HP P2000 Array System Remote Snap SoftwareTechnical white
paper
Table of contents Introduction
..................................................................................................................................3
HP P2000 Array System Remote Snap Software product overview
..................................................3 Whats new?
...............................................................................................................................3
CHAP settings and Remote Snap
....................................................................................................7
Benefits of the HP P2000 Array System Remote Snap Software
..........................................................8
Disaster recovery
......................................................................................................................8
Backup
....................................................................................................................................9
Development
............................................................................................................................9
How the technology works
.............................................................................................................9
The HP P2000 Array System Remote Snap Software setup requirements
........................................11 Network requirements
.............................................................................................................11
The HP P2000 Array System Remote Snap Software basic functions
................................................12 Adding a remote
system
..........................................................................................................13
Creating a replication set
.........................................................................................................13
Deleting a replication set
.........................................................................................................13
Exporting a replication image
..................................................................................................13
Setting the primary volume
.......................................................................................................13
Verifying the remote link
..........................................................................................................13
Types of replications
...................................................................................................................14
Local replication
.....................................................................................................................14
Remote replication
..................................................................................................................18
Disaster recovery operations
....................................................................................................19
Use cases
..................................................................................................................................20
Single office with a remote site for backup and disaster recovery
using iSCSI to replicate data ........21 Single office with a local
site disaster recovery and backup using iSCSI
........................................23 Single office with a
local site disaster recovery and backup using FC
............................................24 Two branch offices
with disaster recovery and backup
................................................................25
Single office with a target model using FC and iSCSI ports
..........................................................26
Multiple local offices with a centralized backup
..........................................................................27
Replication of application-consistent snapshots
...........................................................................28
Replication of the Microsoft VSS-based application-consistent
snapshots ........................................30 Best practices
.............................................................................................................................33
Fault tolerance
........................................................................................................................33
Volume size and
policy............................................................................................................33
License
..................................................................................................................................34
Scheduling
.............................................................................................................................34
Physical media transfer
............................................................................................................35
Replication setup wizard
..........................................................................................................35
Application-consistent snapshots
...............................................................................................35
Max volume limits
...................................................................................................................36
Monitoring
.............................................................................................................................36
Performance tips
.....................................................................................................................37
Troubleshooting
..........................................................................................................................37
FAQs
........................................................................................................................................38
Summary
...................................................................................................................................39
Glossary
....................................................................................................................................39
For more
information...................................................................................................................41
IntroductionThis document provides information for using the HP
P2000 Array System Remote Snap Software. The following topics are
covered:
Benefits of the P2000 Array System Remote Snap Software How the
HP P2000 Array System Remote Snap Software technology works HP
P2000 Array System Remote Snap Software basic functions Types of
replication Use cases Best practices Troubleshooting
HP P2000 Array System Remote Snap Software product overviewThe
HP P2000 Array System Remote Snap Software has array-based
functionality that provides remote replication on the HP P2000 G3
Array. The following array controllers support the HP P2000 Array
System Remote Snap Software:
HP StorageWorks P2000 G3 FC Modular Smart Array Controller HP
StorageWorks P2000 G3 FC/iSCSI Combo Modular Smart Array Controller
HP P2000 G3 10GbE iSCSI Modular Smart Array System HP P2000 G3
iSCSI Modular Smart Array SystemNote: We do not support the remote
snap feature on HP P2000 G3 SAS Modular Smart Array System
The HP P2000 Array System Remote Snap Software is a form of
asynchronous replication that replicates block-level data from a
volume on a local system to a volume on the same system or on a
second independent system. The second system may be at the same
location as the first, or it may be located at a remote site.
Whats new?This section describes new enhancements and support
added to the HP P2000 G3 Modular Smart Array System with the
release of the new TS230 firmware.
Remote Snap support for iSCSI controllers Concurrent Remote Snap
and I/O with iSCSI controllers A new tool to check local system
link Replication-set-connected port fieldRemote snap support for
iSCSI controllers The controllers listed below now support the
Remote Snap feature with TS230 firmware.
HP P2000 G3 iSCSI Modular Smart Array Dual Controller HP P2000
G3 10GbE iSCSI Modular Smart Array Dual Controller
3
Concurrent Remote Snap and I/O with iSCSI controllers You can
use iSCSI host ports for both Remote Snap and I/O with any HP P2000
G3 iSCSI arrays running TS230 firmware. We do not support both
remote snap and I/O at same time using iSCSI host ports with any
pre-T230 firmware. A new tool to check local system link You can
check the connectivity of host ports between controllers in the
local system. When you check links, the panel may show this
information for each linked host port in both controllers. You can
access this tool from the SMU (System Name (P2000 G3 iSCSI) >
Tools > Check Local System Link) and the CLI (run command verify
links)
It will check links from controller A ports to controller B
ports irrespective of where you run the command from(e.g.
controller A or B).
In the CLI, you can use the same command to check remote system
links for the replication purposes; this tests thelinks to be used
for replication from one system to another system.
In the Replication Wizard, you can also enable a local link
check by selecting the check box.Screenshots below explain the
different ways of using the tool (SMU, CLI etc). You can choose any
one of these methods to check the connectivity. Note: Verify links
can be used to check links to the remote system. Figure 1 shows an
example of this command being used to check remote system link.
Figure 1: Check System Link Tool for remote system: CLI
# verify Port Type
links remote-system remote_sys_1 Links
--------------------------------------------A1 iSCSI
A1,A2,A3,A4,B1,B2,B3,B4 A2 iSCSI A1,A2,A3,A4,B1,B2,B3,B4 A3 iSCSI
A1,A2,A3,A4,B1,B2,B3,B4 A4 iSCSI A1,A2,A3,A4,B1,B2,B3,B4 B1 iSCSI
A1,A2,A3,A4,B1,B2,B3,B4 B2 iSCSI A1,A2,A3,A4,B1,B2,B3,B4 B3 iSCSI
A1,A2,A3,A4,B1,B2,B3,B4 B4 iSCSI A1,A2,A3,A4,B1,B2,B3,B4
--------------------------------------------Success: Command
completed successfully. #
Note: While using replication wizard, user can check the local
link by enabling the check links option. Figure 2 is the screenshot
of this step.
4
Figure 2: Check Local System in the Replication Wizard in the
SMU.
5
Note: The previous screen shot showed how to check the local
system links using the wizard. The following shows how to do the
same using the tool.
Figure 3: Check Local System Link Tool: SMU
Note: The verify links command is used to check local link.
Figure 4 shows the command line output of this command, and in this
case all ports are connected.
Figure 4: Check Local System Link tool: CLI
# verify links Port Type Links
--------------------------------------------A1 iSCSI B1,B2,B3,B4 A2
iSCSI B1,B2,B3,B4 A3 iSCSI B1,B2,B3,B4 A4 iSCSI B1,B2,B3,B4
--------------------------------------------Success: Command
completed successfully. #
Connected port field in replication set For a remote primary or
secondary volume, the Connects Ports field shows the IDs of up to
two ports from another array (replication involves two arrays) in
the local system that are connected to the remote system. If two
ports are connected but only one is shown, this could mean that a
problem is preventing half the available bandwidth from being
used.
6
Note: This field shows N/A for a local primary or secondary
volume.
Figure 5: Connected ports displaying ports A1 and A2 as being
used for replication
CHAP settings and Remote SnapIf you configure CHAP with Remote
Snap, you can use CHAP to authenticate iSCSI login requests between
the local system and a remote system:
Create a one-way CHAP record on each system. On the local
system, the CHAP record must refer to the nodename of the remote
system. On the remote system, the CHAP record must refer to the
node name of the local system. Both records must use the same
secret. Use the create chap-record command to create a CHAP
record:# create chap-record name
iqn.1991-05.com.microsoft:myhost.domain secret 0D12x
After the CHAP records are created, enable CHAP on the primary
system, the secondary system, or both. To enable CHAP, use the set
iscsi-parameters command: # set iscsi-parameters chap enabled
7
Table: CHAP settings and corresponding behavior with HP P2000
Array System Remote Snap Software. Local system CHAP Disabled
(Secret: No; CHAP record: No) CHAP Enabled (Secret: SECRET1; CHAP
record: Yes) CHAP Enabled (Secret: SECRET1; CHAP record: Yes) CHAP
Enabled (Secret: No; CHAP record: No) Remote system CHAP Disabled
(Secret: No; CHAP record: No) CHAP Enabled (Secret: SECRET1; CHAP
record: Yes) CHAP Enabled (Secret: SECRET2; CHAP record: Yes) CHAP
Enabled (Secret: No; CHAP record: No) Expected behavior Remote Snap
works fine. No iSCSI authentication Remote Snap works fine. Remote
Snap will fail. Use the same secret for both the local and remote
systems. Remote Snap will fail. Enabling CHAP without specifying a
secret for an iSCSI initiator effectively blocks that initiator.
Remote Snap works fine. Remote Snap will fail. Use the same secret
for both the local and remote systems.
CHAP Disabled (Secret: SECRET1; CHAP record: Yes) CHAP Disabled
(Secret: SECRET1; CHAP record: Yes)
CHAP Enabled (Secret: SECRET1; CHAP record: Yes) CHAP Enabled
(Secret: SECRET2; CHAP record: Yes)
Note: If you are performing a local replication involving iSCSI
ports, CHAP will not be used.
Benefits of the HP P2000 Array System Remote Snap SoftwareHP
P2000 Array System Remote Snap Software replication is an
HP-licensed feature for disaster recovery. It has been designed
with a robust, fault-tolerant design that allows replication to
continue in the event of some system failures involving
communication, controllers, ports, hard drives (depending on the
RAID configuration), or temporary power failure. The HP P2000 Array
System Remote Snap Software can employ either Ethernet or Fibre
Channel (FC) technology. The HP P2000 Array System Remote Snap
Software technology enables the following key data management and
protection capabilities:
Continuity of business systems in the event of a failure on the
primary site Access to data at a remote site, for either dispersed
operations, or development activities Multiple recovery points
using snapshots
Disaster recoveryThe HP P2000 Array System Remote Snap Software
provides access to data at a secondary site when the primary site
experiences a critical failure. The HP P2000 Array System Remote
Snap Software enables a maximum of 16 data volumes to be replicated
to a separate offsite location. Recovery time is reduced at the
primary location because data is still available at the secondary
site. The data stored at the secondary site can then be used to
restore the primary location once it is back online, or the data
can be exported and used by users at the secondary site. A remote
site should be set up to replicate the data at regular intervals.
When a disaster occurs at a local site, applications can switch to
the remote site. The data that was replicated from the local site
can still be accessed. Minimal downtime occurs while switching the
application to the remote site, and data is made available from the
last replication point. Setting up the remote site and replicating
data at regular intervals protects the data in the event of outages
at the local site.
8
BackupThe HP P2000 Array System Remote Snap Software can
replicate up to 16 volumes, with up to 126 replication snapshots
per volume and marginal impact on server performance. The software
can be used by small businesses as a primary backup tool and by
large businesses as a secondary backup tool at data centers. The HP
P2000 Array System Remote Snap Software can be used as interim
storage for backing up to removable media such as tape. The remote
site can be as a centralized backup location for branch offices.
Alternatively, remote offices can replicate to central data centers
where backups occur. The software reduces the overall backup time
by replicating only data that has been modified. Because the
software supports either FC or Ethernet interconnects, businesses
have the flexibility to use the technology that best matches their
current environment.
DevelopmentThe HP P2000 Array System Remote Snap Software
enables different development use cases:
An application administrator can test patches or changes in the
local system by switching the applications to the
remote site. Once the testing of the patch update is completed,
the administrator can switch the applications back to the local
site. database volumes by exporting the snapshots on the remote
system. When the exported snap shot is no longer needed, it can be
deleted.
A database application development team can have access to
regularly scheduled snapshots of the replicated
How the technology worksFunctionality of the P2000 family is
based on the existing snapshot technology offered by the HP P2000
G3 Modular Smart Array Systems. Snapshots are used to track the
data to be replicated by determining the differences in data
updated on the master volume, minimizing the amount of data to be
transferred. The HP P2000 Array System Remote Snap Software enables
snapshots of data to reside on another array at a location other
than the primary site. To perform a replication, the system takes a
snapshot of the volume to be replicated, creating a point-in-time
image of the data. The system replicates this point-in-time image
to the destination volume by copying only the differences in the
data between the current snapshot and the previous one via TCP/IP
(iSCSI) or FC. The snapshot occurs at the volume level and is
block-based. The software functions independently of the vdisk RAID
and drive configuration, so the secondary volume in a given set may
have different underlying RAID levels, drive counts, drive sizes or
drive types than the primary volume, though the volume sizes must
be identical. Since the software functions at the raw block-level,
it has no knowledge of the volumes operating system configuration,
the file system, or any data that exists on the volume. The HP
P2000 Array System Remote Snap Software uses a pull model for
replication. In a pull model, the secondary volume requests data
from the appropriate snapshot on the primary volume. The HP P2000
Array System Remote Snap Software replication process includes the
following steps: 1. Create a snapshot on the primary volume. This
snapshot is a component of a replication image. 2. The local array
sends a notification to the designated remote array that a
replication operation has been started for a given storage volume.
3. The remote array requests a difference list from the local
array. This list contains only the changed storage blocks of the
primary volume since the last replication command (previous sync
point snapshot). The difference list contains all storage blocks
from the first replication of the primary volume. 4. The local
array sends a difference list to the remote array. 5. The remote
array requests all blocks in the difference list from the local
array. 6. When the transfer is complete, the remote array creates a
new snapshot to track newly acquired storage blocks. The remote
array creates snapshots of the secondary volume for each
replication.
9
7. The HP P2000 Array System Remote Snap Software feature
repeats step 1 and queues steps 26 for each new replication command
issued to the same replication set until the prior replication
command is complete. As long as the sync points are maintained, new
replication commands to the same primary volume can be performed,
while one or more previously executed replication commands are
still in process. This enables the user to take snapshots at
discrete intervals without waiting for any previous replications to
complete.
Figure 6: The HP P2000 Array System Remote Snap Software
creation cycle
10
The HP P2000 Array System Remote Snap Software setup
requirementsTo set up the HP P2000 Array System Remote Snap
Software using the P2000 G3 FC Modular Smart Array System, observe
the following requirements:
Two dual-controller arrays with all ports configured are
required and must be connected via a switch: A direct connection
between systems is not supported.
The management ports on the two P2000 G3 arrays must be on the
same network when using the SMU to createreplication sets.
Remote Snap does not supported SAS. Refer to the product
overview section in the introduction (page 3).Note: For more
information on controller types and additional specifications, see
the HP P2000 G3 Modular Smart Array QuickSpecs.
The HP P2000 Array System Remote Snap Software license must be
enabled on the local and remote systems: To explore the HP P2000
Array System Remote Snap Software, enable the 60-day temporary
license. To permanently enable the HP P2000 Array System Remote
Snap Software, purchase a license.
The HP P2000 Array System Remote Snap Software supports up to 16
replication sets per array. If a volume on thesystem is
participating in a replication set, either as a primary volume or
as a secondary volume, it counts against the 16 replication set
limit.
For the combo controller array, configure all of the iSCSI ports
with valid IP addresses. At least one vdisk on the remote system is
required to create replication set. Creating replication sets with
the SMU requires adding the remote system to the local system.
Creating replicationsets with the CLI is easier if the remote
system is added, but is not necessary.
The primary and secondary volumes must be the same size.
Network requirementsThe following is a guideline for users
setting up iSCSI controllers with T230 firmware (1 Gb and 10 Gb
ports) to use remote snap feature. System or environment variables
Hardware type: 10 Gb or 1 Gb Priority (set from set
replication-volume-parameters priority): Low, Medium, or High
Number of concurrent inbound replications (Rp) (from the primary
systems view): User-configured Number of inbound channels (Cp)
(from the primary systems view): User-configured Number of
concurrent outbound replications (Rs) (from the secondary systems
view): User-configured Number of outbound channels (Cs) (from the
secondary systems view): User-configured Packet loss rate (PL): You
may get this from a switch or router, or use a tool such as
PathPing or mtr. Round trip time (RTT) in ms: Get this from ping
Bandwidth (BW) in Kilobytes/second (Kbps): Use a bandwidth speed
test available from many websites. Congestion Avoidance Loss (CAL):
This is difficult to obtain. It is generally around 30% for a WAN,
but higher as distance increases. Throughput requirements Data
Transfer Pending (DTP) depends on the Priority: 1280 for low, 2816
for medium, or 4096 for high. Primary system calculations: Primary
timeout (TOp): 30 ms Throughput required (Tp) = DTP * Rp / TOp
Minimum throughput required (MTp) = 13 KB/s
11
Secondary system calculations: Secondary timeout (TOs): 40 ms 1
Gb throughput required (Ts): DTP * Rs (up to 8) / TOs 10 Gb
throughput required: N/A Minimum throughput required (MTs) = 9.6
KB/s Network Throughput Maximum segment size (MSS): 8960 if Jumbo
frames are enabled, 1460 otherwise TCP window size (TWS) in
Kilobytes: 32 KB for 1 Gb controllers, 64 KB for 10 Gb controllers
Throughput limit by packet loss (Bps): If PL = 0, then 0, else MSS/
(RTT/1000) * (1/SQRT (PL)) Throughput limit by RTT (Bps): TWS/
(RTT/1000) * (1-CAL) Throughput limit by Bandwidth (Bps): BW * 1024
Network throughput limit (NTL): Minimum of throughput limit by
packet loss (if non-zero), throughput limit by RTT and throughput
limit by bandwidth Results Primary system throughput required to
avoid timeout (TOAp) (KB/s) = Tp / Cp For the primary system
If the NTL is greater than the TOAp, the replication should not
timeout. If the NTL is greater than the MTp, timeouts will occur,
but the replication should succeed. If the NTL is less than the
MTp, the replication will fail.Secondary system throughput required
to avoid timeout (TOAs) (KB/s) = Ts / Cs for 1 Gb controllers, N/A
for 10 Gb controllers. For the secondary system with 1 Gb
controllers
If the NTL is greater than the TOAs, the replication should not
timeout. If the NTL is greater than the MTs, timeouts will occur,
but the replication should succeed. If the NTL is less than the
MTs, the replication will fail. If the secondary system contains 10
Gb controllers, then thereplication should not timeout.
The HP P2000 Array System Remote Snap Software basic
functionsThe following functions are highlighted in this
section:
Adding a remote system Creating a replication set Deleting a
replication set Exporting a replication image Setting the primary
volume Verifying the remote linkFor more information on the HP
P2000 Array System Remote Snap Software functions, see the HP P2000
G3 Modular Smart Array System SMU reference guide.
12
Adding a remote systemThis operation adds the remote system to
the local system. The remote system is listed in the navigation
tree. This operation is required when using the SMU to create a
replication set. This operation is not required when using the CLI,
but makes it easier to create a replication set.
Creating a replication setThis operation connects the secondary
volume and the primary volume. To create a replication set, specify
the local and remote systems and the link type (iSCSI or FC). For
the remote volume, you must first specify the system (the local
system is the default), and either a remote vdisk, in which case a
volume is created automatically, or a secondary volume, which must
be of the same size as the primary volume. The secondary volume can
reside on the local or remote system. Users can perform this
operation either with the SMU or CLI. Choose your link type
carefully. The link type of the replication set cannot be changed
later. Also, the primary volume does not have to be created as a
master volume. The process of creating the replication set converts
a standard volume to a master volume. However, if a secondary
volume is created, it must be created as a replication-prepared
volume, using the prepare-replication-volume parameter of the
create volume command in the CLI, or by checking the Replication
Prepare box on the Provisioning > Create Volume page.
Deleting a replication setNot all volumes associated with the
replication set are deleted, but they are automatically converted
to regular master volumes and standard snapshots.
Exporting a replication imageA replication image can be exported
to a standard snapshot, which can be mounted and used to verify
data or for any other purpose. Any change made in this new snapshot
does not affect the original replication snapshot. Snapshots
exported from a replication image count against the snapshot
license limits.
Setting the primary volumeBy default, the source volume is
designated as a primary volume and the destination volume is
designated as a secondary volume. If any failure occurs at the
local site, the secondary volume on the remote system can be
changed to a primary volume (the secondary volume now appears as a
primary volume in the remote system). The result is that this
volume now can be mapped and is accessible to the hosts. This is
necessary for applications to fail over to the remote site. Once
your local site comes up again, you must change the original
primary volume to secondary volume for the following reasons: 1. If
both the volumes participating in the replication set are primary
volumes, replication will not complete. 2. Once the volume at the
local system becomes secondary, the user can replicate data back
from the remote system to the local system. This will synchronize
the two systems. Since a secondary volume cannot be mapped, unmap a
primary volume before changing it to a secondary volume. Once the
data is replicated back to local system from the remote system,
change the primary volume to a secondary volume by giving up its
status as a primary.
Verifying the remote linkThe Tools > Check Remote System Link
function of the SMU and the verify remote-link command in the CLI
are available to verify the link connectivity between the local and
remote systems. This tool should be run before creating a
replication set. Sample outputs of check remote system link in the
Modular Smart Array SMU and Modular Smart Array CLI are provided
here.
13
Figure 7: Check remote system link output with full
connectivity
Figure 8: Check remote system link CLI output. A few FC ports
are not connected.# verify remote-link remote-system Velma
link-type all Port Type Remote Links
-----------------------------------------------------------------------------------------
A1 A2 A3 A4 B1 B2 B3 B4 FC A2,B1 FC A1,B2
iSCSI A3,A4,B3,B4 iSCSI A3,A4,B3,B4 FC FC A2,B1 A1,B2
iSCSI A3,A4,B3,B4 iSCSI A3,A4,B3,B4
--------------------------------------------- Success: Command
completed successfully. #
Types of replicationsThe HP P2000 Array System Remote Snap
Software can perform both local and remote replication.
Local replicationLocal replication occurs when the primary and
secondary volumes reside on the same system. Create the primary
volume on one vdisk and the secondary volume on another vdisk.
Then, create a replication set. Once the set is created,
replications can be initiated. The following figures provide
examples of the reporting of replication sets in the Modular Smart
Array SMU and the Modular Smart Array CLI.
14
Figure 9: Modular Smart Array SMU showing a complete setup of
local replication
15
Figure 10: Modular Smart Array CLI output showing a complete
setup of local replication (same as above SMU illustration)# show
replication-sets Replication Set [Name (rsrvd01_v001) Serial Number
(00c0ffda426100489ed7904c01000000) ] Primary Volume: Name Serial
Number Status Status-Reason Monitor Location Primary-Volume
Primary-Volume-Status MaxQueue MaxRetryTime On Error Link Type
On
Primary-Volume-Serial Collision Monitor Priority
------------------rrvd01_v001
00c0ffda4261004876d7904c01000000 Online N/A Online 32 1800
OK
Local Retry
rvd01_v001 iSCSI Oldest
00c0ffda426100484bba774c01000000 60 Medium Remot
----------------------------------------IP=16.83.136.159:3260
IP=16.83.136.160:3260 IP=16.83.136.161:3260 IP=16.83.136.162:3260
rvd01_v001 00c0ffda426100484bba774c01000000 Online N/A Online 32
1800 OK Local Retry rvd01_v001 iSCSI Oldest 60
00c0ffda426100484bba774c01000000 Medium Remot
----------------------------------------IP=16.83.136.159:3260
IP=16.83.136.160:3260 IP=16.83.136.161:3260 IP=16.83.136.162:3260
Success: Command completed successfully. # show replication-images
rsrvd01_v001 Volume Summary [Name (rrvd01_v001) Serial Number
(00c0ffda4261004876d7904c01000000) Type (Secondary Volume) ]
Images: Image Serial Number
-----------------00c0ffda42610048e3d7904c01000000 rvd01_v001_i01
Status Progress Start Date/Time 00c0ffda4261004881d8904c01000000
Snap12845610250 Suspended Est Complete 2010-09-15 14:30:29 Time
Image Name Snapshot Serial Snapshot Name Creation Date/Time
Last Update
----------------------------Complete N/A 2010-09-15 14:27:47
2010-09-15 14:30:29 N/A N/A 00:02:42 2010-09-15 15:20:24 Time
00c0ffda4261004836e4904c01000000 rvd01_v001_i02 Status Progress
Start Date/Time
00c0ffda4261004837e4904c01000000 Snap12845640230 Suspended Est
Complete
Last Update
---------------------------------------------------------------------------------------------------------------------------------------------Complete
N/A 2010-09-15 15:20:23 2010-09-15 15:20:24 N/A N/A 00:00:01
Volume Summary [Name (rvd01_v001) Serial Number
(00c0ffda426100484bba774c01000000) Type (Primary Volume) ] Images:
Image Serial Number
----------------00c0ffda42610048e3d7904c01000000 rvd01_v001_i01
00c0ffda4261004836e4904c01000000 rvd01_v001_i02 Success: Command
completed successfully. 00c0ffda42610048ced7904c01000000
rvd01_v001_i01 00c0ffda426100482de4904c01000000 rvd01_v001_i02
2010-09-15 14:27:26 2010-09-15 15:20:13 Image Name Snapshot Serial
Snapshot Name Creation Date/Time
We recommend performing an initial local replication before
performing a remote replication as part of a disaster recovery
setup or backup. If there is a huge amount of data in the primary
volume during the initial remote replication, it can take a long
time to complete and can cause bandwidth congestion. In order to
avoid congestion, the user can do a local replication and then
transfer the secondary volume to the remote system (physical media
transfer). Once the initial replication is completed, perform a
physical media transfer to move the secondary volume to the remote
site. The initial replication between a primary volume and a
secondary volume requires a full data copy between the two volumes.
Since the initial replication is a full data copy, every block on
the volume is copied. This can take a significant amount of time
depending on the amount of data and the link between the primary
and remote sites. See Figure 6 for an illustration of local
replication and physical media transfer.
16
Figure 11: Illustration of local replication and physical media
transfer
Physical media transfer Once the first replication is completed,
do the following: 1. Detach the secondary volume (outlined in next
section), which resides on the local system. If the secondary
volumes vdisk contains any other secondary volumes, detach those
volumes also. 2. First, make sure the detach operation is
completed. Stop the secondary volumes vdisk using the stop vdisk
command in CLI or using vdisk > Provisioning > Stop vdisk
function in SMU. If the secondary volume or volumes snap pools are
on other vdisks, stop those vdisks also. 3. Remove the disks or
enclosure containing the disks and move them into the remote
system. 4. At the remote site, insert the disks or attach the
enclosure to the remote system. HP recommend powering down an
enclosure before inserting disks. Make sure that the remote system
supports the chosen link type (iSCSI or FC); the link type cant be
changed once the replication set has been created. 5. Start the
snap pools vdisks using the start vdisk command in CLI or vdisk
> Provisioning > Start vdisk function in the SMU (only if the
secondary volumes snap pools are in a different vdisk). 6. Start
the secondary volumes vdisks. The secondary volume appears on the
system at the remote site. You will need to reattach the secondary
volume to add it to the set. This operation makes the secondary
volume a part of the original set. 7. Continue replicating from the
local site.
17
Detaching a replication volume
You must detach the replication volume before moving the
secondary volume to the remote system. Make sure the initial
replication is completed before detaching the secondary volume.
Once the secondary volumeis detached it remains part of the
replication set but is not updated with any new data.
After detaching the secondary volume stop the secondary volumes
vdisk and associated snap pools vdisk (if thesecondary volume and
its snap pool reside on a separate vdisks) before moving the disks
to remote system.
If you intend to move the disk drive enclosure, add this
enclosure at the end of the chain of connected enclosures.
Before removing the drive enclosure from the system, power off
the enclosure. But if you are removing only the disks there is no
need to power off the enclosure. After the drive enclosure is
powered off, there will be unwritable cache data in the drive
enclosure. Use the clear cache command in the CLI to clear the
unwritable cache. Use the detach replication-volume command via the
CLI or using the SMU: Provisioning > Detach Replication
Volume.
Reattaching a replication volume
After the secondary volume and its snap pool are moved to the
remote system, start the vdisk(s) and reattach it
using the reattach replication volume command or the
Provisioning > Reattach Replication Volume page in the SMU.
After the volume is reattached, the replication set can resume
replication operations. volume resides on one of the following:
Users can perform the physical media transfer by setting up a
local replication in such a way that the secondary On the same
array but on different vdisks (so that disks can be removed and
physically transferred to the remote site). In an attached Modular
Smart Array drive enclosure. This drive enclosure will be attached
to a remote system later.
Remote replicationFor an illustration of remote replication
usage for initialization of the remote site, see Figure 12.
Figure 12: Replication initialization options
Using FC Users can set up local and remote sites connected via
an FC network and can have replication sets performing replications
over FC. This is useful in cases where local and remote sites are
in different blocks of a campus or building.
18
FC ports can be used to transfer regular data while they are
being used for replication. However, this action will result in
decreased performance for both the data path and the replication
transfer. Using iSCSI When the local and remote systems are in
different geographical regions, users can create replication sets
using iSCSI to perform replications over WAN. For example, when the
local system is in New York and you are planning to set up your
backup system (remote system) in Houston, you can create remote
replication sets using iSCSI as the transfer media for performing
replications. Perform a physical media transfer to overcome
bandwidth and latency issues with the initial replication. These
issues can sometimes be caused by a large amount of data in the
primary volume getting replicated to a remote system. (See more on
these issues in the Physical media transfer section above). If user
is running firmware older than T230, host I/O and replication
should not be done at the same time on an iSCSI port.
Disaster recovery operationsIf the local site fails, the
applications need to switch to the remote site. To bring up the
remote site, do the following: 1. Convert the remote volume
(secondary volume) to a primary volume using the set
replication-primaryvolume CLI command or by selecting Provisioning
> Set Replication Primary Volume from the SMU. o During the
conversion to a primary volume, the volume is rolled back, or
synced, to a replication snapshot. By default, the volume syncs to
the latest replication snapshot, but you can choose any of the
previous replication snapshots. Any data that has not been
replicated is lost. A secondary volume can be converted to a
primary via the SMU or CLI.
o o
2. The new primary volume, which resides on the remote site, can
be mapped to a LUN and used as you would use the original volume.
3. Switch the applications to this primary volume and continue
using the application. Once the failure has been addressed at the
local site, complete the following steps: 1. Make the original
primary volume a secondary volume using the set
replication-primary-volume CLI command or by selecting Provisioning
> Set Replication Primary Volume from the SMU.
Important: A secondary volume cannot be mapped, so be sure to
unmap the primary volume before attempting to make it a secondary
volume.
2. Replicate any data written to the remote volume (now acting
as primary volume residing at remote system) to this volume
residing at local system (now acting as secondary volume). This can
be performed in a single replication or in multiple replications.
This ensures that all data has been transferred properly. 3. After
all the data is replicated back to the local site, convert the
volume at the local site to the primary volume and then convert the
remote volume to the secondary volume. To convert a primary volume
to a secondary volume, set the other volume of the replication set
as the primaryperform this operation on both systems. The user can
perform this operation using the CLI command set
replication-primary-volume or using the SMU via Primary Volume >
Provisioning > Set Replication Primary Volume function. For the
HP P2000 G3 SMU screen shot details, see the HP P2000 Remote Snap
Technical Cookbook. 4. Re-establish the replication set to the
remote site. Continue using the scheduler for running remote
replications in regular intervals.
19
Figure 13: Disaster recovery operations
Use casesThis white paper provides examples that demonstrate the
HP P2000 Array System Remote Snap Softwares ability to replicate
data in various situations.
20
Single office with a remote site for backup and disaster
recovery using iSCSI to replicate dataFigure 14: Single office with
a remote site for backup and disaster recovery (iSCSI)
Figure 15: Single office with a remote site for backup and
disaster recovery (iSCSI) CLI output# create vdisk level raid5
disks 1.1,1.2,1.3 vd01 Success: Command completed successfully.
(vd01) - The vdisk was created. # create snap-pool size 10GB vdisk
vd01 sp1 Success: Command completed successfully. (sp1) - The snap
pool was created. # create master-volume snap-pool sp1 size 20GB
vdisk vd01 FSDATA Success: Command completed successfully. (FSDATA)
- The master volume was created. # create remote-system username
manage password !manage 16.83.138.215 Success: Command completed
successfully. (16.83.138.215) - The remote system was created. #
create replication-set remote-system 16.83.138.215 remote-vdisk
remote-vd01 link-type iSCSI FSDATA Info: Verifying that the
secondary volume was added. This may take a couple of minutes...
Info: The secondary volume was created. (rFSDATA) Info: The primary
volume was prepared for replication. (FSDATA) Info: Started adding
the secondary volume. (rFSDATA) Info: Added secondary volume
rFSDATA. (rFSDATA) Success: Command completed successfully.
(FSDATA) - The primary volume is ready for replication. # #
replicate volume FSDATA snapshot image-1 Success: Command completed
successfully. (00c0ffdaff130000b439bc4c01000000) - Initiated
replication of volume image-1.
To configure a single office with a remote site for backup and
disaster recovery (iSCSI): 1. Set up a P2000 G3 hybrid array with
enough disks (according to the application load and users), then
configure the management ports and iSCSI ports with IP addresses.
Install the HP P2000 Array System Remote Snap Software license if
one has been purchased, or install the temporary license from the
Tools > Install License page of the SMU. 2. Create master
volumes FS Data and App A Data, enabling snapshots. Snap pools are
automatically created with the default policy and size, or you can
adjust the settings as necessary. Make sure the volumes are in
different vdisks and that each vdisk has enough space to expand the
snap pool in the future.
21
3. Connect your array to the WAN. If using iSCSI over the WAN as
part of your disaster recovery solution, connect your file server
and application server to the WAN. Connecting the management port
of an array to the WAN helps you to manage the array remotely and
is necessary when using the SMU to create the replication sets. 4.
Map the volumes to the file server and the application server. 5.
Identify a remote location and set up a second P2000 G3 array and
configure both the management ports and the iSCSI ports. This is
the remote system. Configure the vdisks to accommodate destination
volumes at a later stage. 6. In both the local system and the
remote system, add the other system using the Configuration >
Remote Systems > Add Remote System page of the SMU. 7. Verify
the data path between your local system and remote system using the
Tools > Verify Remote Link. Always configure all four iSCSI
ports to facilitate a working connection to the WAN. 8. Set up the
replication sets for the volumes FS Data and APP A Data using the
replication wizard, and choose iSCSI as the link type. 9. After the
setup is complete, schedule the replication in desired intervals,
based on the application load, critical data, replication window
(the time it takes to perform a replication) and so on. This
enables you to have a complete backup and disaster recovery setup.
10. Verify the progress of replications by checking the replication
images. It lists the progress or completed message. 11. Verify the
data at the remote location by exporting the replication image to a
snapshot and mounting it to a host. 12. In case of a failure at the
local site, it is possible to switch the application to the remote
site. Do the following:
Move the remote Modular Smart Array to the local site. Move
disks that contain the destination volume to the local site.
Replace the local Modular Smart Array with the new Modular Smart
Array, and then replicate the data. Use the iSCSI target mode to
provide the host access for the file server and application server
to access theremote site.
For more information on disaster recovery, see Disaster recovery
operations on page 19.
22
Single office with a local site disaster recovery and backup
using iSCSIFigure 16: Single office with a local site disaster
recovery and backup using iSCSI
To configure a single office with a local site disaster recovery
and backup using iSCSI: 1. Set up two HP P2000 G3 combo arrays at
the local site. 2. Connect the file servers and application servers
to the arrays via an FC SAN. 3. Mount the volumes to the
applications. 4. Create multiple replication sets with FS Data and
App A Data as primary volumes to the second HP P2000 G3 system. On
the second system, create the App B volume, map it to the hosts,
and use it as the source of a replication set with the first system
as the destination. These replication sets are created using the
iSCSI link type. We recommend that both of the HP P2000 G3 systems
are connected by a dedicated Ethernet link (LAN). 5. Switch the
applications to the other system if any failures occur on either of
the two systems.
23
Single office with a local site disaster recovery and backup
using FCFigure 17: Single office with a local site disaster
recovery and backup using FC
To configure a single office with a local site disaster recovery
and backup using FC: 1. Set up two HP P2000 G3 combo arrays at the
local site. You can also use the HP P2000 G3 FC controllers. 2.
Connect the file servers and application to these arrays via an FC
SAN. 3. Mount the volumes to the applications. 4. Create multiple
replication sets with FS Data and App A Data as primary volumes to
the second HP P2000 G3 system. On the second system, create the App
B volume, map it to the hosts, and use it as the source of a
replication set using the first system as the destination. These
sets are created using the FC link type. 5. Switch the applications
to the other system if any failures occur on either of the two
systems.
24
Two branch offices with disaster recovery and backupFigure 18:
Two branch offices with disaster recovery and backup
To configure two branch offices with disaster recovery and
backup: 1. Set up an HP P2000 G3 hybrid array with enough disks
(according to the application load and users) and configure the
management ports and iSCSI ports with IP addresses. Install the HP
P2000 Array System Remote Snap Software license, if one has been
purchased, or install the temporary license from the Tools >
Install License page of the SMU. 2. Create the master volumes FS
Data and App A Data in the array, enabling snapshots. Snap pools
are automatically created with the default policy and size, or you
can adjust the settings as necessary. Make sure volumes are in
different vdisks and that each vdisk has enough space to expand the
snap pool in the future. 3. Connect your array to the WAN. If using
iSCSI over the WAN as part of your disaster recovery solution,
connect your file server and application server to the WAN.
Connecting the management port of the array to the WAN helps you to
manage the array remotely and is necessary when using the SMU to
create the replication sets. 4. Map the volumes to the file server
and application server. 5. Identify a remote location and set up a
second HP P2000 G3 array and configure both management ports and
iSCSI ports. This is your remote system. Configure the vdisks to
accommodate destination volumes at a later stage. 6. At the local
site, create remote replication sets using the master volumes FS A
and App A. Corresponding destination volumes are created
automatically in the remote site. 7. Schedule replications at
regular intervals. This ensures that data at the local site is
backed up to the remote site. 8. Set up remote replications at the
remote site for volumes FS B and App B. Destination volumes are
created at the local site. 9. Schedule replications at regular
intervals so that all data at the remote site is backed up to the
local site. In case of failure at either the local site or the
remote site, you can fail over the application to the available
site.
25
Single office with a target model using FC and iSCSI portsFigure
19: Single office with target model using FC and iSCSI ports
To configure a single office with a target model using FC and
iSCSI ports: 1. Set up an HP P2000 G3 hybrid array with enough
disks, according to the application load and number of users, and
configure the management ports and iSCSI ports with IP addresses.
2. Create master volumes App A Data, App B Data, and FS Data in the
array, enabling snapshots. Snap pools are automatically created
with the default policy and size, or you can adjust the settings as
necessary. Make sure volumes are in different vdisks and that each
vdisk has enough space to expand the snap pool in the future. 3.
Map FS Data to the iSCSI port so that the file server can use this
volume via the iSCSI interface. 4. Map App A Data and App B Data
volumes to the FC port so that the application servers can access
these volumes via the FC SAN. Using the HP P2000 G3 hybrid
controllers provides several advantages:
You can leverage both the FC and iSCSI ports for target-mode
operations You connect file servers and other application servers
that are not part of the FC SAN to the array using the iSCSIports
via the LAN or WAN
You can connect new servers with FC connectivity directly
through the FC SANNote: Accessing a volume through both iSCSI and
FC is not supported.
26
Multiple local offices with a centralized backupFigure 20:
Multiple local offices with a centralized backup
1. Setup P2000 G3 combo controller system with enough disks and
configure the iSCSI ports with valid IP addresses. Install the
software license at remote sites A, B, and C. 2. Create volumes
(master volumes) FS A Data and App A Data corresponding to file
server and application server (accessing these volumes) at remote
sites. 3. Setup an HP P2000 G3 system at the centralized location
and make sure that it has enough disks to accommodate data coming
from remote sites A, B, and C and install the HP P2000 Array System
Remote Snap Software license. 4. Connect sites A, B, and C with the
central site using the WAN and make sure iSCSI ports are configured
and connected to this WAN. 5. Make sure the iSCSI ports of the HP
P2000 G3 arrays at site A, B, and C can access the iSCSI ports of
the HP P2000 G3 system at the central site. 6. Create replication
sets for volumes FS A data and App A Data, specifying the remote
central system as remote systems so that secondary volumes will be
created at the central site. 7. Repeat step 8 for site B and C. 8.
Schedule the replication in regular intervals so that data from
site A, B, and C will get replicated to the central site.
27
Replication of application-consistent snapshotsUsers can
replicate application-consistent snapshots on a local Modular Smart
Array HP P2000 G3 array to a remote Modular Smart Array HP P2000 G3
array. Use the Modular Smart Array SMU for manual operation and the
Modular Smart Array CLI for scripted operation. Both options
require you to establish a mechanism that enables all application
I/O to be suspended and resumed (quiesced). Many applications
enable this via a scripting method. For an illustration of the
following steps, see Figure 21. To create application-consistent
snapshots for any supported OS and any application: 1. Create a
Modular Smart Array volume for each required application volume.
When defining the volume names, use a string name variant that will
help identify the volumes as a larger managed group: With the
Modular Smart Array SMU: a) Use the Provisioning Wizard to create
the necessary vdisks and volumes. With the Modular Smart Array CLI:
b) Use the create vdisk command. c) Use the create master-volume
command. 2. Create a replication set for each Modular Smart Array
volume used by the application. a) Use a string name variant when
defining the replication set name. This helps identify each
replication set as part of a larger managed group. b) With the
Modular Smart Array SMU: i. Use the Replication Wizard for each
volume defined in step 1 c) With the Modular Smart Array CLI: ii.
Use the create replication-set command as follows:create
replication-set parm1 parm2 parmN
3. When the application and its volumes are in a quiesced state,
you can create I/O-consistent snapshots across all volumes at same
time. a) With the Modular Smart Array SMU, use the Create Multiple
Snapshots operation. The Modular Smart Array SMU also enables
scheduling snapshots one master volume at a time. For
applicationconsistent snapshots across multiple Master Volumes, we
recommend a server-based scheduling as explained in step 4. 4. For
an automated solution, schedule scripts on the application server
that coordinate the quiescing of I/O, invoking of the Modular Smart
Array CLI snapshot commands, and resuming I/O. Verify that you
defined the desired snapshot retention count. See the Figure 21
below for an example the Modular Smart Array CLI snapshot commands.
a) With the Modular Smart Array CLI: i. Use the create snapshots
command as follows create snapshots parm1 parm2 parmN The time
interval between these snapshot groups will be utilized in the
following steps.
Note: To achieve application-consistent snapshots, you must
ensure application I/O to all volumes at the server level is
suspended, prior to taking snapshots, and then resumed after the
Modular Smart Array snapshot commands are taken. The P2000 Array
System management firmware will only create point-in-time
consistent snapshots of indicated volumes.
28
Figure 21: Setup steps for replication of application-consistent
snapshots
#
To replicate application-consistent snapshots: 1. Ensure all
volumes have established reoccurring snapshots as detailed above.
2. Schedule each replication set. See Figure 17 for an example of
Modular Smart Array CLI replication commands. a) With the Modular
Smart Array SMU do the following: i. After the Replication Wizard
creates the replication set, you will be prompted to select
initiate replication now or schedule replication. Select the
schedule option. ii. In the Scheduler screen, with Replication
Mode, select Replicate Most Recent Snapshot so that the latest
primary volume snapshot will be used.
29
b) With the Modular Smart Array CLI do the following: i. Use the
create task replicate-volume command as follows create task
replicate-volume parm1 parm2 parmN ii. Use the create scheduler
command as follows create scheduler parm1 parm2 parmN 3. Ensure the
reoccurring schedule for each replication set coordinates with the
scheduled snapshots. a) Calculate an appropriate time that falls
between the application volumes snapshot times created in step 4.
b) For an example of snapshot scheduling and Modular Smart Array
replication scheduling, see Best practices on page 29. Create your
own naming scheme to manage your application data volumes, snapshot
volumes, and replication sets. In your naming scheme, include the
ability to establish a recognizable grouping of multiple Modular
Smart Array replication sets. This will help with managing the
instances of your application-consistent snapshots and the
application-consistent replication sets when restore or export
operations are used.
Figure 22: Examples of using the Modular Smart Array CLI for
replication of application-consistent snapshots# create snapshots
volumes FSDATA,APPDATA,LOG fs1-snap,app1-snap,log1-snap Success:
Command completed successfully. (fs1-snap,app1-snap,log1-snap) -
Snapshot(s) were created. # show snapshots
vdisk Serial Number Name Creation Date/Time Status Status-Reason
Type
------------------------------------------------------------------------------------------------------------------------vd02
00c0ffda03590048c6fec54b01000000 app1-snap 2010-04-14 17:43:34
Available N/A vd03 00c0ffda42160048c6fec54b02000000 log1-snap
2010-04-14 17:43:34 Available N/A vd01
00c0ffda42160048c6fec54b01000000 fs1-snap 2010-04-14 17:43:34
Available N/A
------------------------------------------------------------------------------------------------------------------------Success:
Command completed successfully.
# replicate snapshot name repapp1 app1-snap Success: Command
completed successfully. (00c0ffda03590048c6fec54b01000000) -
Initiated replication of snapshot app1-snap.
#
Replication of the Microsoft VSS-based application-consistent
snapshotsUsers can replicate the Microsoft VSS-based
application-consistent snapshots on a local HP P2000 Modular Smart
Array to a remote HP P2000 Modular Smart Array. To create
application-consistent snapshots using VSS: 1. Create the Modular
Smart Array volumes for your application: a) When defining the
volume names, use a string name variant that helps identify the
volumes as a larger-managed group. b) With the Modular Smart Array
SMU, use the Provisioning Wizard to create the necessary vdisks and
volumes. c) With the Modular Smart Array CLI: i. Use the create
vdisk command. ii. Use the create master-volume command. d) With a
VDS client tool, refer to the vendor documentation to create the
necessary volumes. 2. Create a replication set for each Modular
Smart Array volume in the application. a) When defining the
replication set name, use a string name variant that will help
identify each replication set as part of a larger managed group. b)
With the Modular Smart Array SMU: i. Use the Replication Wizard for
each volume defined in step 1.
30
c) With the Modular Smart Array CLI: i. Use the create
replication command as follows: create replication-set parm1 parm2
parmN 3. Determine an appropriate Microsoft VSS backup application,
or VSS requestor, that is certified to manage your VSS-compliant
application. a) The HP P2000 G3 VSS hardware provider is compatible
with Microsoft Windows certified backup applications. b) For a
general scripted solution, see the Microsoft VSS documents for
usage of the Windows Server DiskShadow (Windows 2008 and later) or
VShadow (applicable for Windows 2003 and beyond) tools. 4.
Configure your VSS backup application to perform VSS snapshots for
all of your applications volumes. The VSS backup application uses
the Microsoft VSS framework for managed coordination of quiescence
of VSS-compatible applications and the creation of Modular Smart
Array volume snapshots through the Modular Smart Array VSS hardware
provider. a) Establish a reoccurring snapshot schedule with your
VSS backup application. b) The time interval between these snapshot
groups will be used in the following steps.
Note: The VSS framework, the VSS Backup application (requestor),
the VSS-compliant Application writer, and the Modular Smart Array
VSS hardware provider achieve application-consistent snapshots. The
HP P2000 Array System firmware management features only create
point-in-time snapshots of indicated volumes.
To replicate VSS-generated, application-consistent snapshots: 1.
Ensure all volumes have established reoccurring snapshots as
detailed above. 2. Schedule each replication set. a) With the
Modular Smart Array SMU do the following: i. After the Replication
Wizard creates the replication set, you will be prompted to select
initiate replication now or schedule replication. Select the
schedule option. ii. In the Scheduler screen, with Replication
Mode, select Replicate Most Recent Snapshot so that the latest
source volume snapshot will be used. b) With the Modular Smart
Array CLI: i. Use the create task replicate-volume command as
follows Create task replicate-volume parm1 parm2 parmN ii. Use the
create scheduler command as follows create scheduler parm1 parm2
parmN 3. Ensure the reoccurring schedule for each replication set
coordinates with the scheduled snapshots. Calculate an appropriate
time that falls between the application volumes snapshot times as
created in step 4. For an example of such VSS snapshot scheduling
and Modular Smart Array replication scheduling, see Best practices
on page 33.
31
Figure 23: Setup steps for replication of the VSS-based
application-consistent snapshots
32
Best practicesFault toleranceTo achieve fault tolerance for HP
P2000 Array System Remote Snap Software setup, we recommend the
following:
For FC and iSCSI replications, the ports must be connected to at
least one switch, but for excellent protection it is
recommended that half of the ports be connected to one switch
(for example, the first port on each controller) and half of the
ports be connected to a second switch, with both switches connected
to a single SAN. This avoids having a single point of failure in
the switch. Direct Attach configurations are not supported by
replication over FC and iSCSI. The iSCSI ports must all be routable
on a single network space.
In case of link failure, the replication operation will
re-initiate within a specified amount of time. The amount of
time
is defined by the parameter max-retry time (the default value is
1800 seconds). Set this time to a preferred value according to your
setup. Once the retry time has passed, replication goes into a
suspended state and then needs user intervention to resume. either
0 (retry forever) or 60 minutes for every 10 GB in volume size to
prevent a replication set from suspending when multiple errors
occur. This can be done in the CLI by issuing the following
command:
During a single replication, we recommend setting the maximum
replication retry time on the secondary volume to
set replication-volume-parameter max-retry-time
Replication services are supported on both single-controller and
dual-controller environments, but only between
similar environments. That is, a single-controller system can
replicate to a single-controller system or a dualcontroller system
can replicate to a dual-controller system. Replication between a
single-controller system and a dual-controller system is not
supported. replication continues through the second controller.
We recommend using a dual-controller array to try to avoid a
failure of one controller. If one controller fails,
Volume size and policy While setting up the master volumes, keep
note of the size of the vdisk and the primary/secondary volume.Once
part of a replication set, the sizes of the primary/secondary
volume cannot be changed.
Every master volume must have a snap pool associated with it. If
no space exists on the primary volumes
virtual disk to create the snap pool, or insufficient space is
available to create or maintain a sufficiently large snap pool for
the snapshots to be retained, the snap pool should be created on a
separate vdisk that does have sufficient space.
To help you accurately set a snap pools size, consider the
following: a. What is the master volume size, and how much will the
master volume data change between snapshots? The amount of space
needed by a snap shot in a snap pool depends on how much data is
changed in the master volume and the interval in which snapshots
are taken. The longer the interval, the more data will be written
to the snap pool. b. How many snapshots will be retained? The more
snapshots retained, the more space is occupied in the snap pool. c.
How many snapshots will be modified? Regular snapshots mounted with
read/write will add more data to snap pool. d. How much modified
(write) data will the snapshots have? The more data is modified for
mounted snapshots, the more space is occupied in the snap pool.
Although the firmware allows for more, it is recommended that no
more than four volumes (master volume orsnap pools) be created on a
single vdisk when used for snapshots or replication.
33
If snap pools are considered highly critical, then the snap pool
should be created on a fault-tolerant virtual disk
(for example RAID 6), and if its not considered critical the
user can choose a virtual disk with no fault tolerance (for example
RAID 0). the snap pool size to the desired value. By default, the
snap pool policy is set to automatically expand when it reaches a
threshold value of 90%. Note that the expansion of a snap pool may
take up the entire volume or vdisk, limiting the ability to put
additional data on that vdisk. We recommend that you set the auto
expand size to a preferred value so that snap pools are not
expanded too often. It is also important that you monitor the
threshold errors and ensure that you have free space to grow the
snap pool as more data is retained. policy. When a snapshot is
deleted, all data uniquely associated with that snapshot is deleted
and associated space in the snap pool is freed for use. In order to
accommodate the number of volumes per vdisk limit delete
unnecessary snapshots.
By default, a snap pool is created with a size equal to 20% of
the volume size. An option is available to expand
Snapshots can be manually deleted when they are no longer needed
or automatically deleted through a snap pool
License Use a temporary license to enable the HP P2000 Array
System Remote Snap Software and get a hands-onexperience. For live
disaster recovery setups, we recommend upgrading to a permanent
license. A temporary license expires after 60 days, disabling
further replications. replications with live systems. remote and
source systems.
With a temporary license, users can test local replications and
gain experience before setting up remote To set up remote
replication, you must have an HP P2000 Array System Remote Snap
Software license for both the Updating to a permanent license at a
later stage preserves the old replication images. By default, there
are 64 snapshots that can be upgraded to a maximum number of 512
snapshots. Exporting a replication image to a standard snapshot
uses a license count, so be sure to install a license that
allowsfor the appropriate number of snapshots.
Scheduling The scheduler can be used to replicate data from a
primary volume in regular intervals to the remote system. In order
to ensure that replication schedules are successful, we recommend
users to schedule no more thanthree volumes to start replicating
simultaneously, although as many as 16 can replicate at the same
time. These and other replications should not be scheduled to start
or recur less than one hour apart. If you schedule replications
more frequently, some scheduled replications may not have time to
start. or when using any other tool such as Microsoft VSS and you
want to replicate these snapshots. This helps in achieving
application-consistent snapshots.
The use existing snapshot option can be used when standard
snapshots are manually taken of the primary volume
When older replication images are deleted to accommodate new
replication images (based on retention count atthe local system),
corresponding replication images at the secondary volume are
retained, until the total volume count reaches the maximum volume
limit per vdisk. maximum value of 32.
Users can set the replication image retention count to a
preferred value. A best practice is to set the count to a Schedules
may only be associated with the primary volume. You can see the
status of a scheduler by selecting aschedule from the overview
panel.
Users can modify an existing schedule to change any of the
parameters such as interval time and retention count. When standard
snapshots are taken at the primary volume in regular intervals
(manually or using VSS), select theproper time interval for the
replication scheduler so that the latest snapshot is always
replicated to the remote system. The system restricts the minimum
time interval between replications to 30 minutes.
34
The following diagram provides a summary example.
Figure 24: Tabulation of resources used with replication of
application-consistent snapshots
)
Physical media transfer Power down the enclosure or shut down
the controllers before inserting the disks at the remote system and
prior toreattaching the volume.
With stop and start operations for the vdisks, you dont need to
power down the local system to remove disks whileperforming
physical media transfer.
While performing physical media transfer, its easier to have the
snap pool of the secondary volume on the samevdisk as the secondary
volume. If the snap pool is on a different vdisk, you should first
detach the secondary volume and stop the vdisk containing the
secondary volume before stopping the vdisk with the snap pool.
Before moving the disks to a new system, ensure that the remote
system does not have volumes or vdisks with the
same names as volumes on the newly inserted disks. After
performing the transfer, the system should be checked for volumes
and vdisks with the same names. These should be renamed before
performing any other operation on the volumes.
Always check whether the initial replication is completed before
detaching the secondary volume from thereplication set.
After reattaching the secondary volume, initiate a replication
from the primary volume to continue syncing the databetween the
local and remote systems.
Replication setup wizard The replication setup wizard helps set
up remote or local replication sets. Enable Check Links when
performingremote replication. This will validate the links between
the local and remote systems.
When prompted, manually initiate or schedule the replication
after the setup wizard is completed.
Application-consistent snapshots When snapshots are taken
manually, no I/O quiescing is done at the array level. Use the
create multiple snapshots command to take time-consistent snapshots
across multiple volumes. Use the replication scheduler to initiate
the replication of these snapshots. Select the option use existing
snapshot. Software such as the Microsoft VSS framework enables
quiescing applications and taking
application-consistentsnapshots.
35
Max volume limits Replication images can decrease the volume
count per vdisk. We recommend monitoring the number ofreplication
images created.
Users can restrict the number of replication images at the local
system, where the primary volume is residing, by
using the retention count in the scheduler. Set it to a higher
value (the maximum is 32). This limit does not apply to remote
systems.
At the remote system, delete older replication images as needed.
Older replication images are deletedautomatically once the volume
count of the vdisk reaches the maximum volume limit.
Users take snapshots of up to 16 volumes at a single operation
by using the creating multiplesnapshot command.
A vdisk can accommodate only 128 volumes, including the
replication images in that vdisk.
Figure 25: Configuration limits
MonitoringReplication Users can monitor the progress of an
ongoing replication by selecting the replication image listed in
the navigation tree. The right panel displays the status and
percentage of progress. When the replication is completed, the
status appears as Completed. Events When monitoring the progress of
ongoing replication, view the event log for the following
events:
HP P2000 Remote Snap Software license expiredThis indicates that
the temporary license has expired. The HP
P2000 Array System Remote Snap Software will no longer be
available until a permanent license is installed. All the
replication data will be preserved even after the license has
expired, but you cannot create a new replication set or perform
more replications.
Snap pool thresholdThe snap pool can fill up when there is
steady I/O and replication snapshots are taken atregular intervals.
When a warning event occurs, take action: either remove the older
snapshots or expand the vdisk. initiated is queued. Once the
problem is identified and fixed, you can manually resume the
replications.
Replication suspendedIf the ongoing replication is suspended, an
event is received. Any further replicationFor more related events,
see the HP P2000 G3 Modular Smart Array System Event Descriptions
Reference Guide. SNMP traps and email (SMTP) notifications You can
set up the array to send SNMP traps and email notifications for the
events described above. Using the SMU, use the Configuration >
Services > SNMP Notification or Configuration > Services >
Email Notification pages. For the CLI, use the set snmp-parameters
and set email-parameters commands.
36
Performance tipsFor a gain in replication and host I/O
performance of up to 20 percent, enable jumbo frames on all
infrastructure components (if supported by all) in the path and on
iSCSI controllers. Jumbo frames are disabled by default for the
iSCSI host ports. You can enable them using either the SMU or
CLI.
Note: If your infrastructure does not support jumbo frames,
enabling them only on your controllers may actually lower
performance or even prevent the creation of replication sets or
replications.
With the SMU, enable jumbo frames by going to (P2000 G3 iSCSI)
> Configuration > System Settings > Host Interfaces With
the CLI, enable jumbo frames by using the command set
iscsi-parameters jumbo-frames enabled
TroubleshootingIssue: Replication enters a suspended state
Recommended actions: If performing a local replication, ensure all
the ports are configured and connected via the switch. Check the HP
P2000 Array System Remote Snap Software license status at the local
and remote site. If you are running a temporary license, the
license may have expired. Install a permanent license and manually
resume replication. The connectivity link may be broken. Use the
Verify remote link tool to check the link connectivity between the
local and remote systems. Repair the link and make sure all links
are available between the systems, then manually resume the
replication. Issue: You cannot perform an action such as changing
the external view for a replication set Recommended actions:
Actions performed on a replication set, such as schedule creation
or modification and adding or removing a secondary volume, must be
performed on the system where the primary volume resides. Changing
the primary volume is a coordinated effort between the local and
remote systems. It must first be performed on the remote system,
and then on the local system. To help remember this, the secondary
volume pulls data from the primary volume. To avoid a potential
conflict, do not attempt to have two secondary volumes. Since the
secondary volume cannot be mapped to the hosts, unmap a primary
volume before converting it to a secondary volume. Issue: You cant
delete a replication set involving a secondary volume Recommended
actions: Convert the secondary volume to a primary volume. You can
now delete the replication set.
37
FAQs1. Do we support port failover? Answer: Yes. See examples
below to understand how it works. Example Where the primary volume
is owned by controller A and ports A1 and A2 are connected is part
of the replication set (considering it as a dual-port
controller)
If port A1 fails, replication will go through A2 without any
issues. If port A1 and A2 fail, the replication will continue using
the B1 and B2 ports of controller B (must have B1 and B2ports added
to the replication set). 2. Do we support load balancing with
multiple replications in progress? Answer: Yes. Example Four
primary volumes owned by controller A and both ports (A1 and A2)
are connected and used for replication.
All four sets will try to use both ports A1 and A2, unless the
array doesnt have sufficient resources to use both ports.3. Can
CHAP be added to a replication set at any time after it is created?
For instance, if you have a local set for doing an initial
replication and then media transfer, do you need to set up CHAP
before the set creation? Answer: CHAP is specific to a system and
not specific to the replication set. CHAP is specific to the
local-to-remote system communication path and vice versa. Once you
are done with the initial replication and physical media transfer,
you can enable CHAP before reattaching the secondary volume from
the remote system; the reattach operation should go through fine.
4. Does using CHAP affect replication performance? Answer: CHAP is
just for initial authentication across nodes. Once a login is
successful with another system, CHAP will not be involved in
further data transfer, so replication performance should not be
affected. 5. I created a master volume as the primary and did a
local replication. Can I now do a remote replication with the same
primary volume? Answer: A volume can only be part of one
replication set. So, you need to delete the set and create a new
set or remove the secondary volume from the set and add the other
remote secondary volume to the set. 6. I initiated a remote
replication, and now I am not getting an option to suspend the
replication/abort the replication in the local system. Answer: By
design, suspend and abort operations can only be performed on the
secondary volume. You can access the secondary volume on the remote
system; it has an option to suspend/resume replication. 7. I
deleted the replication set using remove replication and all my
replication images disappeared. Answer: All the replication images
are converted to standard snapshots and can be viewed under the
volume in the Snapshots section.
38
8. I see an option called Enable Snapshot when attempting to
create a volume. Answer: By selecting the box Enable Snapshot, you
automatically create a snap pool for the volume. The created volume
is now a master volume. 9. I am not able to map to the secondary
volume. Answer: Secondary volumes cannot be presented to any hosts.
10. I cannot remove the primary volume from a set. Answer: Only a
secondary volume can be removed. If you want to remove the primary
volume, first make the other volume the primary volume, and then
make the original primary volume a secondary volume. You can then
remove the volume. 11. I cant expand a primary and secondary
volumes in the replication set. Answer: Master volumes cannot be
expanded, because both the primary and the secondary volumes are
master volumes; they cant be expanded even in a prepared state. In
the context of the HP P2000 Array System Remote Snap Software,
volume expansion also causes problems because both the primary and
the secondary volumes must be identical in size. This prohibits
expanding volumes which are part of the set. 12. What is the
MaxRetryTime? Answer: MaxRetryTime is the maximum time in seconds
to retry a single replication. If this value is zero, there will be
an infinite number of retries. This is valid only when error Policy
is set to RETRY. A retry for a replication occurs every five
minutes if an error is encountered. That is, a five-minute delay
occurs in between retry attempts. If the delta time from the
current time to the initial retry time is greater than the
MaxRetryTime, the replication is suspended. While the MaxRetryTime
field does not have granularity in terms of what can be set in the
field, the behavior of the retry will have a five-minute
granularity.
SummaryThe HP P2000 Array System Remote Snap Software provides
array-based, remote replication with a flexible architecture and
simple management, and supports both Ethernet and FC technology.
The software protects against detrimental impacts to application
performance, while the snapshot-based replication technology
minimizes the amount of data transferred. The HP P2000 Array System
Remote Snap Software enables the use of multiple recovery points
for daily backups, access to data in remote sites, and business
continuity when critical failures occur.
GlossarySecondary volumeThe replication volume residing on a
remote system. This volume is also a normal master volume and
appears as a secondary volume once it is part of a replication set.
It cannot be mapped to any hosts. Replication-prepared volumeThe
replication volume residing on a remote system that has not been
initialized with the first initiate replication command. A user can
create the volume using the SMU/CLI and can use it when creating a
replication set. The default option after the replication setup
wizard is to immediately perform the initiate replication command;
its a good idea to perform this during setup, for example, in the
replication of VSS-creating snapshots.
39
Primary volumeA master volume becomes a primary volume when it
is used in the creation of a replication set. It is the source from
which replication snapshots are taken and copied. It is externally
accessible to host(s) and can be mapped for the host I/O.
Replication setThe association between the source volume (primary
volume) and the destination volume (secondary). A replication set
is a set of volumes associated with one another for the purposes of
replicating data. To replicate data from one master volume to
another, you must create a replication set to associate the two
volumes. A replication set is a concept that spans systems. In
other words, the volumes that are part of a replication set are not
necessarily (and not likely) located on the same system. It is not
a volume, but an association of volumes. A master volume is part of
exactly one replication set. Replication snapshotReplication
snapshots are a special form of the existing snapshot
functionality. They are explicitly used in replication and do not
require or count against a snapshot license. Replication imageThe
representation of a replication snapshot at both the local and
remote systems. In the SMU, clicking on the table shown in right
pane displays both the local and Remote Snapshot information
associated with a particular replication image. The name of the
replication image is the time at which it was created. Sync
pointsReplication snapshots are retained both on the primary volume
and the secondary volume. When a matching pair of snapshots is
retained on both the primary and secondary volumes, they are
referred to as sync points. VSS HW ProviderA software driver
supplied by the storage array vendor that enables the vendors
storage array to interact with the Microsoft Server Virtual
Snapshot Services framework (VSS). VSS RequestorA software tool or
application that manages the execution of user Snapshot Service
commands. VSS WriterA software driver supplied by the Windows
Server Application vendor that enables the application to interact
with the Microsoft VSS framework.
40
For more information HP P2000 G3 Modular Smart Array
Systemsh10010.www1.hp.com/wwpc/us/en/sm/WF05a/12169-304616-241493-241493-241493-4118559.html
bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c02520779/c02520779.pdf
h18000.www1.hp.com/products/quickspecs/13551_div/13551_div.pdf
bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c02520791/c02520791.pdf
bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c02520788/c02520788.pdf
HP P2000 G3 Modular Smart Array System CLI Reference Guide HP
P2000 G3 FC Modular Smart Array System QuickSpecs
HP P2000 G3 Modular Smart Array System SMU User Guide
HP P2000 G3 Modular Smart Array System Event Descriptions
Reference Guide
To learn more about using HP P2000 Array System Remote Snap
Software, visit:
http://h18006.www1.hp.com/products/storage/software/p2000rs/index.html
Copyright 20102011 Hewlett-Packard Development Company, L.P. The
information contained herein is subject to change without notice.
The only warranties for HP products and services are set forth in
the express warranty statements accompanying such products and
services. Nothing herein should be construed as constituting an
additional warranty. HP shall not be liable for technical or
editorial errors or omissions contained herein. Microsoft and
Windows are U.S. registered trademarks of Microsoft Corporation
4AA1-0977ENW, Created June 2010; Updated July 2011, Rev. 2