares: All models prior to PS4100/PS6100
Drive
(a)
(a)
qty / 50 100 74
Size
146
250
300
400
450 500
600
750
1000 2000 3000
(b)
-
-
-
-
1111
-
-
-
-
-
-
-
-
-
(c)
-
-
-
-
1333
-
-
-
-
-
-
-
-
-
-
-
792
-
2671
-
4281
-
5354
-
-
-
-
-
16
621
1244
-
1827
3118
3744
4995
5622
6247
7500
9378
12508
25456
38461
48
-
-
-
-
-
-
-
-
19452
23345
-
38916
74580
115394
7
8
14
(d)
Total reported usable storage when using hot spares: PS41x0/PS61x0
Drive
qty / 146
Size
6
(e)
200
300
(e)
400
500
600
900
1000
2000
3000
-
-
-
-
1781
2129
-
3563
7280
11008
12
1300
1843
2672
3696
4454
5355
8038
8929
18186
27473
14
1565
-
3206
-
5353
6426
9645
10719
-
-
24
2744
3880
5620
7772
9379
11253
16855
18728
38184
57698
(a)
(b)
(c)
(d)
(e)
Solid State Disk
Legacy PS50E
PS4000E and PS6000E only
Legacy PS70E,PS100E, PS100X, PS200E, and PS300E
PS6100S (SSD) only
March 2013
Dell EqualLogic Configuration Guide v14.1
4-34
4.5
Array RAID configurations and associated RAID sets
The tables show a logical drive layout when an array is initialized for the first time. The actual physical
layout of drives can change and evolve due to maintenance and administrative actions. Spare drives
can move as they are utilized to replace failed drives and newly added drives become the spares. It is
not possible to determine which physical drives are associated with each RAID set. This information is
dynamic and maintained by the EqualLogic firmware.
Table 15 shows the RAID set relationship for each RAID type in a 24-drive configuration.
Table 15 EqualLogic PS Series array RAID types and RAID set relationships
RAID policy Spare disks
RAID set relationship Best practice
RAID 6
1 Spare Disk
(10+2) (9+2)
RAID 10
2 Spare Disks (6+6) (5+5)
Yes
RAID 50
2 Spare Disks (5+1, 5+1) (4+1, 4+1)
For selected configurations
RAID 5
1 Spare Disk
Not for business critical data
(12+1) (9+1)
Yes
Table 16 shows the RAID set relationship for each RAID type in a 48-drive configuration.
Table 16 EqualLogic PS Series array RAID types and RAID set relationships
Raid
policy
Spare disks Raid set relationship
Best practice
RAID 6
2 Spare
(12+2, 12+2, 12+2) (2+2)
Yes
RAID 10
2 Spare
(7+7, 7+7, 7+7) (2+2)
Yes
RAID 50
2 Spare
(6+1 ,6+,1, 6+1, 6+1, 6+1, 6+1) (3+1)
For selected
RAID 5
2 Spare
(12+1, 12+1, 12+1) (6+1)
Not for business critical
Table 17 shows the RAID set relationship for each RAID type in a 16-drive configuration.
Table 17 EqualLogic PS Series array RAID types and RAID set relationships
March 2013
Raid policy Spare disks
Raid set relationship Best practice
RAID 6
1 Spare Disk
(13+2)
RAID 10
2 Spare Disks (7+7)
Yes
RAID 50
2 Spare Disks (6+1, 6+1 )
For selected configurations
RAID 5
1 Spare Disk
Not for business critical data
(14+1)
Yes
Dell EqualLogic Configuration Guide v14.1
4-35
5
PS Series array concepts
5.1
Groups and pools
A PS Series SAN Group is a Storage Area Network (SAN) comprised of one or more PS Series arrays
connected to an IP network. Each array in a group is called a group member. Each member is
assigned to a storage pool. There can be up to 4 pools within the group.
A group can consist of up to 16 arrays of any family or model as long as all arrays in the group are
running firmware with the same major and minor release number. For example, it is supported to have
different arrays in the same group running different maintenance release levels, as long as their major
and minor revision levels match. An example of this would be running version 4.0.1 and 4.0.3 on
different arrays in the same group. An example of an unsupported configuration would be different
arrays in the same group running 4.0.1 and 4.1.0 (the minor versions differ). The only exception to this
rule is for short term time periods when array firmware versions may be out of sync while upgrading
the firmware on each array within a group. Features available for use within the group are determined
by the lowest version of FW running within the group.
Note: We recommended that all arrays run the same version of PS Series firmware at all
times, except during a firmware upgrade process. (See Section 2.1 for more information.)
5.1.1
Pools
A pool is a container that each member is assigned after being added to the group. A pool can have
between one to eight members. There is always at least one pool in any group and it is called the
“default pool” unless the name is changed. Regardless of the name of this pool, it is always considered
the default storage pool. All newly added members are automatically assigned to the default pool. The
default pool cannot be deleted.
Pools can be used as part of an overall tiered storage strategy for the SAN. Tiered storage is storage
that is differentiated and optimized for the type of applications and data being hosted. Instead of
optimizing all of the storage in the group for maximum storage utilization, a tiered storage system
allows for the administrator to optimize different arrays for a variety of requirements such as
application performance requirements or cost efficiency.
Pools are the containers that provide scope for all automated storage optimization features of the
group. Pools with multiple arrays can implement different RAID policies within the pool. The
EqualLogic automated optimization features will automatically move a volume from an array or set of
arrays with one RAID policy to another array or set of arrays with a different RAID policy if it is
determined that the application needs more (or less) performance.
The following rules apply to pools:
•
•
•
•
Each member can be assigned to only one pool.
Each pool can host up to eight members.
Each group will have at least one pool – the default pool.
Each group can have up to four pools.
March 2013
Dell EqualLogic Configuration Guide v14.1
5-36
•
•
If all members in the pool are running PS Series firmware v5.0 or later then you can mix
PS5500E, PS6500E/X and PS6510E/X models together with other array models in the same
pool.
If you are running PS Series firmware version prior to v5.0 then PS5500E, PS6500E/X and
PS6510E/X models must reside in a separate pool from other array types.
Figure 3 shows a PS Series group with the maximum of four pools. Note the use of Pool 3 for
containing PS5500/PS6500 series arrays only. Also note that Pool 3 contains arrays that implement
both 1GbE and 10GbE controllers.
Figure 11 Tiered pools in a PS Series group
The following best practices should be considered for storage pools unless the automatic
performance load balancer (APLB) is disabled:
•
•
•
Mixing arrays of different drive speeds and types is fully supported and is the primary reason
APLB was created.
Do not mix arrays with different controller speeds (1GbE, 10GbE) within a single pool unless
they are each running unique RAID policies.
To override the automated performance decisions for a specific volume, indicate a “preferred”
RAID type for that volume. If that RAID type exists within the pool, the volume will reside on
those arrays that match the preferred RAID policy for the volume.
If the APLB is disabled, the following recommendations apply:
•
Do not mix arrays with different drive speeds within a single pool unless they are running a
unique RAID policy
March 2013
Dell EqualLogic Configuration Guide v14.1
5-37
•
•
Do not mix arrays with different drive technologies (SATA, SAS, SSD) within a single pool unless
they are running a unique RAID policy.
Do not mix arrays with different controller speeds (1GbE, 10GbE) within a single pool unless
they are each running unique RAID policies.
5.2
Volumes
Volumes provide the storage allocation structure within an EqualLogic SAN. Volumes are seen on the
network as iSCSI targets by hosts and are presented to the user as disk drives once the iSCSI initiator
has been configured and authenticated to the volume. Only computers with an iSCSI initiator and the
correct access credentials can access a volume. Disk space for volumes is allocated from the target
storage pool’s free space.
Volumes are identified within the Group Manager with a unique volume name. The volume name is
created by the administrator during volume creation and can be configured with several additional
characteristics and capabilities. The following sections provide additional details.
5.2.1
Volume attributes
Volumes are created using the “create volume” function in Group Manager or through the Command
Line Interface (CLI). Each volume must have a unique name that is used by Group Manager to identify
and manage the volume. Volume names may be changed via Group Manager at any time. Volume
names must meet the following requirements:
•
•
1 to 64 alpha-numeric characters
A-Z, a-z, 0-9, “.”, “-“, “:” are legal characters
Volumes are assigned a unique iSCSI Qualified Name (iqn) that consists of the following parts:
•
•
•
•
•
“iqn” followed by a “.” (period)
Year and Month of first full month that the naming authority was registered with standards
body. EqualLogic’s value is “2001-5” followed by a “.”
The storage provider’s domain name in reverse order. For example: “com.equallogic”
A colon (“:”)
Vendor specified information to identify an iSCSI target or initiator as unique within the vendor
domain. For EqualLogic iSCSI targets, this part consists of a unique set of numbers and the
user assigned volume name (within Group Manager).
The following is an example of an iSCSI target name for a volume named db3:
iqn.2001-05.com.equallogic:7-8b0900-6d0000000-001ebbc5d80sf0k0-db3
Volumes are visible to iSCSI initiators through one or more of three mechanisms during volume
creation:
•
•
•
iSCSI initiator IP address
A host’s iSCSI iqn well-known name
Mutual CHAP authenticated user identity
March 2013
Dell EqualLogic Configuration Guide v14.1
5-38
Note: IQN names are assigned to volumes automatically when they are created. They
cannot be changed for the life of the volume. If a volume name is changed, the IQN
name associated with the volume will remain unchanged.
5.2.2
Volume features
Each volume supports a set of features and capabilities that include the following:
•
•
•
•
•
Ability to define a volume as thin-provisioned
Support for snapshots
Support for replication
Support for creating clones
Support for multiple host shared access
Note: These block level volume features are not applicable to the component volumes
created by the NAS Service for the FS Series NAS Reserve.
5.3
Snapshots and clones
Snapshots are point in time copies of volumes. Snapshots have some features and properties similar to
volumes as well as some unique capabilities. Like volumes, snapshots can be assigned an iqn and
presented as volumes to a host. This allows hosts to mount a snapshot, potentially change the data in
the snapshot, or convert it to a real volume (clone) that has dedicated space allocated to it from the
free storage pool.
Snapshots require that space be reserved during volume creation (or after volume is created) to store
the snapshot(s) created from that volume. All snapshot reserve space for a specific volume always
resides in the same storage pool as the volume. By default, Group Manager allocates snapshot reserve
space equal to 100% of the host volume’s allocated space. This ensures that a 100% data change in the
volume can be protected by a single snapshot. This value can be set to a lower value based on the
application’s data change rate, snapshot/backup plan, or role that the snapshot will be used. The
following items should be considered when determine the size of a volume’s snapshot reserve:
•
•
•
Data change rate for the application(s) who is using the volume
Defragmentation policy of the OS or application using the volume
The role or purpose for creating the snapshot(s).
Snapshots have the following characteristics:
•
•
•
•
•
They are identified using names that consist of the parent volume name plus a date/time
stamp indicating when the snapshot was created as a default name.
If name of parent volume changes, existing snapshots iqn names do NOT change accordingly,
but retain their original iqn name.
Deleting a snapshot’s parent volume deletes all associated snapshots for that volume.
Snapshots of volumes with a high data change rate will require a larger snapshot reserve space.
Snapshots have access control lists that are inherited from the parent volume by default.
March 2013
Dell EqualLogic Configuration Guide v14.1
5-39
•
•
Snapshot reserve space for any volume can be decreased at any time. The minimum size
allowed is based on the current space used by existing snapshots using the snapshot reserve.
Snapshot reserved space for any volume can be increased at any time assuming there is
available free space in the storage pool hosting the volume.
5.3.1
Clones
Cloning creates a new volume by copying an existing volume. The new volume has the same
reported size, contents and thin-provision settings as the original volume. You can clone a regular
volume, a specific replica of a volume, or a specific snapshot of a volume.
•
•
•
•
Volume clones use non-shared storage (unlike snapshots)
Volume clones reduce the amount of storage pool free space
All regular volume operations can be performed on the clone once the cloning operation is
complete.
The cloning operation supports instant accessibility (while the clone operation is occurring)
Table 18 shows clone sources and properties.
Table 18 Cloning modes and properties
Clone source
Common properties
Clone from Volume
•
•
•
•
•
•
•
Clone from
Snapshot
Clone from Inbound
(a)
Replica
(secondary group)
(a)
5.4
Creates a new volume
Creates a new volume name
Creates a new iSCSI target IQN
Has same reported size as source volume
Has same contents as source volume
Cloned volume is immediately available
Cloning consumes free pool space equivalent to 100% of the
(a)
volume reserve setting for the source volume
Replica clones are created in the secondary group pool and are immediately available at the
secondary group IP address.
Thin provisioning
Thin provisioned volumes provide administrators with an additional option for managing data growth
within the SAN. Thin provisioning allows volumes to present a logical capacity to the host(s)
connecting to the volume that is different (larger) than the actual physical storage resource allocation
used by the volume at any given time.
Note: Volumes automatically created and managed by the NAS Service inside the NAS
Reserve pool used by an FS Series appliance are fully allocated at creation and cannot be
thin provisioned. Once those volumes are created (when the NAS Reserve is created they
cannot be shrunk.
A volume can be provisioned as a thin provisioned volume at creation or after creation. The following
rules apply to thin provisioned volumes:
March 2013
Dell EqualLogic Configuration Guide v14.1
5-40
•
•
•
•
•
•
A minimum physical allocation of 10% of the logical allocation is required
If a volume is converted to a thin provisioned volume, physical allocation cannot be less than
the amount of physical space already used within the volume
Any pool free space allocated to a thin provisioned volume is not returned to the free pool if
the host’s file system usage of that volume is reduced (due to file system defragmentation,
data removal, etc.)
If a thin provisioned volume’s allocated space exceeds the “maximum in-use space” setting,
the volume will go into an offline state.
Converting a volume to a thin provisioned volume may not reduce the physical allocation of
space the volume is using. Actual physical space recovered will depend on previous writing
patterns for the OS file system hosted on the volume.
All initial settings for “minimum volume reserve”, “In-use volume reserve warning level” and
“maximum in-use space” are default recommendations and can be changed by the
administrator within the constraints defined above.
The following best practices should be considered when using thin provisioned volumes:
•
•
•
•
•
•
•
•
Use Pool Free Space, not Group Free Space when making all determinations of thin
provisioned volume physical capacity allocation.
Create regular volumes before creating thin provisioned volumes. This provides the
administrator with a better view of the remaining available free space in the pool.
Set each thin provisioned volume’s “In use Volume Reserve Warning Level” to a level that
allows the administrator to ensure that additional physical storage can be deployed before the
volume uses the remainder of current pool free space.
Ensure that the Sum of all “In use Volume Reserve Warning Levels” for all thin provisioned
volumes does not exceed the current pool free space capacity minus 10%.
The storage administrator and application owner should reach an agreement on storage use
and procedures in determining the best automatic growth policy, reducing the frequency of
monitoring and intervention.
Be aware of the data usage patterns of the application. Some applications perform automated
disk optimization that may cause a thin provisioned volume to use more physical storage than
would be needed for normal operations. For these types of applications, thin provisioned
volumes may not be indicated.
Use “quick format” options when formatting OS file systems that are hosted by thin
provisioned volumes.
Thin provisioned volumes should not be used for hosting OS boot partitions or OS page file
cache.
5.4.1
Template volumes and thin clones
Template volumes and thin clones are new PS Series features introduced with firmware version 5.0.
Thin clones are created from template volumes. With a few exceptions, all normal volume operations
apply to template volumes and thin clones. For details on using template volumes and thin clones, see
the Advanced volume operations  About template volumes and thin clones section of the PS
Series Group Manager documentation.
March 2013
Dell EqualLogic Configuration Guide v14.1
5-41
6
Array firmware features
6.1
Replication
Replication is a powerful feature that can help you manage and implement a disaster recovery strategy
for your business applications and data. By replicating business-critical volumes, you ensure that your
business operations can resume quickly on a partner group in the event of a disaster on the primary
group. You also have the ability to restore the configuration to its original state if the problem on the
original group can be corrected. The replication process does not require placing volumes offline.
Thus you can take advantage of the benefits of replication without causing interruption in application
or data availability.
In EqualLogic PS Series storage, replication is the process of copying volume data from one PS Series
group to another PS Series group. Physical distance between the replication groups is not a concern
as long as they are connected by a reliable TCP/IP network path.
Table 19 provides definitions of EqualLogic PS Series terminology related to replication processing.
Table 19 Replication terminology
6.1.1
Term
Description
Replica
A time synchronized copy of an EqualLogic volume stored in a
Secondary Group.
Replica Set
A collection of all time synchronized replicas for a specific source
volume.
Primary Group
A group containing the source volume(s) to be copied or
replicated.
Source Group
Same as “Primary Group”.
Secondary Group
A group containing the replica or copy of the source volume(s).
Destination Group
Same as “Secondary Group”.
Delegated Space
The amount of space on the secondary group that is “delegated”
to a replication partner, to be reserved for retaining replicas.
Replica Reserve
The space allocated from delegated space in the secondary group
to store the volume replica set for a specific volume.
Local Reserve
The amount of space reserved on the local or Primary Group for
holding temporary snapshots and failback snapshots of the
source volume.
Replication limits
You need to consider the following limits when designing a volume replication topology:
•
A primary group can replicate to multiple partner (secondary) replica groups.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-42
•
•
•
•
•
PS Series groups can have up to 16 replication partners and can support a maximum of 10,000
total snapshots and replicas from all of its replication partners.
A group can have volumes replicating with multiple partners, but an individual volume can
have only one replication partner.
A maximum of 256 volumes per group can be configured for active replication.
All volumes that are part of a volume collection can only replicate with a single partner.
A volume can have up to a maximum of 512 replicas stored on a partner group.
o For a PS4000 only group, you cannot exceed two replication partners and 32 volumes
configured for replication
6.1.2
Replication paths
The example replication paths shown in Figure 12 are described below.
•
•
•
•
Basic partnership: one partner group hosts the primary copy of the volume and a second
partner group hosts the replica copy of the same volume. We also show the reverse direction
of the path for the Fast Failback replication feature, if it is enabled for the replica set.
Single to multiple group partnership: a single group can support replication of multiple
primary volumes to different secondary replication partner groups.
Reciprocal partnership: you can create bi-directional replication paths between partner
groups
Multiple to single group partnership: a common scenario in which multiple primary groups
replicate their volumes to a single secondary group partner in a different location.
Figure 12 Replication partnership paths
March 2013
Dell EqualLogic Configuration Guide v14.1
6-43
6.1.3
Replication process
When a replica is created the first replication process completes transfer of all volume data. For
subsequent replicas, only the data that changed between the start time of the previous replication
cycle and the start time of the new replication cycle is transferred to the secondary group. Dedicated
volume snapshots are created and deleted in the background as necessary to facilitate the replication
process.
A volume replica set is defined as follows:
Volume
Replica Set
A full copy of the primary volume,
with data synchronized to the beginning
of the most current completed replication
=
+
A time sequenced set of replicas, where
each replica corresponds to the state of the
volume at the beginning of a prior replication.
The number of prior replicas in the replica set that can stored on the secondary group is limited by the
size of the Replica Reserve allocated for that volume and the amount of data that changes.
Replication processing occurs in a series of phases. The flowchart in Figure 13 shows the process
phases, focusing on how the process tracks and copies changes that occur between each replica
cycle.
Referring to the phases shown in Figure 13:
Replication Setup (one-time):
Configure replication partnership and volume replication settings.
Replication Processing (repeating):
The primary group checks for availability of sufficient delegated and replica reserve space on the
secondary group at the beginning of each replication processing phase. If adequate space is not
available then the process will pause and generate an event message. Replication will continue
once sufficient space is made available. This part of the process is not shown in the chart.
Primary to secondary volume data replication is completed. The process steps vary based on
replication status (first or subsequent) and fast failback mode (enabled or disabled). During this
process Local Reserve is consumed by a hidden snapshot (and the fast failback snapshot if
enabled). Volume data changes that occur during the replication processing phase are stored by
the hidden snapshot in Local Reserve. Replica Reserve allocated to the volume within delegated
space on the secondary group receives all volume data changes. Replica Reserve is consumed by
the most recent complete volume replica plus all prior replicas stored in the replica set.
Between Replication Events (repeating):
Once first replication has occurred the system continues to keep track of volume data changes
that occur so that subsequent replication processes can copy those changes to the replica set.
This tracking process does not consume additional space.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-44
Figure 13 Replication Process
March 2013
Dell EqualLogic Configuration Guide v14.1
6-45
With replication, it does not matter if the volume is thin provisioned or uses a traditional volume since
in either case only the data that has changed will be copied to the replica. On the secondary side,
volumes are always thin provisioned to conserve available capacity used by the Replica Reserve for
that volume.
6.1.4
Fast failback
With fast failback enabled you can ensure that the volume data preserved by the failback snapshot on
the primary group always matches the volume data in the most recent replica stored on the secondary
group. If you have an event that causes failover to the secondary group and the workload
subsequently writes changes to the replica volume then failback snapshot supports a quicker fail back
to the primary group by replicating only the changes made to the replica volume during the time it
was active as a recovery volume on the secondary group. If the failback snapshot is not enabled, you
must replicate the entire volume contents back to the primary group to complete the failback
operation. Depending on the size of the volume, the failback scenario can take significantly longer to
complete if Fast Failback is not enabled.
6.1.5
Sizing Replica Reserve and Delegated Space
The volume replication process in EqualLogic PS Series storage consumes extra storage space on both
the primary and secondary group partners. In addition to the space consumed by the volume itself,
each volume will require additional space in the primary group for Replication Reserve and Local
Reserve, plus delegated space for storing replica sets in the secondary group. This is illustrated in
Figure 14 below. A single delegated space on the secondary group must be used for all volumes
received from a specific primary group.
Delegated space for a replication partnership must be assigned to a pool. Delegated space for
different replication partnerships can be assigned to different pools on the secondary group.
Note: Any type of application or operating system level process that causes block level
data changes will affect replication data size and time to complete. For example, file
system defragmentation of a large partition will significantly increase the amount of
space consumed in replica reserve, delegated space on the secondary group and time to
complete the next scheduled replication.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-46
Figure 14 Replica Reserve, Local Reserve, and Delegated Space
Auto replication requires reserved disk space on both the primary and secondary groups. The amount
of space required depends on several factors:
•
•
•
•
Volume size.
The amount of data that changes (on the source volume) between each replication period.
The number of replicas that need to be retained on the secondary site.
If a failback snapshot is retained on the primary group.
The default values that appear in Group Manager are sufficient to ensure that enough space is
reserved for at least one successful replication, even if the entire contents of a volume are altered
between replicas. Initial replication of a volume will cause the complete contents of the volume to be
copied to the secondary group. Subsequent replication events will copy only the changed data. The
recommended and space efficient guidelines for sizing replication reserves and delegated space are
presented in Table 20.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-47
Table 20 Replication space sizing guidelines
Recommended
Value
Replication Space
Local Reserve
(Located on Primary Group.)
Space Efficient
Value
No Failback
Snapshot:
100%
5% + CR
Keep
Failback
Snapshot:
200%
10% + CR
(a)
(a)
200%
Replica Reserve
(b)
Ensures there is adequate
space for the last replica and
any replica in progress.
(Located on Secondary Group.)
Delegated Space
(Located on Secondary Group. The
replica reserve space for all replica
sets coming from a single group)
(a)
(b)
Must be large enough to hold
the sum of all replica reserve
sizes for all volumes
replicating to that group.
105%
+ CR x (# of Replicas – 1)
Monitor change rate, adjust to
lower than recommended
value and continue
monitoring.
CR: “Change Rate”. For details on how volume changes affect replication space, see the Centralized Replication
section under Replication Configuration Options in the Group Manager Help documentation.
Start with 105%, then add to that the maximum number of replicas expected to be stored in the replica set minus 1,
multiplied by the expected Change Rate.
6.1.6
Effect of TCP Window Size on Latency across WAN links
In certain cases, increasing the TCP Window Size setting on PS Series controllers can improve
replication throughput across WAN links. The ability to change this setting is currently not supported
using the EqualLogic Group Manager. Customers interested in changing TCP Window size settings
should contact Dell EqualLogic Technical Support for assistance. For more details on how this setting
can affect replication performance across WAN links, see the following publication:
•
Dell EqualLogic Auto-Replication: Best Practices and Sizing Guide :
http://en.community.dell.com/dell-groups/dtcmedia/m/mediagallery/19854181aspx
6.1.7
Replication partner compatibility
Table 14 provides details on replication partner interoperability based on array firmware version. For
the latest information, consult the release notes for your array firmware version.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-48
Table 21 Firmware replication compatibility
6.1.8
Firmware of “Local” Group
Firmware on Replication Partner
V6.0
V5.1.x, V5.2.x, V6.0.x
V5.1.x, V5.2.x
V5.0.x, V5.1.x, V5.2.x
V5.0.x
V4.2.x, V4.3.x, V5.0.x, V5.1.x, V5.2.x
V4.2.x, V4.3.x
V4.1.x, V4.2.x, V4.3.x, V5.0.x
V4.1.x
V4.0.x, V4.1.x, V4.2.x, 4.3.x
V4.0.x
V3.2.x, V3.3.x, V4.0.x, V4.1.x, V4.2.x
V3.2.x, V3.3.x
V3.0.x, V3.1.x, V3.2.x, V3.3.x, V4.0.x
V3.0.x, V3.1.x
V3.0.x, V3.1.x, V3.2.x, V3.3.x
Clustering
To support a shared storage environment, EqualLogic allows concurrent access to any volume.
Concurrent volume access is enabled on a per volume basis within the Group Manager or via CLI. The
EqualLogic array will not manage concurrent access to these volumes. Control of access to a shared
volume must be provided by means of access control software on each host that has access to the
volume.
Warning: Failure to provide access control to a shared volume can result in data loss.
By default, EqualLogic PS Series groups disable multi-host (shared) access to target.
•
•
If needed, you can enable multi-host access for a volume.
If you enable multi-host access to a volume, then proper I/O operation concurrency must be
maintained.
6.2
Synchronous replication
6.2.1
About Synchronous replication
Synchronous replication (SyncRep) is the simultaneous writing of volume data across two different
storage pools in the same PS Series group, resulting in two hardware-independent copies of volume
data. Each write must go to both pools before the write is acknowledged as complete. If one pool is
not available due to a power failure or resource outage, you still can obtain volume data from the
other pool.
You cannot perform traditional replication on a SyncRep-enabled volume.
6.2.2
How Synchronous replication works
SyncRep is enabled on a per-volume basis. In volumes for which SyncRep is not enabled, volume data
and snapshots are located only in the pool to which the volume is assigned.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-49
In SyncRep-enabled volumes, volume data exists simultaneously in two pools:
• SyncActive: The pool to which iSCSI initiators connect when reading and writing volume data.
• SyncAlternate: When volume data is written to the SyncActive pool, the group simultaneously
writes the same data to this pool. You can switch the roles of the SyncActive and
SyncAlternate pools.
When you switch the SyncActive and SyncAlternate pools, the former SyncActive pool then becomes
the SyncAlternate pool, and vice-versa. No iSCSI target configuration changes are required. During the
switch, host connections are logged out. iSCSI initiators can reconnect when the switch has
completed. Depending on its configuration, the initiator may automatically reconnect. If you delete a
volume for which SyncRep is enabled, the system will place the SyncActive volume into the Recovery
Bin. However, the SyncAlternate volume will be deleted outright and cannot be recovered.
6.2.3
Synchronous states
There are three SyncRep states for a volume:
• In-Sync —SyncActive and SyncAlternate pools contain the same volume data.
• Paused—Administrator has paused SyncRep. While SyncRep is paused, the volume is still online,
and initiators can connect to and write to the SyncActive volume. An administrator may pause and
later resume SyncRep. For example, this could happen in a maintenance window during which the
SyncAlternate pool is taken offline. If data is written to the volume while SyncRep is paused, it is
written only to the SyncActive pool, and the two pools are out of sync. The group tracks all
volume writes while SyncRep is paused and, when the administrator resumes SyncRep, writes the
tracked changes to the SyncAlternate pool.
• Out of Sync—SyncActive pool and SyncAlternate pool may not contain the same volume data; the
SyncActive pool contains the most recent volume data. A volume can become out of sync if
SyncRep is paused, or if one of the pools becomes unavailable or has no free space. The volume
can become out of sync when the snapshot reserve in the SyncAlternate pool is full, but only
when the snapshot space recovery policy sets volumes offline when the snapshot reserve is
depleted.
Whenever the volume's state changes to paused or out of sync, the group creates a snapshot of the
volume that reflects the volume's contents at the point in time when the state changed. This snapshot
resides in the SyncActive pool. If there is insufficient room in the snapshot reserve for the snapshot,
the group does not create
6.2.4
Caveats:
Depending on the quantity of tracked changes, activity within the group, and available network
bandwidth, there may be an extended period of time before the two pools become in sync again. The
Group Manager GUI displays the status of this operation. When SyncRep is first enabled, or at any
other time when the volume is writing data to both pools to become in sync, performance
degradation may occur. This effect increases with the quantity of tracked changes, but it is significantly
reduced after the volume becomes in sync.
6.2.5
Requirements for using SyncRep
Before you can configure a volume to use SyncRep, verify that the following requirements are met:
• Two pools, each containing at least one member.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-50
•
•
•
•
Adequate network bandwidth between pools.
Free space in each pool to accommodate the volume and snapshot reserve for the volume.
You cannot enable SyncRep on a volume for which traditional replication is configured, and
you cannot enable traditional replication on a volume for which SyncRep is configured. See
the "Disabling Replication" topic in the online help for instructions on disabling traditional
replication on a volume.
You cannot enable SyncRep on a volume that is bound to a group member.
6.2.6
How SyncRep protects volume availability
In normal SyncRep operation, in which the volume is in sync, the pools containing the SyncActive and
SyncAlternate volumes contain identical data. Volume writes are accepted as shown in Fig 15.
Figure 15 SyncRep
1.
2.
3.
4.
The iSCSI initiator sends a write to the group.
Data is simultaneously written to both the SyncActive and SyncAlternate volumes.
The SyncActive and SyncAlternate volumes confirm to the group that the writes are complete.
The write is confirmed to the iSCSI initiator.
6.2.7
SyncAlternate volume unavailable
If the group can write to the SyncActive volume, but can no longer write to the SyncAlternate, initiator
access to the volume continues without disruption, as shown in Figure 16.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-51
1.
2.
3.
4.
The SyncAlternate volume becomes unavailable.
Initiator access to the volume continues without interruption; the volume is out of sync.
The group tracks all changes written to the volume.
When the SyncAlternate volume becomes available, tracked changes are written to the
SyncAlternate volume.
Figure 16 SyncAlternate volume unavailable
Performance will be temporarily degraded while changes are being tracked or when tracked changes
are being written back to the SyncAlternate volume.
6.2.8
Tracked changes written to the SyncAlternate volume
1. The SyncAlternate volume becomes available.
2. Changes tracked while the SyncAlternate volume was unavailable are written to it. Until all
tracked changes are written, the data in the SyncAlternate volume is valid up to the point in
time when the volume went out of sync. During this time, performance may be temporarily
degraded. When all tracked changes are written, the volume goes back in sync.
3. New writes are simultaneously written to both the SyncActive and SyncAlternate, and normal
SyncRep operations resume as shown in Figure 17.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-52
Figure 17 Tracked changes written to the SyncAlternate volume
6.2.9
SyncActive volume unavailable
If a malfunction occurs in the SyncActive pool, or some other event has occurred causing the volume
to go offline, you can safely switch or failover to the SyncAlternate volume by following one of the
procedures listed below.
• Volume In Sync: If the volume is in sync, you may switch to the SyncAlternate as documented in
the online help. Although host access to the volume is disrupted during the switch, no initiator
changes are required.
• Volume Out of Sync: If the volume is out of sync, the group administrator performs the steps
outlined in the online help to safely restore access to the volume:
Table 22 below gives a comparison between traditional replication and SyncRep.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-53
Table 22 Comparing SyncRep and traditional replication
Replication
consideration
Traditional replication
SyncRep
Typical Use
Case
A point-in-time process that is conducted
between two Groups, often in
geographically diverse locations.
Replication provides protection against a
regional disaster such as an earthquake or
hurricane.
Traditional replication has the advantage of
providing multiple recovery points.
A disadvantage of traditional replication is
that the state of the data between recovery
points is unknown; if any changes are
made to the volume since the last replica
was created, they could be lost.
A real-time process that keeps two
identical copies of volume data in
two different pools within the same
PS Series group.
SyncRep is useful for maintaining two
copies of a volume's data in the same
data center, or dispersed to two
different facilities on the same
campus or in the same metropolitan
area.
An advantage of SyncRep is that it
captures a duplicate copy of every
write. One disadvantage of this is that
if an application writes bad data to
the volume, the bad data is
simultaneously written to both the
SyncActive and SyncAlternate pools.
Recovery Time
If a disaster occurs in the primary group,
you can promote the replica set on the
secondary group to a recovery volume.
After the promotion, you must reconfigure
initiators to discover and log in to the iSCSI
target now hosted by the secondary group,
or switch to an alternate set of server
resources that have been preconfigured to
use the secondary group storage. See the
Impact on Applications row of this table for
more information.
If a disaster involving the SyncActive
pool occurs, you can manually
switch the volume to the
SyncAlternate pool.
After the switch, the SyncAlternate
pool becomes the SyncActive pool
and hosts the volume.
Host access to the volume is
disrupted by the switch, but iSCSI
initiators do not need to be
reconfigured.
Recovery Point
The recovery volume contains point-intime data that is current as of the most
recent replica. Replication can be
scheduled to take place as frequently as
once every five minutes.
You can also restore to the point in time
when any previous replicas were created,
provided that the replicas have been
retained.
SyncRep provides a single recovery
point: the most recent
acknowledged write to the volume.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-54
Replication
consideration
Traditional replication
SyncRep
Network
Requirements
Replication requires that the network
connection between the primary and
secondary groups must be able to handle
the load of the data transfer and complete
the replication in a timely manner.
Because writes are not
acknowledged until they are written
to both the SyncActive and
SyncAlternate pools, SyncRep is
sensitive to network latency.
The network must be able to handle
the load of the data transfer from the
SyncActive pool to the SyncAlternate
pool and complete the replication in
a timely manner, or application
performance may suffer.
Snapshots
Replication is functionally similar to
SyncRep creates snapshots of the
snapshots, creating point-in-time copies of volume whenever the SyncActive and
the volume. If the "keep failback" option is
SyncAlternate pools are switched.
enabled, the group creates a "failback
snapshot" on the primary group every time
a replica is created. This allows for "fast
failback" capabilities.
In addition, you can schedule the creation
of snapshots, or create them on demand,
just as you would with any other volume.
See About SyncRep and Snapshots on page
126 for more information.
Scheduling
Replication operations can be scheduled
using the same mechanism used for
scheduling snapshots.
Replication between the SyncActive
and SyncAlternate pools is
continuous. Therefore there is no
need or mechanism to schedule
SyncRep.
Pool Space
Requirements
The primary group must have enough
space for the volume reserve and local
replication reserve, in addition to any
snapshot reserve.
The secondary group must have enough
free space delegated to the primary group
for the volume reserve and the replicas that
record changes to the volume's data over
time.
Both the SyncActive pool and the
SyncAlternate pool must have
enough space for the volume and
snapshot reserve.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-55
Replication
consideration
Traditional replication
SyncRep
Impact on
Applications
iSCSI initiators must be reconfigured to
connect to the secondary group after the
failover, or an alternate set of host
resources must be brought online, both of
which may cause application disruptions.
If you are using the Host Integration Tools,
you can coordinate replication with host
software to quiesce applications on a
schedule and create application consistent
Smart Copies.
Replication can help protect against the
corruption of application data: depending
on when the replica occurred and what
your replica retention policies are, you may
be able to restore the volume to a point in
time before the corruption occurred.
Pool switches may cause disruptions
in host access to the volume, but no
change to the iSCSI initiator
configuration is required to restore
access.
Writes must be committed to both
pools before they are acknowledged
to the host, so the application must
be able to tolerate whatever
additional delay is caused by the
simultaneous writes.
When SyncRep is first enabled, or at
any other time when the volume is
writing data to both pools to become
in sync, performance degradation
may occur. This effect is diminished
after the volume becomes in sync.
PS Series Group
Requirements
Two PS Series groups, each of which must
contain at least one member.
One PS Series group containing two
storage pools, each of which must
contain at least one member.
6.3
Protecting your EqualLogic group with Internet Protocol
Security
Internet Protocol Security (IPsec) is a set of standardized protocols designed to allow systems on IPbased networks to verify each other's identities and create secured communication links. IPsec uses
cryptographic security mechanisms for authentication and protection. IPsec validates the identity of
devices communicating over IP-based networks, encrypts all data passing between participating
systems, and protects against disclosure, modification, eavesdropping, and attack. IPsec is supported
for both IPv4 and IPv6 networks.
In the context of an iSCSI SAN that uses EqualLogic PS Series storage arrays, IPsec secures
communications between group member arrays and also between iSCSI initiators and the group. You
can use policies to configure your IPsec implementation to protect iSCSI traffic based on initiator IP
address, initiators in a specific subnet, or network protocol. IPsec authentication is handled using
certificates or pre-shared keys.
IPsec is supported only for PS Series array models PS6xxx, PS41x0, and PS-M4110, and can only be
enabled for a group if all members support IPsec. See the Dell EqualLogic PS Series Storage Arrays
Release Notes for more information.
6.3.1
Types of protected traffic
The types of traffic protected by IPsec are shown in Figure 18 and the sections that follow. Any
incoming or outgoing IP traffic that travels between hosts and the group can be protected with IPsec.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-56
Figure 18 IPsec protected traffic
6.3.2
Protected Intra-Group Traffic
Once IPsec is enabled, all network traffic between group members is automatically protected with
IPsec using IKEv2. No further configuration is required.
6.3.3
IPsec and Replication
The PS Series Firmware provides no mechanism for using IPsec to protect traffic between replication
partners. It is technically possible to create IPsec polices on both the primary and secondary group in
which each group treats the other as an iSCSI initiator and traffic is protected accordingly. However,
this is an unsupported configuration, and Dell recommends against implementing it in a production
environment.
6.3.4
About IPsec Security Parameters
IPsec security parameters control the authentication and key negotiation carried out using the Internet
Key Exchange IKEv1 or IKEv2 protocol.
Security parameters specify the following:
Using IKEv1, IKEv2, or manual keying.
You can configure IPsec to use manual keys. However, manual keying provides significantly weaker
security than IKEv1 or IKEv2, and is also significantly more difficult to configure. Consequently, Dell
strongly discourages the use of manual keying in any production environment. IKEv1 or IKEv2 are the
preferred keying methods. Refer to the Dell EqualLogic Group Manager CLI Reference Guide for more
information about using manual keys.
6.3.5
About IPsec Certificates
Certificates are used in an IPsec configuration as one method of authenticating secured connections
between iSCSI initiators and the group. Implementation of an IPsec-enabled SAN requires both a rootCA (Certificate Authority) certificate from the issuing authority and a local certificate to authenticate
the group.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-57
You can generate certificates suitable for use in IPsec connections to the PS Series using any
Windows, OpenSSL, or other commercial Certificate Authority product.
You can use the Group Manager CLI to import, display, and delete certificates. using the IPsec
certificate commands. See the Dell EqualLogic Group Manager CLI Reference Guide for more
information.
6.3.6
About IPsec Pre-Shared Keys
In addition to using certificates, you can use pre-shared keys to authenticate secured connections.
Pre-shared keys are identical strings that are specified at both ends of the communications pathway.
This allows the systems to correctly identify each other.
You can use either ASCII or hexadecimal strings. ASCII can be used in most situations. However, you
can also use hexadecimal strings if your organization mandates their use, if you have systems that do
not support the use of ASCII strings, or if you want to use unsupported characters.
6.3.7
About IPsec policies
Traffic that meets the conditions stipulated by the policy can either be passed, dropped, or protected
using an IPsec security parameter associated with the policy.
You can use IPsec policies to apply IPsec protection to traffic that meets one or more of the following
criteria:
• Data traveling to or from specific IP addresses, or a range of IP addresses defined by a specific
subnet or netmask
• IPv4 or IPv6 traffic
• Specific network protocols: TCP, UDP, or ICMP (either IPv4 or IPv6)
Unless explicitly specified by the policy, traffic is allowed to pass. If you want to drop all traffic that is
not explicitly protected or passed, you must create an IPsec policy that drops traffic by default.
If there are multiple IPsec policies in place, the system determines their priority by the order in which
they were created; policies created first take precedence over policies created later.
You can also use IPsec policies to determine what traffic is being protected using IPsec, and what
traffic is being passed or dropped without encryption.
IPsec policies are managed using the IPsec policy commands. See the Dell EqualLogic Group Manager
CLI Reference Guide for more information.
6.3.8
IPsec considerations and limitations
The limitations listed in the sections below apply when implementing IPsec.
Configuration limitations
• IPsec is only supported for certain PS Series array models, and can only be enabled for a group if
all members support IPsec. See the Dell EqualLogic PS Series Storage Arrays Release Notes for
more information.
• IPsec can only be enabled and configured using the Group Manager CLI. The Group Manager GUI
provides no facility for configuring or monitoring IPsec.
• The PS Series array does not serve as an IPsec-secured gateway; it only behaves as an IPsecsecured host.
• You cannot use the save-config CLI command to preserve the group's IPsec certificates and preshared keys. The save-config command saves the CLI commands that were used to configure
IPsec, but it does not save certificates that have been transferred to the array using FTP. Therefore,
when you restore a configuration, you must manually restore any configuration options set using
the IPsec certificate load, IPsec security-params create certificate, and IPsec security-params preshared-key commands.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-58
•
•
•
•
•
•
•
•
•
•
Kerberos-based authentication is not supported.
Multiple Root Certificate Authorities (CA) are not supported.
Certificate Revocation Lists (CRL) are not supported.
Only users with group administrator privileges can configure IPsec.
Perfect Forward Secrecy (PFS) is not supported.
Encrypted private keys are not supported for X.509 format certificates.
Dell recommends using a minimum of 3600 seconds and 10GB lifetime rekey values.
IKE mobillity is not supported
NAT Traversal (NAT-T) is not supported. Dell recommends against placing a firewall that performs
address translation between the PS Series group and its IPsec peers.
If you use the Windows default IPsec lifetime rekey values, the high rekey rates may be disruptive
for protected iSCSI traffic. Values in the range of 1GB to 100GB, depending on iSCSI traffic, are
recommended instead.
6.3.9
Performance considerations
The performance impact of IPsec varies by host and network configuration, and increases with the
number of IPsec-protected iSCSI connections to the group. Even if IPsec is only used to protect traffic
between group members, I/O performance is still affected. Based on these factors, you can expect
that using IPsec may degrade I/O performance.
Although PS Series group members use hardware to accelerate cryptographic operations, many
initiators perform these operations in software, which can cause a further reduction in the speed of
communications between iSCSI initiators and the group.
6.3.10
Host Connectivity Considerations
Enabling or disabling IPsec for the group using the IPsec enable and IPsec disable commands might
disrupt host connectivity to the group for several minutes. To prevent unplanned outages, IPsec
should be enabled or disabled during a planned maintenance window when there are no active iSCSI
connections to any volumes.
Consult the documentation for your host operating systems, HBAs, and iSCSI initiators to verify that
they support IPsec. There might also be known issues and idiosyncrasies with the initiators' IPsec
support that require additional planning or configuration.
When configuring IPsec with Windows hosts, note the following:
IPsec traffic is not always handled correctly if the IPsec policy is configured to protect only a subset of
traffic between the host and the group. For example, if the IPsec policy protects only iSCSI traffic on
port 3260, the Windows host may not perform reliably when connecting to the group. As a
workaround, IPsec policies should apply to all traffic passing between the group and Windows
systems. Microsoft KB article 2665206 discusses this in greater detail.
IPsec must be configured using the Windows Firewall with Advanced Security. Do not use the IPsec
option in the Microsoft iSCSI initiator, which does not have the capability to fully configure an IPsec
configuration between the host and the group. Further, if you attempt to configure an IPsec
connection using the iSCSI initiator, the system might not allow you to remove the partial
configuration and replace it with a complete configuration created with Windows Firewall.
IPsec policies defined using the Local Security Policy Manager are not supported.
March 2013
Dell EqualLogic Configuration Guide v14.1
6-59
7
EqualLogic SAN design
An EqualLogic iSCSI SAN can be operated in any network that supports the industry standards and IP
subnet design guidelines described in this section. Because of this flexibility, there are many network
design and configuration choices that can affect SAN performance. The following sections provide
details related to network design and configuration to support the use of an EqualLogic SAN.
Note: With the addition of FS Series NAS functionality, the EqualLogic product family now
provides an iSCSI based unified file and block storage platform. An EqualLogic SAN can
now be used to provide block level access (direct iSCSI access to PS Series arrays) or file
system level access (via the FS Series appliance) using NFS or CIFS protocols and the Dell
FluidFS scale-out file system.
Unless otherwise stated, recommendations in this document are applicable to both file and block
environments.
7.1
General requirements
7.1.1
Implementation of standards
EqualLogic SANs are based on industry standards. The following standards are required to support all
host to target communications and member to member communications in an EqualLogic SAN:
•
•
IETF Standards
o IETF RFC1122 “Requirements for Internet Hosts – Communications Layers”
o IETF RFC1123 “Requirements for Internet Hosts – Application and Support”
o IETF RFC3270 “Internet Small Computer Systems Interface (iSCSI)”
IEEE Standards
o 802.1
o 802.3
iSNS support
1
An Internet Storage Name Service (iSNS) server can support discovery, management and
configuration of group members by dynamically updating information about the iSCSI target names
for group volumes. Once the IP address of an iSNS server is entered in an iSCSI initiator’s configuration
utility, the setting is persistent across initiator sessions. A PS Series group can be configured to register
with up to three iSNS servers.
Note: Starting with Firmware V4.1.4, volume and snapshot identifiers are no longer
automatically published to iSNS servers. This applies to new volumes and snapshots as
well as volumes and snapshots that existed before the group was upgraded to V4.1.4.
1
The Internet Storage Name Service (iSNS) specification: http://tools.ietf.org/html/rfc4171
March 2013
Dell EqualLogic Configuration Guide v14.1
7-60
7.1.2
General requirements and recommendations
For EqualLogic PS Arrays, the following general SAN design requirements apply:
•
For all members (arrays) in a given SAN Group all ports should be connected to the same subnet.
This allows the arrays to communicate with each other as a group of peer members. The arrays
must be in the same subnet as the group’s “well known” IP address.
Note: Hosts can be in a different subnet as long as those hosts have layer 3 routing
available to the subnet containing the arrays and the group’s well known IP address.
•
It is strongly recommended that a physically separated network be used for iSCSI traffic and that
this network not be shared with other traffic types.
Note: If there is a requirement to share the same physical networking infrastructure with
other non-iSCSI traffic then Data Center Bridging (DCB) is the recommended method for
sharing networking resources.
•
•
Rapid Spanning Tree Protocol must be enabled if the SAN infrastructure has more than two
switches in a non-stacked configuration, and portfast must be enabled on all edge device ports
(hosts, FS Series appliances and arrays).
Port density requirements to support fully redundant configurations and maximum SAN
throughput are as follows:
•
•
•
•
•
•
•
•
•
•
•
•
•
PS4x00 family: 2x 1GbE ports per controller = 4x 1GbE ports total
PS4110 family: 1x 10GbE port per controller = 2x 10GbE ports total
PS5x00 family: 3x 1GbE ports per controller = 6x 1GbE ports total
PS6x00 family: 4x 1GbE ports per controller = 8x 1GbE ports total
PS6510 family: 2x 10GbE ports per controller = 4x 10GbE ports total
PS6110 family: 1x 10GbE port per controller = 2x 10GbE ports total
FS7500 NAS: 12x1GbE ports per controller node (four client LAN and eight iSCSI SAN) + 1
100/1000Mb port per controller node for IPMI interconnection
FS7600 NAS Applicance: 16x1GbE Ethernet ports per appliance = 8 x 1Gb Ethernet ports per
NAS controller for client connectivity and 8 x 1GbE Ethernet ports per NAS controller for
SAN connectivity
FS7610 NAS Appliance: 8x10GbE Ethernet ports per appliance = 4 x 10GbE SFP+ Ethernet
ports per NAS controller for client connectivity and 4 x 10GbE SFP+ Ethernet ports per NAS
controller for SAN connectivity
At least two iSCSI SAN ports per host (block level iSCSI access) are required for fully redundant
SAN connectivity. Host ports can be 1GbE or 10GbE.
Quality of Service (QoS) based on what is traditionally designated as IEEE 802.1p is not currently
supported for use with EqualLogic SANs. QoS and Class of Service designations must be disabled.
All switches within the SAN must be interconnected such that there is always a path from any
Ethernet port on one array to all other Ethernet ports on all other arrays in the group.
All switches and host network controllers within the infrastructure must have flow control enabled
for optimal performance.
March 2013
Dell EqualLogic Configuration Guide v14.1
7-61
•
•
•
•
•
•
Any EqualLogic SAN group that is required to send or receive replication traffic to/from another
SAN group must have an uninterrupted communications path (ie. “visibility”) between each group.
To prevent a switch failure from also disabling all paths between a host and its connected
volumes, all ports from each controller need to be connected to at least two different switches
For PS4100/PS6100 family arrays, split the vertical port pair connections between two switches to
ensure 100% bandwidth capability is maintained in the event of a vertical port failover event. See
the configuration diagrams in Section 7.3.6.
The above guideline regarding path redundancy across multiple switches also applies in FS Series
NAS appliance configurations. (Refer to the connection diagrams in Section 13.1)
Do not block IPv6 traffic on the SAN internal switches when utilizing FS Applicances
FS76x0 uses IPv6 for internal communication and setup only; not for user data.
Client communication happens over IPv4
For SANs connecting to an FS Series appliance, all switches in the SAN must have jumbo frames
enabled.
For EqualLogic PS Series Arrays, the following general SAN design recommendations apply:
•
•
•
Take advantage of your switch’s VLAN capabilities. You should create a VLAN dedicated to iSCSI
traffic (even on dedicated switches). If necessary, create a second VLAN for management traffic.
The actual VLAN configuration of your iSCSI SAN will be dictated by your SAN network design
requirements and the features of the iSCSI SAN switches being used.
Jumbo frames should be enabled. If you choose to use jumbo frames then all nodes in the SAN
fabric must have jumbo frames enabled.
For best performance and reliability, we recommend that all interconnection paths between nonstacking switches (LAGs) use a dynamic link aggregation protocol such as LACP
7.1.3
Quality of service (qos)
Quality of service is described as either of the following:
•
•
The ability to provide different priority levels to different applications, users, or data flows, or to
guarantee a certain level of performance to a data flow.
A network function implemented in some routers and switches that provides high priority for
certain types of latency-sensitive traffic (for example, VoIP) and lower priority for other types of
traffic (for example, web or http).
PS Series arrays are designed to provide I/O as fast as your network can support it. Therefore, using
QoS with iSCSI traffic does not produce expected or desirable results on the SAN. Also, QoS rules can
affect how well–or even whether–replication succeeds between PS Series groups. If you plan to use
QoS, Dell recommends that you use it only on VLANs that do not carry iSCSI traffic, or on WANs,
where bandwidth is shared with other applications and the PS Series array uses it for time-insensitive
replication traffic.
Dell recommends against using QoS on the SAN.
7.2
Ethernet switches and infrastructure
Any switch used in an EqualLogic SAN should meet the requirements listed in this section.
March 2013
Dell EqualLogic Configuration Guide v14.1
7-62
Note: A detailed and frequently updated list of recommended switches is maintained in a
separate document:
Validated Components List for EqualLogic PS Series SANs
http://www.delltechcenter.com/page/EqualLogic+Validated+Components
Note: The FS7500 NAS appliance requires the use of 1Gb switches that meet the
requirements in this section.
An EqualLogic SAN consists of one or more hosts connected to one or more PS Series arrays through
a switched Ethernet network.
Note: The minimum network configuration for a PS Series Array consists of a connection
between Ethernet0 on each control module and a network switch. To increase
performance and availability, configure multiple network interfaces on an array and
connect them to multiple switches. EqualLogic does not support Direct Attached
Storage (DAS) configurations.
To support a high performance Dell EqualLogic SAN, switches must meet the following general
requirements:
•
•
•
•
•
•
Low latency: Switches with relatively high latency may cause SAN throughput performance to
degrade, and under high load conditions they could increase the risk of dropped connections.
Non-blocking backplane design: SAN Switches should be able to provide the same amount of
backplane bandwidth to support full duplex communication on ALL ports simultaneously.
Adequate buffer space per switch port: In addition to supporting data transfers between the hosts
and the SAN, Dell EqualLogic arrays also use the SAN to support inter-array communication and
data load balancing. For this reason, the more buffer space per port that a switch can provide the
better. Due to the multitude of buffer implementations used by switch vendors, Dell cannot
provide definitive guidelines as to how much is enough. Port buffers should be designed such that
data is not lost when traffic reaches extreme levels. Due to the clustered storage traffic patterns
used by the EqualLogic SAN architecture, switches that support “cut-through” mode are not
suitable for use in an EqualLogic SAN and may actually result in lower overall SAN performance.
Support for IEEE 802.3x flow control (passive and/or active) on ALL ports: Switches and network
interface controllers used in an EqualLogic SAN must be able to passively respond to any “pause”
frames received. If possible, you should use switches that have the ability to transmit “pause”
frames to external devices in the event that the device cannot adequately forward traffic in a timely
fashion.
Support for Jumbo Frames: This is not a requirement. But, the use of jumbo frames may yield
desirable results. Most iSCSI SAN implementations should benefit from using jumbo frames. The
actual impact on SAN throughput when using jumbo frames will depend on your workload’s I/O
characteristics.
Support for Rapid Spanning Tree protocol (IEEE 802.1w), or Cisco “portfast” functionality if the
SAN infrastructure will consist of more than two switches: For SAN infrastructures consisting of
March 2013
Dell EqualLogic Configuration Guide v14.1
7-63
•
•
•
more than 2 non-stacking switches, R-STP must be enabled on all ports used for inter-switch
trunks. All non-inter-switch trunk ports should be marked as “edge” ports or set to “portfast”.
Support for unicast storm control: iSCSI in general, and Dell EqualLogic SANs in particular can
send packets in a very “bursty” profile that many switches could misdiagnose as a virally induced
packet storm. Since the SAN should be isolated from general Ethernet traffic, the possibility of
actual viral packet storms occurring is non-existent. In an EqualLogic SAN, the switches must
always pass Ethernet packets regardless of traffic patterns.
Support for Inter-Switch Trunking (IST) or Stacking: IST support is required to link all switches in
SAN infrastructure together. For stacking capable switches, the use of stacking ports for IST is
assumed. A good rule of thumb for stacking link bandwidth would be a minimum 20 Gbps fullduplex.
Support for VLAN functionality.
Note: VLANs can be used as a means of isolating out-of-band EqualLogic management
traffic. Dell recommends this strategy where networking resources are limited.
Otherwise, it is recommended to use a physically separated network dedicated to iSCSI
traffic that is not shared with other traffic. If sharing the same physical networking
infrastructure is required, then use Data Center Bridging (DCB) for EqualLogic SAN.
•
Support for creating Link Aggregation Groups (LAG): For non-stacking switches, the ability to
bind multiple physical ports into a single logical link for use as an inter-switch trunk (IST) is
required. The switch should support designating one or more ports for IST (via Link Aggregation
Groups). The switch should support creation of LAGs consisting of at least eight 1Gbps ports or at
least two 10Gbps ports.
Note: For 1GbE SANs, using non-stacking switches to connect three or more EqualLogic
arrays into a single group may negatively impact SAN I/O throughput performance.
7.2.1
Connecting SAN switches in a Layer 2 network
When more than one SAN switch is required, each switch connected to the array group members will
be in the same subnet. These switches must be interconnected to provide a single switched Ethernet
fabric. Figure 19 shows the two common methods for interconnecting switches, using either stacking
switches or non-stacking switches.
Figure 19 Switch Interconnects
March 2013
Dell EqualLogic Configuration Guide v14.1
7-64
7.2.1.1
Stacking Switches
Stacking switches provide the preferred method for creating an inter-switch connection within a Layer
2 network infrastructure. Stacking is typically accomplished using a vendor proprietary, highbandwidth, low-latency interconnect that allows two or more switches to be connected in such a way
that each switch becomes part of a larger, virtual switch. A stackable switch will provide a set of
dedicated stacking ports. Installation of an optional stacking module may be required. Considerations
for stacking link bandwidth:
1Gb Switches
The stacking link bandwidth should be at least 10Gb/s in each direction on
each wire (full-duplex) to provide adequate throughput to support an
EqualLogic SAN consisting of 1Gb arrays.
10Gb Switches
The stacking link bandwidth should be at least 40Gb/s in each direction on
each wire (full-duplex) to provide adequate throughput to support an
EqualLogic SAN consisting of 10Gb arrays or a mix of 1Gb and 10Gb arrays.
7.2.1.2
Non-Stacking Switches
Non-stacking switches do not have a dedicated switch interconnect. In this case to create an
interconnect between the switches you must utilize one or more ports on each switch to create a Link
Aggregation Group (LAG). This type of Inter-switch connection should utilize link aggregation
functions (if provided by the switch) to aggregate multiple Ethernet ports into a single, logical high
bandwidth interconnect. There are several options depending on the vendor of the switch. Table 23
describes the most common options.
Note: For best performance and reliability, we recommend that all interconnection paths
between non-stacking switches (LAGs) use a dynamic link aggregation protocol such as
LACP.
March 2013
Dell EqualLogic Configuration Guide v14.1
7-65
Table 23 Link aggregation types
Link aggregation type
Notes
Static
Static link aggregation defines a set of links that provide a point to point
connection between two switches. These links may or may not provide
failover redundancy or traffic load management.
LACP
Link Aggregation Control Protocol is based on IEEE 802.3ad or IEEE
802.1AX. LACP is a dynamic LAG technology that automatically adjusts
to the appearance or disappearance of links within the defined LACP
group.
PAgP
Port Aggregation Protocol (PAgP) is a Cisco Systems® proprietary
networking protocol, which is used for the automated, logical
aggregation of Ethernet switch ports, known as an etherchannel. This
means it can only be used between Cisco switches and/or switches from
licensed vendors.
Vendor Proprietary
Several switch vendors may provide additional link aggregation options
that are completely proprietary or may be extensions to one of the two
previously mentioned solutions. In most cases, this type of link
aggregation solution is designed to reduce or eliminate the need – and
the overhead – of the Spanning Tree Protocol that must be used in the
two previous options. If available, these proprietary options should be
considered. They may be very useful in allowing the network
administrator to create a more efficient multi-switch layer 2 network
infrastructure for a SAN. Be aware that these proprietary solutions must
support the transmission of IEEE 802.1x flow control and jumbo frames
(if used) to properly support an EqualLogic SAN.
Using a LAG to Connect Stacked Switches
7.2.1.3
In some situations it may become necessary to expand the EqualLogic SAN by using more than one
single switch stack. For example, you can link multiple stacks by creating a multi-link LACP based LAG
between the switch stacks. A simplified stack plus LAG switch configuration is illustrated in Figure 20.
You should consider the following recommendations when designing this type of SAN:
•
•
•
•
If possible, use 10Gb connections for all links between each stack.
Distribute the links across multiple switches within each stack (this is known as a “crossstack” link aggregation group).
Use LACP or another type of dynamic link aggregation protocol.
Perform tests to determine the best hashing algorithm to use within the link aggregation
group (port channel).
Note: A multi-stack SAN infrastructure as described in this section may provide added
reliability to the SAN environment. But, it may also introduce additional latency and the
potential for lower throughput. The SAN designer will have to carefully consider the
performance and reliability implications.
March 2013
Dell EqualLogic Configuration Guide v14.1
7-66
Figure 20 Using a LAG to Interconnect Switch Stacks
7.2.2
Sizing inter-switch connections
Use the guidelines in Table 24 as a starting point for estimating Inter-switch connection sizes.
Table 24 Switch Interconnect Design Guidelines
Connection Speeds
Interconnection Guidelines
1GbE switches attached to
1GbE array controllers
1-5 arrays: 1Gb of IST bandwidth per active
array controller port (up to the aggregated
maximum bandwidth of the IST.
6+ arrays: Use 1-5 array rule, then add 1Gb of
additional bandwidth for each array added
10GbE switches attached to
10GbE array controllers
PS6010/PS6510 or PS4110/6110 (Random
Small Block Workloads):
1-5 arrays: 20 – 30Gb of IST bandwidth
between each switch
6+ arrays: At least 6Gb of IST bandwidth per
array between each switch
PS4110/PS6110 (Large Block Sequential
Workloads):
1-5 arrays: 10Gb of bandwidth per active array
controller port (up to the aggregated maximum
bandwidth of the IST).
6+ arrays: Use 1-5 array rule, then add 10Gb of
additional bandwidth for each array added
1GbE switches connecting to
10GbE switches in a mixed speed SAN
7.2.3
Actual requirements will vary. For more details
see Section 8.
Comparing inter-switch connection types
Table 25 provides details and recommendations for selecting interconnect options.
March 2013
Dell EqualLogic Configuration Guide v14.1
7-67
Table 25 Stacking vs. Inter-Switch Trunking
Interconnect
type
7.3
Primary purpose
Analysis
Stacking
Create a larger,
logical switch within
an isolated physical
location.
Advantages:
Easier to manage multiple switches as single
switch
Higher bandwidth than using link aggregation
and Ethernet
Not limited by Ethernet standards
Concerns:
Proprietary, cannot be used to interconnect
switches from different vendors
Increases cost of switch
Stack bandwidth overload risk
Recommendation:
Best way to scale a storage network in a single
location on a single subnet
Provides lower latency and higher bandwidth
than inter-switch trunking
Understand the stacking technology and limit
stack size in accordance with total throughput
requirements and connection counts.
Link
Aggregation
Groups (LAG)
Create a data path
between switches in
one location or
subnet with those in
another location or
subnet
Advantages:
Leverages Ethernet standard extensions
Can be used to interconnect switches from
different vendors
Can use Link Aggregation Protocols
(LACP/EtherChannel) to pool multiple 1GbE or
10GbE links into a single logical link providing
bandwidth and redundancy
Concerns:
Most solutions limited to 8 port link
aggregation group
Spanning Tree Protocol must be used if more
than two switches are used causing some links
to be “blocked” reducing bandwidth availability
LAG bandwidth overload risk
Recommendation:
Use when stacking is not available
Use when connecting to aggregation/core
switching infrastructure
Use when switches are from different vendors
Building a high-availability SAN
Designing a redundant SAN requires the availability of redundant NICs or HBAs on each server. A
redundant NIC configuration on the server requires at least two NICs installed into separate PCI slots
in the server. Table 26 below shows how to achieve redundant server NIC connection configurations
for a server with three installed NICs.
March 2013
Dell EqualLogic Configuration Guide v14.1
7-68
Table 26 Redundant Server NIC Configurations
LOM
NIC
NIC Connections
to SAN
7.3.1
Installed
NIC 1
Installed
NIC 2
X
X
-
-
X
X
X
-
X
Design guidelines for host connectivity in a redundant SAN
Using the Dell PowerEdge R610 server as an example, you configure redundant connection paths to
the SAN switches as shown in Figure 21 below. The R610 server shown in Figure 21 has one additional
dual-port PCI-E NIC installed. This configuration leaves two unused ports on the LOM controller for
connecting to the server LAN.
Note: As a best practice, we recommend using the same NIC devices in your server for all
connection paths to the iSCSI SAN. This will minimize the complexity of server
configuration management.
For maximum performance, ensure that the PCI express slot hosting the network
controller has the same specifications as the network controller. For example, if the
network controller has a PCIe x8 interface then ensure that it is installed into a PCIe slot
that can support 8 or more PCIe lanes.
Figure 21 Redundant NIC Connections from Server to SAN using one installed PCI-E NIC and LOM
An R610 server with two additional dual-port PCI-E NICs installed is shown in Figure 22 below. This
configuration leaves all four ports on the LOM controller available for other connections.
March 2013
Dell EqualLogic Configuration Guide v14.1
7-69
Figure 22 Redundant NIC Connections from Server to SAN using two installed PCI-E NICs
7.3.2
Multi-path I/O
There are generally two types of multi-path access methods for communicating from a host to an
external device. For general networking communications, the preferred method of redundant
connections is the teaming of multiple NICs into a single, virtual network connection entity. For
storage, the preferred method of redundant connection is the use of Multi-Path IO (MPIO). Though
some storage solution can and do support either method for iSCSI connectivity, EqualLogic requires
the use of MPIO to enable multiple NIC/HBA connections to be utilized for access to an EqualLogic
SAN.
EqualLogic MPIO Requirements
The following host port requirements must be met to use MPIO with EqualLogic SANs:
•
•
•
•
•
At least two (2) Ethernet ports are required on each host.
The host operating system must have a supported MPIO driver or service available.
The ports used for MPIO cannot be “teamed” to other ports.
The ports must be the same speed
The ports must be assigned IP addresses on the same subnet
EqualLogic MPIO General Recommendations
Follow this general set of guidelines for configuring MPIO on a host:
•
•
•
•
•
Configure volume access controls to use standard iSCSI IQN names (See Section 5.2.1 for
details). For a more secure configuration you can use the IQN name plus the CHAP security ID.
On each array enable at least two(2) ports for host connectivity.
Install the Dell provided MPIO extension features if available for the host operating system.
For Microsoft Windows, install the Device Specific Module (DSM) found in the Host Integration
Toolkit for Windows.
For VMware vSphere 4.1, install the EqualLogic Multipathing Extension Module.
March 2013
Dell EqualLogic Configuration Guide v14.1
7-70
•
For other operating systems, use the native MPIO functionality.
Configuring Microsoft Windows MPIO
Configure Microsoft Windows MPIO with the following initial configuration settings. Customized
settings may be required depending on the supported application(s).
•
•
•
•
•
•
•
Change the “Subnets included” field to include ONLY the subnet(s) dedicated to the SAN
network infrastructure.
Change the “Subnets excluded” field to include all other subnets.
The “Load balancing policy” should remain set to the default value of “Least queue depth”.
“Max Sessions per volume slice” should be set to the number of network ports dedicated to
SAN subnet (maximum of 4).
“Max sessions per entire volume” should be set to three (3) times the value of “Max sessions
per volume slice” (maximum of 12).
“Use MPIO for snapshots” should remain at the default setting.
“Use IPv6 or IPv4” should be set to IPv4 unless your network is configured to use IPv6 as the
default communications protocol.
Configuring VMware vSphere 4.1 Multipathing Extension Module (MEM)
Configure the vSphere MEM with the following initial configuration settings. Customized settings may
be required depending on the supported application(s).
•
•
•
•
Change the “Subnets included” field to include ONLY the subnet(s) dedicated to the SAN
network infrastructure.
“membersessions” should be set to the number of network ports dedicated to SAN subnet
(maximum of 4).
“volumesessions” should be left at the default value of 6.
“totalsessions” should be left at the default value of 512.
7.3.3
Equallogic iSCI SAN Design
This section will combine all of the SAN components and information provided so far to present a
redundant EqualLogic PS Series SAN design. We also include a series of examples illustrating partially
redundant and non-redundant SAN configurations.
Note: For FS Series NAS design, the only SAN design patterns presented in this section
that we recommend in support of the NAS Service are the fully redundant and fully
cabled examples shown in Section 7.3.4. Fully cabled and redundant connection
diagrams for the FS7500 are shown in shown in Section 13.1.
The information provided here does not address all of the possible variations in a customer network
environment. All information is presented using a set of basic reference designs that make the
following assumptions:
•
•
The SAN network is physically isolated from all other network traffic
The examples are based on best practice design principles.
March 2013
Dell EqualLogic Configuration Guide v14.1
7-71
•
•
•
Unless otherwise stated, all reference designs will provide end-to-end host to volume redundant
paths
A minimal number of switches will be illustrated to allow for the design concept to be understood.
Actual implementations will vary depending on your network infrastructure requirements.
If sharing physical switches with other non-SAN traffic, we assume all switches are VLAN capable.
7.3.4
Redundant SAN configuration
In a redundant iSCSI SAN, each component of the SAN infrastructure has a redundant connection or
path. The following figures show example connection paths necessary to create a Redundant SAN.
Figure 23 Redundant SAN Connection Paths: PS3000 to PS6000 Family Arrays
March 2013
Dell EqualLogic Configuration Guide v14.1
7-72
Figure 24 Redundant SAN Connection Paths: PS4100
Figure 25 Redundant SAN Connection Paths: PS6100
Note: For a production environment, the configuration examples shown above will
protect your access to data. These are the ONLY SAN configurations recommended by
Dell.
March 2013
Dell EqualLogic Configuration Guide v14.1
7-73
7.3.5
Partially redundant SAN configurations
Each of the SAN configurations shown in this section will allow host connectivity to data stored in the
SAN. These configurations are for reference only, and the methods shown apply to both PS3000PS6000 family controllers and PS4100/PS610 family controllers. They are not recommended for
production deployment since they do not provide end-to-end redundant connection paths.
Single array controller configurations
7.3.5.1
Table 27 below shows configurations using a single array controller.
Table 27 Single controller array configurations
Single NIC
Single Switch
Single Controller
Dual NIC
Single Switch
Single Controller
Dual NIC
Dual Switch
Single Controller
March 2013
Dell EqualLogic Configuration Guide v14.1
7-74
Dual array controller configurations
7.3.5.2
You can configure a Dell EqualLogic array to run using dual controllers. Table 28 below shows
configurations using a single array controller.
Table 28 Dual controller array configurations
Single NIC
Single Switch
Dual Controller
Single NIC
Dual Switch
Dual Controller
Dual NIC
Single Switch
Dual Controller
7.3.6
Minimum cabling scenarios: PS4100 and PS6100
The vertical port failover feature (described in Section 1.4.4) allows you to cable dual controller
PS4100 and PS6100 family arrays for maximum I/O bandwidth and controller redundancy while using
March 2013
Dell EqualLogic Configuration Guide v14.1
7-75
only one half of the available controller ports. The diagrams in Figure 26 and Figure 27 show how to
connect these arrays to accomplish this.
Note: The example configurations shown in this section only apply to the PS4100 and
PS6100 family arrays and are recommended only when you do not have available SAN
switch ports necessary to support fully cabled configurations.
Figure 26 Minimum cabling scenario: PS4100
March 2013
Dell EqualLogic Configuration Guide v14.1
7-76
Figure 27 Minimum cabling scenario: PS6100
March 2013
Dell EqualLogic Configuration Guide v14.1
7-77
8
Mixed speed environments - Integrating 1GbE and
10GbE SANs
With the introduction of 10GbE, there will be situations that require 1Gb arrays and 10Gb arrays
coexisting in the same SAN infrastructure. EqualLogic PS Series arrays support operation of 1Gb and
10Gb arrays within the same group. This section summarizes mixed speed SAN design guidelines that
are presented in much more detail in the following publications:
•
•
Best Practices for Deploying a Mixed 1 Gb/10 Gb Ethernet SAN using Dell EqualLogic Storage
Arrays:
http://en.community.dell.com/techcenter/storage/w/wiki/2640.deploying-mixed-1-gb-10gb-ethernet-sans-using-dell-equallogic-storage-arrays-by-sis.aspx
Integrating EqualLogic PS6x10 Arrays with Existing SANs:
http://www.equallogic.com/resourcecenter/assetview.aspx?id=9447
The potential advantages in running a mixed speed (1GbE and 10GbE) EqualLogic SAN include:
•
•
•
Not all of the application workloads on a SAN will require storage I/O performance that the
10Gb arrays provide. Thus, SAN administrators will have additional storage tiering flexibility
based on array I/O performance.
The PS Series Group Manager will allow the SAN administrator to still manage both types of
arrays within the same SAN group.
The ability to mix 1Gb and 10Gb arrays supports seamless operational coexistence during
during migration to a 10Gb SAN.
8.1
Design considerations
To properly implement a mixed speed SAN, you must pay close attention to the following design and
implementation considerations:
•
•
•
•
•
Ethernet switch feature set, port density and required configuration settings
Optimization of Rapid Spanning Tree Protocol behavior
Optimal switch interconnect pattern
Awareness of I/O workload patterns coming from 1Gb and 10Gb initiators vs. target volume
locations in the SAN
I/O performance implications when using mixed speed vs. segregated speed pools
To create a redundant, mixed speed iSCSI SAN, at a minimum we recommend that you start with dual
1GbE and dual 10GbE switches. Figure 28 shows an example SAN design, where two switches of each
type are used.
Referring to Figure 28:
•
•
The design is based on using features provided by the Dell PowerConnect 6248 1Gb Ethernet
switch and the Dell PowerConnect 8024 10Gb Ethernet switch.
The Dell PowerConnect 8024 is not a stackable switch, so a link aggregation group (LAG) is
used to create the inter-switch trunk paths.
March 2013
Dell EqualLogic Configuration Guide v14.1
8-78
•
•
•
Each of the 1Gb switches is configured with one dual-port 10GbE uplink module and one
stacking module. The 10GbE uplink modules are used for creating 20Gb LAG uplinks to the
10Gb switches.
Split Interconnect –The 20Gb LAG uplinks between the 1Gb and 10Gb switches are crossconnected so that each 10Gb switch physically connects to both switches in the 1Gb stack.
FS7500 Connection Path – The initial FS Series NAS product (FS7500) is a 1Gb only solution.
The NAS Reserve pool for the FS7500 could also be stored in a mixed 1Gb/10Gb PS Series
SAN. In this case the FS Series appliance must connect via the 1Gb switch path as shown in
Figure 28 and Figure 29.
The 10Gb switches are connected together using multiple 10Gb links. This provides us a path for
devices connected directly to the 10Gb switches to communicate without having to go through the
slower 1Gb switches.
Figure 28 Mixed speed redundant SAN using split interconnect between 1Gb and 10Gb switches
8.1.1
Optimizing Rapid Spanning Tree Protocol behavior
The LAG between the 10Gb switches in Figure 28 creates a loop in the network. Rapid Spanning Tree
Protocol (RSTP) will compensate for this by blocking paths as necessary. The optimal spanning tree
strategy for this network is to prevent RSTP from blocking the inter-switch trunk between the 10Gb
switches, thus causing some 10Gb traffic to traverse the slower 1Gb switches. To accomplish this you
need to be aware of which switch is acting as the root bridge in the spanning tree. For the mixed
speed SAN design shown in Figure 28 the root bridge is 1Gb SW#1. Based on this information you can
assign a link cost to ensure that the desired link configuration is achieved. Note the location of the
RSTP blocked path in Figure 28. We manually assigned a high link cost to one of the 20Gb uplink LAGs
so that it became the RSTP blocked path.
The straight uplink pattern must be used when using non-stackable switches
March 2013
Dell EqualLogic Configuration Guide v14.1
8-79
If you are using switches that do not support a stacking mode then you must use the straight
interconnect uplink pattern shown in Figure 29. Note the following design differences between the
split uplink pattern in Figure 28 and the straight uplink pattern in Figure 29:
•
•
A LAG is used to create the connection between 1Gb SW#1 and 1Gb SW#2.
A high rapid spanning tree link cost is assigned to the 1Gb switch LAG (note the location of the
RSTP blocked path in Figure 29). Doing this prevents 10Gb inter-switch traffic from having to
pass through the 1Gb switch LAG.
Figure 29 Mixed speed redundant SAN using straight interconnect between 1Gb and 10Gb switches
8.2
Mixed speed SAN best practices
The following list summarizes the important SAN design considerations for integrating 10Gb
EqualLogic arrays into existing 1Gb EqualLogic SANs.
•
When integrating 10Gb switches into your existing 1Gb environment, how you interconnect the
mixed speed switches (split vs. straight uplink) does not have a significant impact on performance
as long as the uplinks are sized appropriately to your workloads.
•
•
•
•
If your 1Gb switches are configured as a stack then you should use the split interconnect
pattern shown in Figure 28 by default.
If your 1Gb switches are not stacked, then you must use the straight interconnect pattern
shown in Figure 29.
When connecting 1Gb switches and 10Gb switches together you must always be aware of where
Rapid Spanning Tree is going to block links to make sure that 10Gb traffic (i.e. EqualLogic interarray data flow) never crosses the 1Gb switch.
You must configure pools and volumes in a way that minimizes impact to IO performance.
March 2013
Dell EqualLogic Configuration Guide v14.1
8-80
•
•
•
If you have predominately 1Gb initiators, start upgrading your arrays to 10Gb for comparable
or better performance across almost all situations.
If you have predominately 10Gb initiators, you should only access data and volumes residing
on 10Gb arrays (from those initiators). You may see high latency and retransmit rates when
10Gb initiators connect to 1Gb targets.
When adding 10Gb arrays, place them in separate pools from your 1Gb arrays.
March 2013
Dell EqualLogic Configuration Guide v14.1
8-81
9
Blade server chassis integration
Integrating the PowerEdge M1000e Blade Server Solution (or any third party blade chassis
implementation) requires additional SAN design considerations. Each M1000e can support up to three
separate networking “fabrics” that interconnect ports on each blade server to a pair of blade I/O
modules within each chassis fabric through an intervening chassis midplane interface. Each fabric is
associated with different interfaces on a given blade server as described in Table 29. Each blade server
has a “LAN on Motherboard” capability that is mapped to the I/O modules located in the Fabric A I/O
modules slots on the M1000e chassis and only supports 1Gb or 10Gb Ethernet networking depending
on the blade server model. In addition, each blade server has two “mezzanine” sockets for adding
additional networking options such as 1Gb or 10Gb Ethernet, Infiniband, or Fibre Channel cards. These
mezzanine cards are mapped to either the Fabric B or the Fabric C I/O modules.
Figure 30 illustrates the layout of the three fabric I/O modules located on the back of the M1000e
chassis.
Table 29 M1000e fabric mapping
Fabric
March 2013
LOM/MLOM
Mezzanine B
Mezzanine C
A
B
C
Dell EqualLogic Configuration Guide v14.1
9-82
Figure 30 Blade I/O modules and M1000e Chassis
There are three primary methods of integrating blade chassis with EqualLogic PS Series arrays.
•
•
•
Directly attaching EqualLogic arrays to the blade I/O modules on each chassis.
Using Pass-Through module to external switches
Utilizing a two-tier design by creating an external SAN switch infrastructure to host all array
connections and using the blade I/O modules as “host access” switches.
For each of these three general SAN design strategies, the user must make design decisions with
respect to the type of interconnects to use between the blade I/O modules and/or the external standalone switches. Depending on the design recommendation, stacking, link aggregation, or a
combination of both types of interconnect may be used. In Section 9.1 we will discuss strategies for
directly attaching arrays to blade I/O modules within one or more M1000e chassis. In Section 9.2 we
discuss connecting M1000e blade chassis to an external switch using pass-through module. In Section
9.3 we discuss strategies for connecting M1000e blade chassis to an external SAN infrastructure
utilizing a two-tier design.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-83
9.1
Designing a SAN using blade chassis I/O modules with
arrays directly attached
The following points should be taken into consideration when planning a SAN solution that requires
directly attaching EqualLogic arrays to M-Series blade I/O modules:
•
Limited number of externally accessible ports typically available on blade I/O modules
Current M-Series blade I/O modules have limited numbers of externally accessible Ethernet
ports. For a 1GbE SAN solutions each array will require up to four ports per I/O module and
for a 10GbE SAN solution each array will require up to two ports per I/O module. Based on
these maximum requirements, the table below provides a breakdown on the number of
externally accessible ports on each model of PowerConnect M-Series blade I/O modules and
the maximum number of arrays that could be directly attached to a single M1000e blade
chassis.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-84
Table 30 Blade I/O Module options for EqualLogic
Maximum Available
External Ports
Ports Recommended
for Interconnect
Arrays Supportable per
M1000e
PowerConnect
M6220
8x 1GbE
0(Stackable)
2
PowerConnect
M6348
16x 1GbE
0(Stackable)
4
PowerConnect
M8428-k
8x 10GbE
2x10GbE(PS6010/6510)
3x10GbE(PS4110/6110)
3(PS6010/6510)
5(PS4110/6110)
PowerConnect
M8024-k
8x 10GbE
2x10GbE(PS6010/6510)
3x10GbE(PS4110/6110)
3(PS6010/6510)
5(PS4110/6110)
2x40GbE Fixed +
2x option slots
(2x40GbE, 4xSFP+,
4x10GbaseT)
2x40GbE (Multiple
chassis, Stacked)
16(PS4110/6110)
16(14x PS6010/6510)
(+ 2x PS-M4110)
Cisco Catalyst Blade
Switch 3032
8x 1GbE
4x1GbE
1
Cisco Catalyst Blade
Switch 3130G
4x 1GbE
0(Stackable)
1
Cisco Catalyst Blade
Switch 3130X
4x 1GbE
0(Stackable)
1
16
n/a
n/a
Force10 MXL
1GbE or 10GbE
Pass-Through
Module
•
•
•
•
•
•
All switches within the SAN infrastructure, including blade I/O modules, must provide a path
from all ports hosting array connections to all other ports within the SAN.
An important networking requirement for EqualLogic SAN infrastructure design is that all
switches must be part of the same Layer 2 Ethernet fabric if they will be hosting connections
from EqualLogic arrays. For this reason, if arrays are going to be directly attached to one or
more blade chassis, then the I/O modules within the Fabric that will be hosting the arrays must
be interconnected. Since the I/O modules are not interconnected via the chassis midplane, the
only alternative is to use external ports (stacking or Ethernet) to make these inter-switch
connections.
Maximum stack size versus practical stack size
Depending on the I/O module model, the maximum number of switches allowed in a single
stack may be different. In addition, the number of switches supported by the switch may not
be optimal for a SAN using EqualLogic arrays. SAN solutions tend to be sensitive to latency
and adding multiple hops between a host and target may create situations where latency
becomes unacceptable for your application. Testing is advised.
Each M1000e enclosure will contribute two switches to the stack placing an upward limit on
the number of enclosures that can be interconnected into a single SAN network. Regardless
of the maximum number of I/O modules supported in a single stack, Dell recommends not
going above six switches in the stack due to possible hop related latency.
M-Series I/O module Stacking Compatibility
March 2013
Dell EqualLogic Configuration Guide v14.1
9-85
•
Not all M-Series I/O modules can be stacked together or with external switches. Typically,
each M-Series I/O modules model can only stack with modules of the same exact model. It
may also be possible to stack M-Series I/O modules with some “stack compatible” stand-alone
PowerConnect switch models. The table below provides stack compatibility information.
Table 31 Stack Compatibility for M-Series I/O modules
Stack Compatible Switches
Maximum
Stack Size
Recommended
Stack Size
PowerConnect M6220
PowerConnect M6220
PowerConnect 6224
PowerConnect 6248
10
4
PowerConnect M6348
PowerConnect M6348
PowerConnect 7024
PowerConnect 7048
12
6
PowerConnect M8428-k
N/A
N/A
N/A
PowerConnect M8024-k
M8024-k
6
6
Force10 MXL
6
4
N/A
N/A
N/A
Cisco Catalyst Blade Switch 3130G
Catalyst BS3130G
Catalyst BS3130X
8
4
Cisco Catalyst Blade Switch 3130X
Catalyst BS3130G
Catalyst BS3130X
8
4
Force10 MXL
Cisco Catalyst Blade Switch 3032
SAN Design for a Single M1000e Enclosure
For a single M1000e enclosure, there is a requirement to use two blade I/O modules for redundancy.
Depending on the blade I/O module model, you may use either stacking or link aggregation as
described earlier in Table 30. Where stacking is supported, this will most likely be the primary method
of interconnecting the blade I/O modules to create a single switch fabric. SAN switches can be
located in any of the three enclosure fabrics and it is not required that the blade I/O modules be in the
same enclosure fabric, though it is typical that the two I/O modules would be in the same enclosure
fabric.
Placing the blade I/O modules in the same fabric does remove one aspect of high availability in that
each blade server will have both of the SAN ports located on the same fabric mezzanine card. This
creates a potential single point of failure if the mezzanine card as a whole fails.
One alternative configuration would be to place the two blade I/O modules into enclosure slots
associated with two different enclosure fabrics (B1 and C1 for example). This has the advantage that
each blade server will have its Ethernet ports connected to these blade I/O modules on two different
mezzanine cards. This ensures that even if a single mezzanine card fails, there is still an active port on
the SAN network.
Table 32 and Table 33 illustrate the stacking strategies between the two I/O modules in each of these
two configurations as described.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-86
Table 32 Single M1000e enclosure stacking single fabric
Single Fabric Stacked Configuration
Advantages
•
•
Concerns
Consistent M1000e fabric management that
adheres to M1000e recommended practices
for blade IO configuration
Reduces administration overhead
March 2013
•
•
All blade server Ethernet ports used for SAN
reside on the same mezzanine card resulting
in a potential single point of failure
Upgrading switch FW will require scheduled
downtime for SAN network
Dell EqualLogic Configuration Guide v14.1
9-87
Table 33 Single M1000e enclosure stacking dual fabric
Dual Fabric Stacked Configuration
Advantages
•
Concerns
Ensures that Ethernet SAN ports on each
blade server will be distributed across two
different mezzanine cards for a more highly
available solution
•
•
Does not adhere to recommended practices
for M1000e blade IO configuration
Upgrading switch FW will require scheduled
downtime for SAN network
As discussed earlier, one of the concerns when creating a SAN that consists of a single stack is that a
single stack does not provide for a highly available solution when attempting to perform switch
maintenance such as switch firmware updates. If there is a requirement that the SAN solution cannot
be brought down for regularly scheduled maintenance, then stacking may not be the right option and
you should consider link aggregation methods for switch interconnect.
Table 34 and Table 35 illustrate the non-stacking strategies between the two I/O modules in each of
these two configurations as previously described.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-88
Table 34 Single enclosure link aggregation single fabric
Single fabric non-stacked configuration
Advantages
•
•
Concerns
Consistent M1000e fabric management that
adheres to M1000e recommended practices
for blade IO configuration
Switch FW can be upgraded without
requiring network to be brought offline
•
•
All blade server Ethernet ports used for SAN
reside on the same mezzanine card resulting
in a potential single point of failure
Spanning Tree must be considered if
uplinking SAN to external switches or
additional M1000e enclosures
Table 35 Single Enclosure Link Aggregation Dual Fabric
Dual fabric non-stacked configuration
Advantages
•
•
Concerns
Ensures that Ethernet SAN ports on each
blade server will be distributed across two
different mezzanine cards for a more highly
available solution
Switch FW can be upgraded without
requiring network to be brought offline
March 2013
•
•
Does not adhere to recommended practices
for M1000e blade IO configuration
Spanning Tree must be considered if
uplinking SAN to external switches or
additional M1000e enclosures
Dell EqualLogic Configuration Guide v14.1
9-89
The 10GbE PowerConnect M-Series I/O modules do not provide a dedicated stacking interface and
must be interconnected using available, external Ethernet ports in conjunction with a link aggregation
protocol such as LACP or “front-port” stacking. Due to the limited number of external ports available
on Dell’s PowerConnect M-Series blade I/O modules, SAN growth can be limited. For 10Gb SAN
solutions that will require multiple arrays and/or multiple M1000e chassis, it is recommended that you
do not consider directly attached arrays and follow the guidelines described in Section 9.2.
If reducing administrative overhead is the goal, the Direct Attached design with a inter-fabric stack is
the best option, with the storage direct attached design requiring fewer cables thatn the host passthrough design.
Direct Attached – Single Chassis Stacked and/or LAG
Figure 31 SAN modules in single enclosure fabric
9.1.1
SAN design for multiple M1000e enclosure
For a multiple M1000e enclosure SAN solutions, there are two primary strategies to consider, the first
– described here – is directly attaching arrays to the individual M1000e chassis and interconnecting
the chassis M-Series I/O modules in each chassis, and the second strategy – described in Section 9.2
– is to build an external network for array consolidation and using chassis M-Series I/O modules as
March 2013
Dell EqualLogic Configuration Guide v14.1
9-90
“host access” switches. In most cases, the latter method of SAN design is preferable to the former, but
this section will describe strategies to build a SAN that consists of multiple M1000e enclosures.
As in the previous section, there are two methods that must be considered when interconnecting MSeries I/O modules: stacking and non-stacking. Due to the limited number of external ports available
on M-Series I/O modules, stacking should be the primary method of interconnecting M1000e
enclosures as these external ports will need to be used for attaching EqualLogic arrays.
As just indicated stacking of M-Series I/O modules should be the primary method of building a
multiple enclosure SAN with arrays directly attached to the M-Series I/O modules. As with the single
enclosure discussion, stacking has some advantages and disadvantages from a network management
perspective and network availability perspective. Take this into consideration before deciding to
implement a multi-enclosure SAN using directly attached arrays.
Table 36 illustrates the primary method for building a multi-enclosure SAN with directly attached
arrays. Each M1000e enclosure should be able to host one to four arrays depending on the M-Series
I/O module used. By stacking each enclosure’s fabric I/O modules into a single stack, a single SAN
network infrastructure can be constructed.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-91
Table 36 Single stack multiple M1000e enclosure direct connect SAN
Single stacked, multiple M1000e enclosure configuration
Advantages
•
•
Concerns
Simplified switch management
High-bandwidth interconnect
•
•
Not highly available, single stack solution
cannot be maintained without scheduled
downtime.
Additional scalability will require SAN
network redesign and possible downtime.
Other notes
•
Limited scalability due to potentially excessive hop-counts and latency. Recommend no more
than 2-3 enclosures in this design
March 2013
Dell EqualLogic Configuration Guide v14.1
9-92
Table 37 Non-stacked, multiple M1000e enclosure direct connect SAN
Non-Stacked, Multiple M1000e Enclosure Configuration
Advantages
•
•
Concerns
Required for non-stacking M-Series I/O
modules to support direct attached arrays.
Array firmware can be updated without SAN
being brought down for maintenance
•
•
•
Spanning tree protocol will block one link to
prevent a loop
Spanning tree may change if M-series I/O
modules are uplinked to other switches
(such as when attempting replication to
remote data center).
Requires more cables than stacked solutions
Other notes
•
•
Should only use 10GbE links for this strategy to ensure adequate bandwidth between modules
and enclosures.
Limited scalability due to potentially excessive hop-counts and latency. Recommend no more
than 2-3 enclosures in this design
If large SAN solutions are required, the design just discussed may not be the adequate. While the
stacking interfaces on the M-series I/O modules is higher than we could obtain using a link
aggregation group consisting of 1Gb ports, it is still somewhat limited. Also, if you foresee needing
more M1000e enclosures than the two to three enclosure maximum recommended, then a different
SAN design strategy should be considered.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-93
9.2
Designing a SAN using blade Pass-through module
This SAN design include configurations where the blade server host ports are directly connected to
TOR switches using 1GbE pass-through IOM in the M1000e blade chassis. The storage ports are also
connected to the TOR switches,
Note that because an ISL stack is not a recommended configuration due to a lack of SAN availability
during stack reloads, only one single switch tier design with ISL stack was tested – Blade IOM switch
only – and the TOR switch only with ISL stack design was excluded.
This SAN design provides sufficient ISL bandwidth between the two switches by creating a LAG
between the external switches. Since there is only a single tier of switches, there is no uplink from the
blade IOM switches. The remaining ports on each switch can accommodate the connection of
multiple PS Series array members. The host/storage port ratio with the maximum number of array
members is 1:1.
The following diagram illustrates how to directly connect the two switches and how the two switches
are connected by an ISL LAG. Also note that the corresponding port on the passive controller is should
be connected to a different switch than the port on the active controller, ensuring that the port-based
failover of the array member will connect to a different switch upon port, cable or switch failure.
Management and host LAN networks are shown for reference.
Pass-Through modules are supported for use with EqualLogic SAN solutions. The Pass-Through
module provides a simple, direct path from each blade server’s optional Ethernet mezzanine card to
an externally accessible port. These ports can then be connected to one or more external switches
that are hosting PS Series arrays.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-94
Table 38 M-Series I/O module Pass-Through to external switch infrastructure
M-Series I/O module Pass-Through to external switch infrastructure
Advantages
•
•
•
•
•
•
•
Concerns
Fewer switch hops
Fewer switches to manage
SAN can support M1000e enclosures and rack or tower
stand-alone servers
SAN traffic isolated outside of M1000e enclosure
Less expensive that multi-tiered switch solution
Supports non-dell branded switching as a single vendor
solution
No firmware to update on the pass-through modules.
•
Complex cabling in a rack
environment.
Other Notes
•
•
•
•
Additional M1000e enclosures are added in similar fashion to diagram above
Supported with stacked or non-stacked external SAN infrastructures
This strategy works for a single fabric or dual fabric M-Series I/O module configuration (See
Section 0 for details)
This is the simplest choice if external switches are from a different vendor.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-95
9.3
Designing a SAN using blade chassis I/O modules as
host access to external switches for array connection
When attempting to have multiple M1000e blade chassis connected to EqualLogic SAN arrays, or if
there is a need to also attach traditional rack or tower servers to the SAN, it is strongly recommended
that you consider creating a SAN infrastructure that does not directly attach arrays to the blade I/O
modules, but rather consider using a set of external switches to host array connections.
The basic strategy is a slight modification from the basic SAN design guidelines used for SANs that do
not have M1000e enclosures discussed in Section 7.3.3 Equallogic iSCI SAN Design, so this discussion
will start with that basic SAN reference architecture. The difference will be the connection of each
M1000e enclosure’s M-Series I/O modules to at least two switches in the external SAN infrastructure
using either Link Aggregation (the primary method) or Stacking interfaces – if the external switches are
stack compatible (See Stack Compatibility for M-Series I/O modules). Even if the M-Series I/O
modules and the external switches are stack compatible, it is recommended that stacking not be used
due to the inability to upgrade switch firmware without taking the SAN network offline.
The following tables provide various methods for uplinking M1000e enclosures to an external SAN
network infrastructure and provide insights into the advantages and concerns that may affect your
configuration, depending on your business environment. In the end, it depends on your specific needs
to determine which configuration is best for your solution.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-96
Table 39 Horizontal Stack with Vertical LAG
Connecting M1000e to Stacked external switch infrastructure
Advantages
•
•
•
•
•
•
•
Concerns
Support configuring LAGs between M-series I/O modules
across both external switches in stack (cross-switch LAG)
SAN Array-to-array traffic is isolated outside of the M1000e
enclosures
Supports ability to scale arrays
Supports ability to scale host.
Supports sharing SAN with non-M1000e servers.
External stacked switches can be expanded by adding
additional switches to stack as more ports are needed.
Can support different vendor switches in the external tier
•
•
Must bring external switch
stack down to apply switch
firmware updates.
Must bring internal switch
stack down to apply switch
firmware updates.
Other notes
•
•
If available, use 10Gb links between M-series I/O modules and external switches
Additional M1000e enclosures are added in similar fashion to diagram above.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-97
Table 40 3-Way LAG
Connecting M1000e to Non-Stacked external switch infrastructure
Advantages
•
•
•
•
•
•
Concerns
Switch Firmware updates can be made on all
switches without downtime
SAN Array-to-array traffic is isolated outside
of the M1000e enclosures
Supports ability to scale arrays
Supports ability to scale host.
Supports sharing SAN with non-M1000e
servers.
Can support different vendor switches in the
external tier
•
•
•
Limited scalability of LAG bandwidth.
Spanning Tree may block some links if a
loop is created.
All switches must be managed separately
Other Notes
•
•
•
•
Recommend using 10GbE or larger links within inter-switch LAGs to provide better bandwidth
scalability.
Care needs to be taken to ensure that addition of switches does not cause a spanning tree
reconfiguration.
Do not interconnect the M-series I/O modules together.
Additional M1000e enclosures are added in similar fashion to diagram above.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-98
Table 41 4-Way Stack
Connecting M1000e to Non-Stacked external switch infrastructure
Advantages
•
•
•
•
•
•
Concerns
SAN Array-to-array traffic is isolated outside
of the M1000e enclosures
Fixed hops between hosts and array volumes
regardless of number of M1000e enclosures
Supports ability to scale arrays
Supports ability to scale host.
Supports sharing SAN with non-M1000e
servers.
Ease of Administration
•
•
•
Must bring external switch stack down to
apply switch firmware updates.
Must bring internal switch stack down to
apply switch firmware updates.
Switches must be stack compatible
Other Notes
•
•
Recommend using 10GbE or larger links within inter-switch LAGs to provide better bandwidth
scalability.
Additional M1000e enclosures are added in similar fashion to diagram above.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-99
Table 42 Vertical Stack with Horizontal LAG
Connecting M1000e to Non-Stacked external switch infrastructure
Advantages
•
•
•
•
•
•
Concerns
Switch Firmware updates can be made on all
switches without downtime
Fixed hops between hosts and array volumes
regardless of number of M1000e enclosures
Supports ability to scale arrays
Supports ability to scale host.
Supports sharing SAN with non-M1000e
servers.
High Availability
•
•
Limited scalability of LAG bandwidth.
Spanning Tree may block some links if a
loop is created.
Other Notes
•
•
•
•
Recommend using 10GbE or larger links within inter-switch LAGs to provide better bandwidth
scalability.
Care needs to be taken to ensure that addition of switches does not cause a spanning tree
reconfiguration.
Do not interconnect the M-series I/O modules together.
Additional M1000e enclosures are added in similar fashion to diagram above.
March 2013
Dell EqualLogic Configuration Guide v14.1
9-100
Table 43 Dual M-Series Stacks to Stacked external switch infrastructure
M-Series I/O module Stack(s) to Stacked external switch infrastructure
Advantages
•
•
•
•
•
•
•
Concerns
Each M-Series I/O module stack can be updated independently
without bringing SAN down.
SAN Array-to-array traffic is isolated outside of the M1000e
enclosures
Fixed hops between hosts and array volumes regardless of
number of M1000e enclosures
Supports ability to scale arrays
Supports ability to scale host.
Supports sharing SAN with non-M1000e servers.
Fewer “switches” to manage since each stack is managed as a
single switch
•
•
•
Firmware updates on
external stack requires
scheduled downtime
Limited scalability of
LAG bandwidth.
Spanning Tree may
block some links if a
loop is created.
Other Notes
•
•
•
•
If available, use 10Gb links between M-series I/O modules and external switches
M-series I/O module stacks do not need to be interconnected since no arrays are attached
directly to M-series I/O modules.
Additional M1000e enclosures are added in similar fashion to diagram above. Enclosure
scalability is greater if 10Gb links are used between M-series I/O module stacks and the external
stack
This strategy works for a single fabric or dual fabric M-Series I/O module configuration (See
Section 0 for details)
March 2013
Dell EqualLogic Configuration Guide v14.1
9-101
Table 44 M-Series I/O module Stack(s) to Non-Stacked external switch infrastructure
M-Series I/O module Stack(s) to Non-Stacked external switch infrastructure
Advantages
•
•
•
•
•
•
•
Concerns
Each M-Series I/O module stack can be updated independently
without bringing SAN down.
SAN Array-to-array traffic is isolated outside of the M1000e
enclosures
Fixed hops between hosts and array volumes regardless of
number of M1000e enclosures
Supports ability to scale arrays
Supports ability to scale host.
Supports sharing SAN with non-M1000e servers.
Fewer “switches” to manage since each stack is managed as a
single switch
•
•
•
Firmware updates on
external stack requires
scheduled downtime
Limited scalability of
LAG bandwidth.
Spanning Tree may
block some links if a
loop is created.
Other Notes
•
•
•
•
If available, use 10Gb links between M-series I/O modules and external switches
M-series I/O module stacks do not need to be interconnected since no arrays are attached
directly to M-series I/O modules.
Additional M1000e enclosures are added in similar fashion to diagram above. Enclosure
scalability is greater if 10Gb links are used between M-series I/O module stacks and the external
stack
This strategy works for a single fabric or dual fabric M-Series I/O module configuration (See
Section 9.1 for details)
March 2013
Dell EqualLogic Configuration Guide v14.1
9-102
10
Fluid File system
The Dell EqualLogic FS Series NAS Appliance adds scale-out unified file and block Network Attached
2
Storage (NAS) capabilities to any EqualLogic PS Series iSCSI SANs. The key design features and
benefits provided by the EqualLogic FS Series Appliance include:
•
•
•
•
A scalable unified (block and file), virtualized IP based storage platform.
Seamless capacity and performance scaling to meet growing storage demands. By increasing the
number of EqualLogic arrays and FS Series controllers you can scale both storage capacity and I/O
performance as needed.
A single configuration and management UI (the EqualLogic Group Manager) from which both
block and file storage features are controlled.
The Dell Fluid File System (FluidFS) is a high-performance scale-out file system capable of
presenting a single file-system namespace through a virtual IP address, regardless of the number
of NAS appliances in the cluster. FluidFS is designed to optimize file access performance and
hardware utilization while eliminating capacity constraints.
Additionally, the FS Series Firmware V2.0 adds the following features:
•
•
•
•
Support for FS7600 and FS7610 NAS appliances
NAS container replicaton: Version 2.0 of the FS Series Firmware supports NAS container
replication, in which a file-system-consistent copy of a container is replicated from one NAS
cluster to another. NAS container replication requires FS Series Firmware Version 2.0 on the
EqualLogic FS Series NAS appliances.
NAS antivirus service for CIFS shares: NAS clusters support virus scanning by using an antivirus
server. You can configure up to four antivirus servers to scan files stored in CIFS shares (NFS is not
supported). The NAS cluster solution contains integration with industry standard ICAP-enabled
antivirus software to ensure files written from CIFS clients are virus-free. The antivirus host must
run Symantec ScanEngine 5.2, which is ICAP-enabled.
New monitoring features: Version 2.0 of the of the FS Series Firmware introduces support for the
following new monitoring panels:
•
•
•
•
2
NAS Snapshot Schedules – Provides information about NAS container snapshot operation
according to the regular schedules that you configure by using the Create Schedule (snapshot)
activity. You can enable, disable, modify, and delete schedules from this panel.
NAS Replication Schedules – Provides information about NAS container replication operations
according to the regular schedules that you configure by using the Create Schedule (replica)
activity. You can enable, disable, modify, and delete schedules from this panel.
Outbound NAS Replication – Provides information about replicated NAS containers where the
source container exists on the local group, replicating to a replica container on a partner
group.
Inbound NAS Replication – Provides information about replicated NAS containers where the
replica container exists on the local group, replicated from a source container on a partner
group.
Any new or existing EqualLogic PS Series array running controller firmware version 5.1 or later
March 2013
Dell EqualLogic Configuration Guide v14.1
10-103
10.1
FS Series architecture
The FS Series appliance connects to an EqualLogic PS Series SAN via standard iSCSI connection paths
with iSCSI based block level I/O taking place between the FS Series appliance initiators and the target
volumes on the PS Series arrays. The FS Series controllers host the FluidFS based NAS file system and
all front-end client protocol connection management and load balancing, as well as manage all backend data protection and high-availability functions.
Figure 32 illustrates the EqualLogic Unified Storage Architecture.
•
•
•
It is a unified solution because the system is providing both block and file level storage.
The storage pool shows a single pool that contains a NAS Reserve (managed by the NAS
Appliance) plus other volumes that are providing block I/O storage targets not used by the FS
NAS appliance. To emphasize this flexibility, we show a group of servers connected to the NAS
Appliance through the LAN Switches (acting as CIFS/NFS clients) and connecting directly to
the iSCSI SAN switches for block I/O.
The EqualLogic Group Manager provides a single user interface for configuring and managing
PS Series and FS Series functions.
Figure 32 EqualLogic Unified Storage architecture
10.1.1
FS Series solution for file only storage
In a file only scenario the initiators on the FS Series appliance are the only iSCSI clients connecting to
the PS Series arrays. The pools and volumes in the arrays provide storage for FS Series appliance file
I/O only. This scenario is shown in Figure 33.
March 2013
Dell EqualLogic Configuration Guide v14.1
10-104
Figure 33 FS Series NAS (file only)
10.2
Dell FluidFS
Dell EqualLogic FS Series appliances, coupled with PS Series arrays, offer a high performance, high
availability, scalable NAS solution.
FS Series Firmware V2.0 adds the following features:
•
•
•
•
NAS Container Replication
NAS Antivirus Service for CIFS Shares
New Monitoring Features
Support for FS7600 and FS7610 Appliances
3
Dell FluidFS is a high performance clustered file system. It provides fully interoperable multi-protocol
file sharing for UNIX, Linux, and Windows® clients using standard CIFS and NFS file access protocols
and authentication methods (Active Directory, LDAP, NIS). Some of the key features provided in
FluidFS include:
Single namespace: FluidFS presents a single file system namespace (up to 509TB usable capacity),
accessible via a virtual IP address, regardless of cluster size.
High availability and reliability: The FS Series architecture includes a high speed cluster interconnect
between controller nodes, write cache mirroring, failsafe journaling, and data integrity checks. In a
FluidFS based FS Series cluster, a controller node can fail without affecting data availability or causing
data loss, even if write operations were occurring at the time of failure.
3
FluidFS technical whitepaper: http://www.dellstorage.com/WorkArea/DownloadAsset.aspx?id=1578
March 2013
Dell EqualLogic Configuration Guide v14.1
10-105
Snapshots: FluidFS snapshots are read only and redirect-on-write. They are created and managed by
the FS Series appliance to provide file system level snapshot capability. They function independently of
and have no impact on the operation of PS Series array based snapshots.
Note: FS Series FluidFS snapshots and PS Series volume based snapshots function
independently and have no impact on each other.
Please see the following whitepaper for more information on Dell FluidFS:
Dell Fluid File System:
http://www.dellstorage.com/WorkArea/DownloadAsset.aspx?id=1578
NAS Replication: NAS replication is point-in-time and uses file system snapshot technology. NAS
replication can be configured on a per NAS container basis.
March 2013
Dell EqualLogic Configuration Guide v14.1
10-106
11
FS Series NAS Appliances
The FS7500 is the premier offering in the Dell FS Series product line. Table 45 lists the basic functional
details for each FS Series product.
Table 45 FS Series Models
11.1
FS Series Model
System Components
I/O Connections
FS7500
2 x 1U Intel/Linux based
NAS appliance; dual
active-active controller
configuration
1 x 1U Battery backup
power supply unit (BPS)
Client LAN: 8x1GbE per
appliance (4 per controller
node)
iSCSI SAN: 8x1GbE per
system (4 per controller
node)
Controller Interconnect:
8x1GbE per system (4 per
controller node)
FS7600
2U NAS appliance with
two active/active NAS
controllers.
2 x Intel 5620 Quad Core
processors per NAS
controller
24GB memory per NAS
controller
Redundant hot pluggable
power supples and fans
Client LAN: 8x1GbE per
appliance (4 per controller
node)
iSCSI SAN: 8x1GbE per
system (4 per controller
node)
FS7610
2U NAS appliance with
two active/active NAS
controllers.
2 x Intel 5620 Quad Core
processors per NAS
controller
24GB memory per NAS
controller
Redundant hot pluggable
power supples and fans
Client LAN: 4x10GbE per
appliance (2 per controller
node)
iSCSI SAN: 4x1G0bE per
system (2 per controller
node)
Equallogic NAS appliance supported configuration limits
The features and supported configuration limits for the NAS system are provided in Table 46.
March 2013
Dell EqualLogic Configuration Guide v14.1
11-107
Table 46 NAS cluster configuration limits
March 2013
Attribute
Single NAS appliance cluster
(Two NAS controllers)
Dual NAS appliance cluster
(Four NAS controllers)
Maximum NAS reserve
509 TB usable
509 TB usable
Minimum NAS reserve
512 GB
1024 GB
Maximum file size
4 TB
4 TB
Maximum local groups
300
300
Files in a cluster
64 billion
128 billion
Directories in a cluster
34 billion
68 billion
Containers in a cluster
256
512
Minimum container size
20 MB
20 MB
Maximum container size
Available NAS Reserve
Available NAS Reserve
Snapshots in a cluster
10,000
10,000
Snapshot schedules for a
container
64
64
Snapshot schedules in a
cluster
1024
1024
Snapshots per container
512
512
NFS exports in a cluster
1024
1024
CIFS shares in a cluster
1024
1024
CIFS active client
connections in cluster
1500
3000
Local users in a cluster
300
300
Maximum quota rules per
NAS container (user
quotas)
512
512
Maximum quota rules per
cluster (user quotas)
100,000
100,000
Maximum number of
containers enabled for
replication
100
100
Maximum number of
active container
replications in progress
(rest are queued)
10
10
Dell EqualLogic Configuration Guide v14.1
11-108
Single NAS appliance cluster
(Two NAS controllers)
Dual NAS appliance cluster
(Four NAS controllers)
Maximum replication
partners
16
16
Number of NAS
controllers in replication
source and destination
clusters
Each NAS cluster in the
replication partnership must
contain the same number of
NAS controllers.
Each NAS cluster in the
replication partnership must
contain the same number of
NAS controllers.
Attribute
The limits shown for files and directories, containers, NFS exports, CIFS shares, and container snapshots are the maximum supported limits. The group will not prevent you from exceeding a limit and
will not warn you when you reach or exceed a limit.
11.2
Initial NAS cluster valid configurations
The table below describes the configurations allowed during initial configuration.
Table 47 Initial valid configurations
March 2013
Cluster
Controller Pair 1
Controller Pair 2
Two-controller
FS7500
NA
Two-controller
FS7600
NA
Two-controller
FS7610
NA
Four-controller
FS7500
FS7500
Four-controller
FS7500
FS7600
Four-controller
FS7600
FS7600
Four-controller
FS7610
FS7610
Dell EqualLogic Configuration Guide v14.1
11-109
Table 48 Valid add-controllers pairs
11.3
Controller Pair 1
Controller Pair 2
FS7500
FS7500
FS7500
FS7600
FS7600
FS7600
FS7600
FS7500
FS7610
FS7610
FS7500 system components
4
The system components in the initial release of the EqualLogic FS Series NAS appliance (the FS7500)
consist of two controller node units and one backup power supply (BPS) shelf containing two
redundant power supplies. The system components and required power cabling paths are shown in
Figure 34.
Figure 34 FS7500 system components and power cabling
4
For detailed information on racking and power supply cabling see the FS7500 Installation and Setup Guide.
March 2013
Dell EqualLogic Configuration Guide v14.1
11-110
System controllers
The system contains dual active-active controller nodes with large onboard battery-backed caches
and 24 GB of battery protected memory per node. They operate in an active-active environment
mirroring the system’s cache. Each node regularly monitors the BPS battery status. They require the
BPS to maintain a minimum level of power stored during normal operation to ensure that they can
execute a proper shutdown procedure in the event of power interruption.
Backup power supply
The BPS is designed to provide power necessary for controllers to execute a proper shutdown
procedure. The BPS also enables the FS7500 controllers to use their onboard cache as NVRAM. This
provides the clustered solution with enough time to write all the data from the cache to disk in the
event a controller node experiences a loss of power.
11.4
FS7500 file system operation on controller failover
In the FS7500 system, writes to one controller node cache are mirrored to the other controller node
cache before the write is acknowledged and committed. This prevents loss of any in-flight write I/O if
a controller failure occurs. If a controller node failure occurs the cluster is put in journal mode. This
triggers all I/O normally written to the mirror cache to be written to a journal file instead. Client load
balancing in the FS7500 make this process transparent from a client point of view.
11.5
FS7600 components
The FS7600 and FS7610 NAS Appliance consist of two hot pluggable redundant NAS controllers per
appliance, redundant hot pluggable power supplies, and fans. Each NAS controller contains a built-in
battery backup to protect cache in case of controller power failure. The system components are
shown in Figure 35 and Figure 36.
Figure 35 Rear view of the FS7600 NAS appliance
March 2013
Dell EqualLogic Configuration Guide v14.1
11-111
11.6
FS7610 components
FS7610 hot pluggable NAS
Controller 1
FS7610 hot pluggable
NAS Controller 2
Figure 36 FS7610 NAS Appliance
March 2013
Dell EqualLogic Configuration Guide v14.1
11-112
12
FS Series file level operations
In this section we provide an overview of key FS Series Appliance features along with some operational
limits. Please refer to the NAS Service Management Tasks section of the EqualLogic PS Series Group
Administration Manual (version 5.1 or later) for detailed descriptions and all administrative task
procedures.
12.1
NAS cluster
The NAS cluster is the logical run-time container in which one or more NAS containers are created.
From the point of view of NAS clients, the NAS cluster is a virtual file server that hosts multiple CIFS
shares or NFS exports. You can only have one NAS Cluster per EqualLogic group. NAS clients connect
to file storage through a single NAS client virtual IP address. Client connections are load balanced
across the available NAS controllers in the cluster. When you configure a NAS cluster, you specify the
network configuration for the cluster and the amount of storage pool space consumed by the NAS
Reserve.
12.2
NAS reserve
The NAS reserve is the storage space allocated to the NAS Cluster from an EqualLogic PS Series
storage pool. The NAS reserve has the following properties:
•
•
•
•
•
It resides within a single EqualLogic storage pool. Only one storage pool in the EqualLogic
storage group can be used for allocating the NAS reserve.
It is formatted with the Dell FluidFS.
A fixed amount of storage space in the pool (512GB per controller node pair) is consumed by
NAS cluster metadata. You should add 512GB per controller node pair to the calculated size of
the NAS Reserve to compensate for this.
The NAS reserve can be expanded later when more storage pool space is available.
At the array level, the NAS reserve is comprised of a variable number of automatically created
volumes (from a minimum of two volumes up to a maximum of 34). The actual number of
storage volumes created within the pool depends on the reserve size setting.
Note: You cannot decrease the size of a NAS Reserve once it has been created.
12.2.1
Relationship between PS Series groups, pools and NAS reserve
The relationships are illustrated in Figure 37. As shown in the figure, the NAS Reserve pool can exist in
a group hosting a single pool, or in a group hosting up to three other pools simultaneously.
March 2013
Dell EqualLogic Configuration Guide v14.1
12-113
Figure 37 Containment Model: PS Series Storage and FS Series NAS Reserve
The addition of the FS Series NAS appliance to a PS Series group does not change the functional
behavior of PS Series groups and pools. PS Series groups and pools are explained in more detail in
sections 5.1 and 5.1.1.
12.3
NAS Container
To provision NAS storage, you need to create NAS containers within the NAS cluster. Inside a NAS
Container you can create multiple CIFS shares and/or NFS exports. Access to all shares and exports is
through a single client virtual IP address.
You can create a single large container, or you can create many separate containers. You can apply
different backup, snapshot, security, and quota policies to each container. Creating multiple
containers gives you the flexibility to apply different management policies to different containers. The
number and size of the containers in a NAS cluster depends on the storage needs of your NAS clients
and applications. You can increase and decrease the size of containers as needed. The relationship
between the NAS Cluster, NAS Reserve, NAS Containers, CIFS shares and NFS exports is shown in
Figure 38 below. Note that you can setup “mixed protocol” access to portions of a container. This
means that the same portion of the container can be simultaneously mounted by NFS clients and
mapped by CIFs clients.
March 2013
Dell EqualLogic Configuration Guide v14.1
12-114
Figure 38 Containment Model: FS Series NAS Reserve, Containers, Shares and Exports
12.3.1
NAS Container security styles
There are three security style options that can be applied to a NAS Container:
•
•
•
UNIX: Controls file access using UNIX permissions in all protocols. A client can change
permissions only by using the chmod and chown commands on the NFS mount point. You can
also specify the UNIX permissions for files and directories created in the container by Window
clients. See About UNIX Permissions for Windows Directories and Files. With UNIX file security
style, Windows clients cannot change any file or directory permissions. Read and write access is
controlled by the UNIX permissions for Windows files and directories, which you set in Group
Manager.
NTFS: The default security style for NAS containers is NTFS. Controls file access by Windows
permissions in all protocols. A client can change the permission and ownership by using the
Windows Security tab. With NTFS file security style, all access permissions are controlled by the
Windows administrator by using access control lists or share level permissions.
Mixed: Supports both NTFS and UNIX security styles. The permissions and ownership for a file or
directory will be the last one set. Permissions and access rights from one protocol to another are
automatically translated. Mixed is a combination of NTFS and UNIX security style. A Windows user
can override UNIX user settings, and a UNIX user can override Windows user settings. This security
style is only recommend for scratch storage where maintaining permissions is not important.
The default security style for NAS containers is NTFS. Referring to File System 1 in Figure 38, you could
assign any of the three container security styles to it. Given that portions of it are simultaneously
March 2013
Dell EqualLogic Configuration Guide v14.1
12-115
accessible as CIFs shares and NFS exports, a general rule of thumb for how to assign the security style
would be as follows:
•
•
•
If your users are predominantly Linux/UNIX based, use the UNIX style.
Likewise, if your users are predominantly Windows/CIFS based, use the NTFS style.
If you have mixed clients, use the style applicable to majority of users. Then create a user mapping
of your Windows to Linux users or vice versa. The user permissions are equivalent to the mapped
user. Note that the Mixed Mode container security style should not be used for cases where you
have a mix of both CIFS and NFS clients.
12.4
NAS Container snapshots
Dell FluidFS snapshots are created and managed by the FS Series appliance to provide a container level
snapshot capability. They are read-only and are created using a redirect-on-write method. This
method is also referred to as allocate-on-write. This approach requires only one I/O operation and
delivers higher write performance.
Note: FS Series FluidFS snapshots and PS Series volume based snapshots function
independently and have no impact on each other.
Please see the following whitepaper for more information on Dell FluidFS snapshot
behavior:
Dell Fluid File System:
http://www.dellstorage.com/WorkArea/DownloadAsset.aspx?id=1578
Additional FluidFS snapshot limits and considerations:
•
•
•
•
•
Each NAS Container has its own snapshot policy.
The minimum unit a snapshot can act on is a NAS Container.
The maximum number of snapshots you can retain per container is 512. The maximum
snapshots retained per FS appliance cluster is 10,000.
When you create a snapshot, all shares and exports within that container are included. If a
particular share or export will require frequent snapshots, you should group it with others that
have the same requirement, or dedicate a container to that particular share or export.
The FluidFS snapshot implementation allows for end users to individually select and restore
previous versions of files. The .snapshot directory contains all the snapshots taken for the
container. Browse to the correct snapshot directory to access the files needed to be restored
to the share. Windows users can also use the “Previous Version” feature of Windows files and
directories to restore individual files or entire folders.
Sizing limits and space utilization considerations for FluidFS snapshots:
•
•
•
The size of the snapshot reserve is specified as a percent of the container size. It is set at
container creation and changes according to the % reserve setting whenever the size of the
container is changed.
The default snapshot reserve size is 50%. It is common to select a reserve size in the range of
25%. The maximum snapshot reserve size is 90% of the FS size.
The snapshot reserve capacity is counted against the NAS container reserve space. However,
the snapshot reserve space is not enforced. Container data can fill the snapshot reserve space.
March 2013
Dell EqualLogic Configuration Guide v14.1
12-116
•
•
•
Snapshot reserve space utilization is a function of the data change rate in the container.
Old snapshots are deleted to make room for new snapshots if enough snapshot reserve is not
available.
NDMP based backup automatically creates a snapshot, from which the backup is created.
12.5
NAS Snapshots and Replication
The FS76X0 further ensures business continuity through support for file system, point-in-time
snapshots and snapshot-based, asynchronous replication. It is important to note that Group Manager
block-level replication cannot be used as a data protection strategy for FS76X0 data.
12.5.1
Snapshots
Fluid File System leverages redirect-on-write snapshots that only require one I/O per write operation
and thus avoid the performance degradation of the more traditional copy-on- write approach. Fluid
File System freezes an image and maintains a point-in-time backup of the data. Each individual NAS
container can have its own snapshot policies. The snapshots can be recovered instantly by
administrators and are also available to end-users through the NAS protocols as read-only views.
12.5.2
Replication
The FS76X0 and FS7500 (using FluidFS series firmware 2.0) now support snapshot-based,
asynchronous NAS replication to or from another FS7500 or FS76X0 system in a different EqualLogic
Group at a peer (local or remote) location. An EqualLogic Group containing FS appliances forms a
partnership with another EqualLogic group to replicate data across the two groups.
The NAS replication uses snapshots to replicate only the changed blocks across sites. NAS replication
works at the NAS container level and can be scheduled (hourly, daily, weekly, etc.) per NAS container.
The replication site must have the free capacity in the NAS reserve to replicate the NAS containers
chosen for replication. All NAS container snapshots on the primary are also replicated to the
secondary site.
Note: NAS replication uses the SAN side network to replicate file data. Refer to the NAS
Network Setup Guide for details on making you’re the customer network is ready to
support NAS replication.
March 2013
Dell EqualLogic Configuration Guide v14.1
12-117
Figure 39 NAS replication
Replication requirements:
•
•
•
•
•
Both sites must be EqualLogic FS appliances. Replication from FS7600 to any other Dell product
using FluidFS is not supported.
Each site (or EqualLogic Group) must have the same number of NAS controllers or NAS appliances.
FS appliances must run firmware version 2.0 or higher and PS arrays must run version 6.0 or
higher.
There must be a network link between the two sites (or EqualLogic groups) via the SAN side
network.
If the client and SAN networks do not have a route to each other, a static route needs to be
configured on FS NAS appliances on both the sites. Refer to NAS Replication Network Setup Guide
for details on configuring network for NAS replication.
Important notes:
•
•
•
•
•
•
•
You can replicate between FS7500 and FS76x0 as long as all FS appliances are running version 2.0
of FS Series Firmware (FluidFS).
Replication of NAS containers within the same EqualLogic group or NAS cluster is not supported
by FS appliances.
One EqualLogic group can form replication partnerships with 16 other groups; however one NAS
container in a group can be replicated to only a single replication partner.
NAS replication operates at block level. After the first complete data replication, only changed
blocks of a file and not the whole file (if any changes are made to a file) is replicated.
There is no Manual Transfer Utility (MTU) available for NAS replication.
The NAS replica is consistent to the last successful replication. If a you want multiple point in time
copies of the NAS container at the replica site, snapshots should be created on the primary side.
All the snapshots and the NAS container point in time copies will be available at the secondary site.
The NAS replication does not use any replication reserve space on the pool.
Accessing data at the replication site:
•
Read Only Access: A NAS replica can be made available to clients on secondary site by enabling
Read Only access
March 2013
Dell EqualLogic Configuration Guide v14.1
12-118
•
•
Temporary Promotion: Promote Read/Write with an ability to demote. However all writes are lost
when container is demoted to resume replication from primary site.
Permanent Promote: Promote Read/Write in case primary site or container is not available and
clients need to be failed over to secondary site. All writes are preserved
March 2013
Dell EqualLogic Configuration Guide v14.1
12-119
13
FS Series NAS Configuration
In Section 11.3 we presented a high level view of the FS7500 NAS component architecture (See Figure
34). In this section we provide detailed connection diagrams demonstrating how to setup fully
connected iSCSI SAN and client LAN connection paths for the FS7500 and FS7600/FS7610 appliances.
Note: It is recommended to keep the client and SAN side networks physically separate
and deploy two switches on both sides to provide redundancy in the event of a switch
failure.
13.1
FS7500 connection paths
The FS7500 appliance is comprised of two peer system controller nodes. In the required fully cabled
configuration, each controller node requires thirteen separate Ethernet cable connections. Thus a
single FS7500 appliance comprised of two controller nodes requires a total of 26 Ethernet
connections (four connecting to the client LAN switches and nine connecting to the iSCSI SAN fabric
switches.)
Figure 40 shows the client LAN connection paths.
Note: While it is possible to operate an FS7500 appliance in a partially cabled
configuration, this configuration is not supported by Dell. You should use a fully cabled
configuration in a production environment. You will also need to provision the required
switch port count on the iSCSI SAN and client LAN sides of the system to support a fully
connected configuration.
March 2013
Dell EqualLogic Configuration Guide v14.1
13-120
Figure 40 Connection Paths for FS7500 Client LAN
Figure 41 below shows the iSCSI SAN, IPMI, and node interconnect paths. Pay careful attention to how
the controller ports alternate between redundant switch paths.
Note: With the exception of the IPMI connection paths, corresponding ports on each
controller node must connect to the same SAN switch. This connection pattern is shown
in Figure 41.
Sizing the iSCSI SAN Inter-Switch Connection for FS7500
The inter-switch connection sizing guidelines provided in Section 7.2.2 also apply to FS7500 SAN
design. The FS7500 mirrors the write cache between controller nodes. To accomplish this, all write
operations are transmitted across the controller interconnect. Thus, it is very important that you follow
connection pattern shown in Figure 41 to ensure corresponding ports are connected to the same
switch. This connection pattern prevents controller interconnect traffic from crossing the inter-switch
connection.
March 2013
Dell EqualLogic Configuration Guide v14.1
13-121
Figure 41 Connection Paths for FS7500 iSCSI SAN, IPMI and Controller Interconnect
13.2
FS7600/7610 connection paths
The Dell EqualLogic NAS appliances require the following networks:
•
•
Client network: Used for client access to the NFS exports and CIFS shares hosted by the NAS
cluster.
SAN/internal network: Used for internal communication between the controllers and
communication between the controllers and the EqualLogic PS Series SAN. The SAN and Internal
networks use the same set of switches.
March 2013
Dell EqualLogic Configuration Guide v14.1
13-122
It is recommended to keep the client and SAN side networks physically separate and deploy two
switches on both sides to protect against a switch failure.
Figure 42 FS7600 AND FS7610 networks
See figures below for network connections.
Figure 43 FS7600 network
March 2013
Dell EqualLogic Configuration Guide v14.1
13-123
Figure 44 FS7610 network
March 2013
Dell EqualLogic Configuration Guide v14.1
13-124
Installation/Expansion
•
•
•
If installing FS7500/FS76x0 into an existing EqualLogic SAN, verify the existing customer network
meets the minimum requirements. Refer to FS76x0 installation guide for more information on
network requirements. Early validation helps avoid issues during and after the install.
FS7500/FS76x0 installation service is mandatory.
All NAS appliances in a NAS cluster need to have either 1Gb or 10Gb connectivity. Appliances with
different connectivity cannot be mixed in a NAS cluster.
•
•
FS7500 cluster can be expanded by using an FS7500 or FS7600 and NOT FS7610.
There is no way to upgrade FS7500 to 10Gb Ethernet by replacing the network controllers or the
NAS appliance. 10Gb functionality is provided by FS7610.
Network
•
The two iDRAC IPMI ports on FS7500 are 100Mb. The SAN switches must be able to support
100Mb in addition to 1Gbps speed. Most 10Gb SFP+ switches (including 8024F and F10 S48) do
not support 100Mb.
•
•
It is best to connect FS7500 to 1 Gb switches and then uplink 1Gb switches to 10Gb switches
OR connect 100Mb iDRAC ports to 1Gb switches (that are connected to 10Gb switches) and
FS7500 SAN/Internal to 10Gb switch ports that can run at 1Gb speed.
FS appliances do not support client VIPs in multiple subnets and use of netgroups.
Data Protection
•
•
FS76x0/FS7500 does not support replication to local NAS cluster
EqualLogic block replication/clones/snapshots cannot be used to protect data managed by FS
appliances. Use NAS replication to protect file data.
March 2013
Dell EqualLogic Configuration Guide v14.1
13-125
14
Data Center Bridging (DCB)
The enhancement to the Ethernet Specifications (IEEE 802.3 specifications) called Data Center
Bridging (DCB) enables bandwidth allocation and lossless behavior for storage traffic when the same
physical network infrastructure is shared between storage and other traffic.
The network is the fundamental resource that connects the assorted devices together to form the
datacenter Ethernet infrastructure. These devices include the server hardware (along with the
operating system and the applications that run on the host) and the storage systems that host
application data. Sharing this Ethernet infrastructure with multiple traffic types (LAN and SAN) requires
a fairness mechanism to provide bandwidth allocation and flow control for each type of traffic.
Without such a mechanism, the Local Area Network (LAN) traffic and the Storage Area Network (SAN)
traffic would have to be separated onto their own dedicated networks.
When the SAN and LAN networks are shared, all traffic is equal, unless Quality of Service (QoS) and/or
Class of Service (CoS) is used.
Traditional or Non-DCB QoS ( IEEE 802.1p) is not supported for EqualLogic implementations.
In a shared network environment, LAN and SAN traffic can impact each other and QoS may not solve
this because of QoS implementation differences with vendors. These different implementations
include:
•
•
•
Number of Queues managed
Relative priorities between queues
Bandwidth reserved per queue
Also traditional QoS lacks selective flow control for each traffic type. Regular IEEE 802.3 PAUSE will
pause the entire link and not selectively pause LAN or SAN traffic. This flow control ability is important
for reducing congestion on switches and to enable fair sharing of resources between traffic types.
Other methods of network sharing:
•
•
VLANs offer port-level security and segregation, but do not provide guaranteed bandwidth
or quality of service.
NIC partitioning (NPAR) manages traffic within the host. Once the network data exits the
NIC to the switch, any QoS/bandwidth management enforced by NPAR is not honored
by the switch.
Note: It is not recommended to share network infrastructure without DCB. DCB is the
only recommended method of converging SAN and LAN in an EqualLogic Storage
environment.
14.1
DCB Overview
DCB is a collection of Standards designed to improve QoS networking and management in the Data
Center that enables iSCSI SANs or FCoE SANs or both to converge with regular server LAN traffic on
March 2013
Dell EqualLogic Configuration Guide v14.1
14-126
the same physical infrastructure to increase operational efficiency, constrain costs, and ease network
management.
There are primarily three progressive versions of DCB:
•
•
•
Cisco, Intel, Nuova (CIN) DCBX
Converged Enhanced Ethernet (CEE) DCBX or baseline DCBX
Institute of Electrical and Electronic Engineers (IEEE) DCB.
DCB technologies based on standards include:
•
•
•
•
PFC – Priority based Flow Control (802.1Qbb)
ETS – Enhanced Transmission Selection (802.1Qaz)
CN – Congestion Notification (802.1Qau)
DCBx – Data Center Bridging Capability eXchange
• Support for iSCSI application protocol priority with DCBX (also known as iSCSI TLV)
Note: DCB uses 10Gb and higher Ethernet only.
I/O convergence using DCB for Dell EqualLogic iSCSI storage is the future direction for “Converged
iSCSI” in a lossless Ethernet environment.
Without DCB, it is typical to dedicate a separate physical network infrastructure for SAN traffic to
guarantee the infrastructure for bandwidth and performance as shown in the figure below. For iSCSI, it
is a best practice to dedicate a SAN comprised of Ethernet switches for server and storage interconnection. In this case, the server and storage systems are configured with dedicated NICs to
communicate on the iSCSI SAN.
March 2013
Dell EqualLogic Configuration Guide v14.1
14-127
Figure 45
Physically separate dedicated network infrastructures
A converged network includes carrying both SAN and other network traffic such as server LAN on a
single network infrastructure as shown in Figure 46. iSCSI can be converged with non-storage based
server LAN traffic, allowing the network ports and the inter-connecting links to carry multiple traffic
types or protocols. A DCB enabled converged Ethernet infrastructure includes the NIC/CNA on the
end-devices (servers and storage arrays) along with the switching infrastructure.
March 2013
Dell EqualLogic Configuration Guide v14.1
14-128
Figure 46
14.2
DCB enabled converged network infrastructure
DCB requirements for EqualLogic
It is required that all devices in the EqualLogic SAN support DCB for iSCSI when this functionality is
enabled. If any device in the SAN does not support DCB, then DCB needs to be disabled at the
switches for the entire SAN. Once all devices in the SAN are DCB compliant, then DCB can be reenabled. Switches and server CNAs/NICs that are designated as DCB Supported in the EqualLogic
Compatibility Matrix have been fully validated by Dell to ensure compatibility for EqualLogic SANs.
March 2013
Dell EqualLogic Configuration Guide v14.1
14-129
The minimum switch and server CNA/NIC requirements to support an end-to-end DCB solution with
EqualLogic are:
•
•
•
•
Data Center Bridging Exchange (DCBx) -DCB protocol that performs discovery, configuration,
and mismatch resolution using Link Layer Discovery Protocol (LLDP )
Application Priority ( iSCSI TLV ) - Switches must support configuration of a priority value for
iSCSI protocol and advertisement to peer ports. Server NICs/CNAs: Must support iSCSI
protocol in application priority (learned from the switch) and must support tagging outgoing
iSCSI frames with this priority.
ETS - Requires a dedicated traffic class or priority group for iSCSI priority
PFC - Requires enabling PFC (no drop or lossless behavior) for iSCSI priority
• Switches: Configure dedicated TC/PG for iSCSI priority with allocated bandwidth and PFC
enabled for iSCSI priority
• Server NICs/CNAs: Adhere to TC/PG mapping for iSCSI priority and PFC for iSCSI priority
(learned from the switch)
Designing a converged network deployment with components that have no DCB support or partial
DCB support is not recommended for end-to-end converged I/O. Configurations with components
that have no DCB support or partial DCB support may result in loss of expected functionality or may
be functional without meeting the converged network objectives.
Note: It is not recommended to share network infrastructure with DCB and Non-DCB
components.
For a complete list of components that support the DCB standards, see the: EqualLogic Compatibility
Matrix
Additionally, there is a partner verified support program (PVSP) between Dell and VMWare.
This agreement is to support Internet SCSI (iSCSI) over Data Center Bridging (DCB) Solution
for ESX in a DCB environment with EqualLogic storage.
VMWare PVSP on VMWare website: EqualLogic SAN over DCB Solution for VMware ESXi 5.1
Note: It is important to verify that all components in the SAN are listed in the EqualLogic
Compatibility Matrix as DCB Supported, or that the components support all EqualLogic
requirements for DCB.
14.3
Methods for configuring DCB
Configuration propagation: If this mechanism is supported by the switch model, then peer switches,
downstream switches, and I/O Aggregator modules should accept the DCBX configuration from the
source switches through certain Inter Switch Link (ISL) ports set as upstream ports capable of receiving
DCBX configuration from peers (enabled with willing mode turned on the ports).
Manual configuration: If the internal propagation method is not supported by your switch model,
then peer and intermediate layer switches must be manually configured with the same DCB
parameters as the source switches with willing mode off across all ports.
March 2013
Dell EqualLogic Configuration Guide v14.1
14-130
The recommended operational mode is that switches use non-willing DCBx mode, while server
NIC/CNAs and Storage ports operate in willing mode. This is the default behavior on most switches
and the NICs/CNAs.
The DCB parameters (ETS, PFC, iSCSI application priority) are then configured in the switch devices
and learned by the end-devices.
14.4
Basic Deployment Topology Example
This topology is an example topology of a single layer switching with rack servers.
Figure 47 Single Layer Switching example with rack servers
PowerConnect 8132F supports the auto-up port role for learning DCBX configuration and the autodown port role for internally propagating to a downstream device. Similary, Force10 S4810 supports
the auto-upstream and auto-downstream port roles.
14.5
Tested SAN designs
This section shows three different end-to-end DCB SAN designs that were tested using the M1000e
blade chassis. The topologies are tested deployments that can be used with EqualLogic
storage arrays.
There are three categories of SAN designs for M1000e blade chassis integration:
•
•
Blade IOM switch only
ToR switch only
March 2013
Dell EqualLogic Configuration Guide v14.1
14-131
•
Blade IOM switch with ToR switch.
14.5.1
Blade IOM switch only
Network ports of both the hosts and storage are connected to the M1000e blade IOM switches. No
ToR switches are required. The switch interconnect can be a stack or a LAG, and no uplink is required.
Figure 48 Blade IOM switch only
14.5.2
ToR switch only
Network ports of both the hosts and the storage are connected to external ToR switches. 10 GbE
pass-through IOM switches are used in place of blade IOM switches in the M1000e blade chassis. The
switch interconnect can be a stack, a LAG, or a VLTi.
March 2013
Dell EqualLogic Configuration Guide v14.1
14-132
Figure 49 ToR switch only
14.5.3
Blade IOM switch with ToR switch
Host network ports are connected to the M1000e blade IOM switches and the storage network ports
are connected to ToR switches. The switch interconnect can be a stack, a LAG, or a VLTi and should
March 2013
Dell EqualLogic Configuration Guide v14.1
14-133
connect the ToR switch to better facilitate inter-array member traffic. An uplink stack, LAG, or VLT LAG
from the blade IOM switch tier to the ToR switch tier is also required.
Figure 50 Blade IOM switch with ToR switch (Two tier design)
14.6
Data Center In A Chassis DCB design
Find guidance for incorporating a data center-in-a-chassis with the new Dell EqualLogic PS-M4110
Blade Array, Force10 Blade IO Modules, and PowerEdge M-Series blade servers to form a reliable,
stable, and well performing full end-to-end DCB solution.
March 2013
Dell EqualLogic Configuration Guide v14.1
14-134
Figure 51 M1000e Blade Enclosure full DCB solution
The components used in the solution are listed here:
•
•
•
•
•
M1000e chassis
M620 blade server
PS-M4110XV array
Force10 MXL (2)
Broadcom 57810 CNA
14.7
VLANs for iSCSI
A non-default VLAN is required for operating prioritized lossless iSCSI traffic in a DCB enabled Ethernet
infrastructure. Switch ports that are based on the IEEE 802.1Q VLAN specification forward frames in
the default or native VLAN without tags (untagged frames). Normally these ports also receive frames in
the default or native VLAN as untagged. Such a port is typically termed a “tagged port” or “trunk port”
and all non-default VLAN frames are forwarded with the tags intact. Since the DCB priority information
(PCP value) is encoded in the VLAN tag, this information will be lost if an end-device sends iSCSI traffic
in a default or native VLAN and the switch receives it in a tagged or trunk port. As a result, DCB
prioritization and traffic classification will be lost.
All devices in the iSCSI data path must have the same VLAN ID configured on the respective ports
participating in the iSCSI network to ensure proper functioning. These devices include the server iSCSI
NIC/CNA ports, EqualLogic arrays, and all switches on the iSCSI SAN.
Note: The VLAN ID for iSCSI can be set in the EqualLogic Group Manager interface or the
storage array CLI. When DCB is enabled on the switch, it is necessary to configure a nondefault VLAN on the array, switch, and all host ports that are part of the EqualLogic SAN.
VLAN IDs 0 and 1 are not recommended, because these may be the default or reserved
VLAN for some switches, and as such, may forward frames untagged (e.g. no VLAN
tagging). VLAN tagging is required to fully support DCB.
March 2013
Dell EqualLogic Configuration Guide v14.1
14-135
For more information on the DCB requirements and configuration guidelines, see the following white
paper: Data Center Bridging: Standards, Behavioral Requirements, and Configuration Guidelines with
Dell EqualLogic iSCSI SANs
For more information on the DCB requirements for EqualLogic, to ensure that DCB is properly enabled
and/or disabled across all devices, and to assist with identifying and resolving basic DCB configuration
issues in the EqualLogic SAN, see the following white paper: EqualLogic DCB Configuration Best
Practices
For a comprehensive understanding of integrating an M1000e Blade Chassis in a Full End-to-End Data
Center Bridging environment, see the following white paper: M1000e Blade Enclosure and EqualLogic
Arrays SAN Design using Force10 Switches
March 2013
Dell EqualLogic Configuration Guide v14.1
14-136
Appendix A Network ports and protocols
PS Series groups use a number of TCP and UDP protocols for group management, I/O operations, and
internal communication. If you have switches or routers set to block these protocols, you may need to
unblock them to allow management or I/O operations to work correctly. The required and optional
protocols are listed in the following sections.
A.1 Required ports and protocols
Table 49 lists the ports and protocols required for operating an EqualLogic iSCSI SAN.
Table 49 Required ports and protocols
Type
Port
Protocol
Access
iSCSI
TCP
3260
iSCSI
To the group IP address and all individual member IP
addresses
EqualLogic Internal
UDP
161
SNMP
Management operations
TCP
9876
Internal
iSCSI intra-system control
TCP
25555
Internal
Group communication
TCP
20002
Internal
Event logging
A.2 Optional ports and protocols
Table 50 lists the optional ports and protocols used for management and alerts. They are not required
for correct array operation.
March 2013
Dell EqualLogic Configuration Guide v14.1
14-137
Table 50 Optional ports and protocols
Type
Port
Protocol
Access
CLI Management
TCP
23
Telnet
To group IP address
TCP
22
SSH
To group IP address
Web Based Management
TCP
80
HTTP
To group IP address
TCP
3002
GUI communication
To group IP address
TCP
3003
GUI communication
(encrypted)
To group IP address
SNMP
UDP
161
SNMP
To and from group IP address
Syslog
UDP
514
Syslog
From group IP address
EqualLogic Diagnostics
March 2013
TCP
20
FTP
Software update and diagnostic procedures; to all
individual member IP addresses
TCP
25
SMTP
E-mail and diagnostic notifications; from all individual
member IP addresses to the configured SMTP server
Dell EqualLogic Configuration Guide v14.1
14-138
Appendix B Recommended switches
The list of recommended switches is now maintained in a separate document.
•
EqualLogic Compatibility Matrix
http://en.community.dell.com/dellgroups/dtcmedia/m/mediagallery/19856862/download.aspx
March 2013
Dell EqualLogic Configuration Guide v14.1
14-139
Appendix C Supported iSCSI initiators
The list of supported iSCSI initiators is now maintained in a separate document.
•
EqualLogic Compatibility Matrix
http://en.community.dell.com/dellgroups/dtcmedia/m/mediagallery/19856862/download.aspx
March 2013
Dell EqualLogic Configuration Guide v14.1
14-140
Appendix D Upgrade paths for EqualLogic PS
SeriesArrays
Table 51 EqualLogic Upgrade Paths
End of sales life arrays:
Latest available
conversion model
1Gb to 10Gb
conversion availability
Drive upgrades
availability
PS-50 thru PS3000
None
None
None
PS4000
Yes, PS6000 – Dellstar
Yes, PS6010 – Dellstar
Yes – cus kit tool
PS5000
Yes, PS6000 – Dellstar
Yes, PS6010 – Dellstar
Yes – cus kit tool
PS6000
None
Yes, PS6010 – Dellstar
Yes – cus kit tool
PS5500
Yes, PS6500 – Dellstar
Yes, PS6510 – Dellstar
Yes – cus kit tool
PS6010
None
Currently shipping
arrays:
Latest available
conversion model
1Gb to 10Gb
conversion availability
Drive upgrades
availability
PS6500
None
Yes, PS6510 – Dellstar
Yes – cus kit tool
PS6510
None
PS4100
TBD
PS4110
TBD
PS6100
None
PS6110
None
Yes – cus kit tool
PS-M4110
None
Yes – cus kit tool
March 2013
Yes – cus kit tool
Yes – cus kit tool
TBD
Yes – cus kit tool
Yes – cus kit tool
TBD
Dell EqualLogic Configuration Guide v14.1
Yes – cus kit tool
14-141
Related publications
The following locations provide additional background and technical details supporting configuration
of EqualLogic SANs.
•
•
•
•
•
EqualLogic Compatibility Matrix
http://en.community.dell.com/dellgroups/dtcmedia/m/mediagallery/19856862/download.aspx
EqualLogic Technical Content
http://en.community.dell.com/techcenter/storage/w/wiki/2660.equallogic-technicalcontent.aspx
Rapid EqualLogic Configuration Portal
http://en.community.dell.com/techcenter/storage/w/wiki/3615.rapid-equallogicconfiguration-portal-by-sis.aspx
Switch Configuration Guides
http://en.community.dell.com/techcenter/storage/w/wiki/4250.switch-configuration-guidesby-sis.aspx
Storage Infrastructure and Solutions Team Publications
http://en.community.dell.com/techcenter/storage/w/wiki/2632.storage-infrastructure-andsolutions-team-publications.aspx
March 2013
Dell EqualLogic Configuration Guide v14.1
14-142
THIS WHITE PAPER IS FOR INFORMATIONAL PURPOSES ONLY, AND MAY CONTAIN
TYPOGRAPHICAL ERRORS AND TECHNICAL INACCURACIES. THE CONTENT IS PROVIDED AS
IS, WITHOUT EXPRESS OR IMPLIED WARRANTIES OF ANY KIND.

Similar documents

×

Report this document