Implementing a two Metro Mirror topology
This chapter describes the creation of a two Metro Mirror configuration and explains recovery scenarios for failures at the remote or local site. It includes the following topics:
40.1 Overview of a two Metro Mirror topology
In a two Metro Mirror configuration, there are two Metro Mirror pairs with the same source volume. They mirror data synchronously to both secondary sites.
These are basic steps of creating a Metro Mirror topology of H1:H2 and H1:H3:
1. Establish Peer-to-Peer Remote Copy (PPRC) paths between all sites.
2. Establish the H1:H2 pairs.
3. Establish the H1:H3 pairs.
The terms H1, H2, and H3 refer to the different sites. Each site can contain many volumes and many volume pairs. Each site can also contain more than one storage system. There is no required order for establishing the pairs. “H1:H2” and “H1:H3” refer to the relationships between the sites and are independent of the order in which they are established.
To simplify the descriptions, these sites are often referred to as though they were a single volume, but it must be understood that the terms refer to the entire set of volumes being mirrored by PPRC. For example, “Failover H2:H1” means to fail over all of the H2 volumes to their corresponding H1 volumes.
|
Note: The scenarios in this chapter refer to an outage at one of the sites. This can be either a failure condition or a planned event for testing or maintenance.
|
40.2 Creating a two Metro Mirror topology
This section describes the detailed steps to create a two Metro Mirror topology.
40.2.1 Terms used in examples
For the examples in this section, the following terms are used:
•H1 is the current primary site, where the production applications are running.
•H2 is the first Metro Mirror secondary site, to which H1 is mirroring data.
•H3 is the second Metro Mirror secondary site, to which H1 is mirroring data.
Table 40-1 identifies the DS8000 storage controllers used in the examples in this chapter.
|
Note: The volume range on each DS8000 is different in the following examples only to help clarify the different sites used. It is not a requirement that they be different.
|
Table 40-1 Identifications used in DS CLI examples
|
Site
|
Role
|
Dev
|
WWNN
|
Volume range
|
|
H1
|
Current primary
|
IBM.2107-75CZM21
|
5005076305FFD75A
|
5000 - 500F
|
|
H2
|
Metro Mirror secondary
|
IBM.2107-75CYM31
|
5005076305FFD71E
|
6000 - 600F
|
|
H3
|
Metro Mirror secondary
|
IBM.2107-75CYK71
|
5005076305FFD71A
|
7000 - 700F
|
40.2.2 Establish PPRC paths
PPRC paths are required before the PPRC pairs can be established. It is best to establish all paths that might be needed so that they are available when they are required.
For example, as shown in Figure 40-1, a two Metro Mirror topology has Metro Mirror relationships H1:H2 and H1:H3. There are also Multi-Target Incremental Resync relationships (MTIR pair), also known as Multi-Target Internal Relationships, H2:H3 and H3:H2. In most cases, after a move or HyperSwap to H2, a failback H2:H1 is done after the H1 site is recovered, so paths for H2:H1 are needed. Similarly, in the event of a move or HyperSwap to H3, a failback of H3:H1 requires paths for H3:H1.
Therefore, it is advisable to establish the PPRC paths in all possible directions among the H1, H2, and H3 sites. For simplicity, the direction of the PPRC paths is not indicated in these diagrams.
Figure 40-1 Establish PPRC paths
Create the following PPRC paths:
•Establish PPRC Paths H1:H2
•Establish PPRC Paths H1:H3
•Establish PPRC Paths H2:H1
•Establish PPRC Paths H2:H3
•Establish PPRC Paths H3:H1
•Establish PPRC Paths H3:H2
Example 40-1 shows an example of a DS CLI command to create PPRC paths between H1 and H2 and then H1:H3. Similar commands are used to create the remaining PPRC paths.
Example 40-1 Establish PPRC paths H1:H2 and H1:H3
dscli> mkpprcpath -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYM31 -remotewwnn 5005076305FFD71E -srclss 50 -tgtlss 60
CMUC00149I mkpprcpath: Remote Mirror and Copy path 50:60 successfully established.
dscli> mkpprcpath -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYK71 -remotewwnn 5005076305FFD71A -srclss 50 -tgtlss 70
CMUC00149I mkpprcpath: Remote Mirror and Copy path 50:70 successfully established.
40.2.3 Create H1:H2 Metro Mirror pairs
After establishing all of the required PPRC paths, create the Metro Mirror volume pairs H1:H2. These pairs can be initially created as asynchronous Global Copy pairs and later converted to Metro Mirror, or they can be created directly as Metro Mirror pairs.
In some large configurations, it is preferable to start the replication as Global Copy because it is less likely to impact the production systems during the initial copy phase. After all pairs have copied nearly 100% of the data, the pairs can be converted to synchronous Metro Mirror.
Example 40-2 shows the use of the DS CLI mkpprc command to create Metro Mirror pairs H1:H2.
Example 40-2 Establish PPRC pairs H1:H2
dscli> mkpprc -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYM31 -type mmir 5000-500f:6000-600f
CMUC00153I mkpprc: Remote Mirror and Copy volume pair relationship 5000:6000 successfully created.
CMUC00153I mkpprc: Remote Mirror and Copy volume pair relationship 5001:6001 successfully created.
. . .
CMUC00153I mkpprc: Remote Mirror and Copy volume pair relationship 500E:600E successfully created.
CMUC00153I mkpprc: Remote Mirror and Copy volume pair relationship 500F:600F successfully created.
At this point, Metro Mirror is running H1:H2.
40.2.4 Establish H1:H3 pairs
After establishing the H1:H2 pairs, establish the H1:H3 pairs, using the same procedure used for establishing H1:H2. There is no requirement to wait for the H1:H2 pairs to reach the full duplex state before creating the H1:H3 pairs. These pairs can also be established as Global Copy and later converted to Metro Mirror, or they can be established directly as Metro Mirror pairs.
After the creation of the H1:H2 and H1:H3 pairs, a two Metro Mirror topology exists.
As noted earlier, the order of establishing the pairs does not matter, and H1:H3 can be established before H1:H2.
Creation of MTIR pairs
As described in 39.2, “Cascaded PPRC after Multiple Target PPRC failover” on page 443, when the DS8000 detects that a Multiple Target PPRC configuration exists, it creates MTIR pairs between H2 and H3.
The creation of these MTIR pairs is initiated by the DS8000 at the primary H1 site by sending commands to the secondary H2 and H3 storage systems.
After establishing the PPRC paths and creating the pairs, the two Metro Mirror topology appears, as shown in Figure 40-2.
Figure 40-2 Two Metro Mirror pairs
Example 40-3 is an lspprc query to a volume, with the primary H1 showing that it has two Metro Mirror relationships, one to H2 and the other to H3. (The output has been truncated on the right in the example output for better readability.)
Example 40-3 lspprc for H1 showing two Metro Mirror pairs on each primary volume
dscli> lspprc -dev IBM.2107-75CZM21 5000-500f
ID State Reason Type SourceLSS ...
===========================================================================================================
IBM.2107-75CZM21/5000:IBM.2107-75CYK71/7000 Full Duplex - Metro Mirror IBM.2107-75CZM21/50 ...
IBM.2107-75CZM21/5000:IBM.2107-75CYM31/6000 Full Duplex - Metro Mirror IBM.2107-75CZM21/50 ...
...
IBM.2107-75CZM21/500F:IBM.2107-75CYK71/700F Full Duplex - Metro Mirror IBM.2107-75CZM21/50 ...
IBM.2107-75CZM21/500F:IBM.2107-75CYM31/600F Full Duplex - Metro Mirror IBM.2107-75CZM21/50 ...
Example 40-4 shows an lspprc query to the secondary H2, first without the -multtgt option and again with the -multtgt option to list the MTIR pairs. Notice that the MTIR pairs are shown as Global Copy suspended pairs in addition to being marked as Multi-target Internal.
Example 40-4 lspprc query to the secondary H2
dscli> lspprc -dev IBM.2107-75CYM31 6000-600f
ID State Reason Type SourceLSS
===========================================================================================================
IBM.2107-75CZM21/5000:IBM.2107-75CYM31/6000 Target Full Duplex - Metro Mirror IBM.2107-75CZM21/50...
...
IBM.2107-75CZM21/500F:IBM.2107-75CYM31/600F Target Full Duplex - Metro Mirror IBM.2107-75CZM21/50...
dscli> lspprc -dev IBM.2107-75CYM31 -multtgt 6000-600f
ID State Reason Type SourceLSS Timeout (secs) Critical Mode First Pass Status
========================================================================================================================
IBM.2107-75CYM31/6000:IBM.2107-75CYK71/7000 Suspended Multi-target Internal Global Copy IBM.2107-75CYM31/60...
IBM.2107-75CYM31/600F:IBM.2107-75CYK71/700F Suspended Multi-target Internal Global Copy IBM.2107-75CYM31/60...
IBM.2107-75CZM21/5000:IBM.2107-75CYM31/6000 Target Full Duplex - Metro Mirror IBM.2107-75CZM21/50...
IBM.2107-75CZM21/500F:IBM.2107-75CYM31/600F Target Full Duplex - Metro Mirror IBM.2107-75CZM21/50...
If there is no connectivity between H1 and H2 at the time that H1:H3 is created, the creation of the MTIR pairs from H2:H3 will be delayed, as explained in 39.3.2, “Delayed creation of MTIR pairs” on page 445.
As a further step you might want to enable Preserve Mirror on one of the two Metro Mirror legs as described in Chapter 9, “Remote Pair FlashCopy” on page 59.
40.3 Outage at H2 or H3
An outage at one of the secondary sites, either H2 or H3, can cause the PPRC pairs to that site to suspend. Because of the design, a suspension of one relationship on a volume does not affect other relationships on the volume, so the second Metro Mirror pairs will remain active.
For example, Figure 40-3 on page 457 shows a two Metro Mirror configuration where there has been an error at the H3 site, causing the H1:H3 pairs to suspend.
Figure 40-3 Failure at H3, causing a suspension of H1:H3
A query of the H1 volumes shows that the pairs to H2 are still active and the pairs to H3 are suspended, as shown in Example 40-5. The H1:H2 pairs remain full-duplex.
Example 40-5 lspprc showing pairs H1:H2 and H3
dscli> lspprc -dev IBM.2107-75CZM21 5000-500f
ID State Reason Type SourceLSS
=======================================================================================================
IBM.2107-75CZM21/5000:IBM.2107-75CYK71/7000 Suspended Host Source Metro Mirror IBM.2107-75CZM21/50...
IBM.2107-75CZM21/5000:IBM.2107-75CYM31/6000 Full Duplex - Metro Mirror IBM.2107-75CZM21/50...
...
IBM.2107-75CZM21/500F:IBM.2107-75CYK71/700F Suspended Host Source Metro Mirror IBM.2107-75CZM21/50...
IBM.2107-75CZM21/500F:IBM.2107-75CYM31/600F Full Duplex - Metro Mirror IBM.2107-75CZM21/50...
When the remote site is recovered, the suspended pairs can be resumed, restoring the Multiple Target PPRC configuration. The MTIR pairs are not used for this case.
Example 40-6 shows an example of a DS CLI resumepprc command.
Example 40-6 Resume suspended pairs H1:H3
dscli> resumepprc -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYK71 -type mmir 5000-500f:7000-700f
CMUC00158I resumepprc: Remote Mirror and Copy volume pair IBM.2107-75CZM21/5000:IBM.2107-75CYK71/7000 relationship successfully resumed. This message is being returned before the copy completes.
...
CMUC00158I resumepprc: Remote Mirror and Copy volume pair IBM.2107-75CZM21/500F:IBM.2107-75CYK71/700F relationship successfully resumed. This message is being returned before the copy completes.
40.4 Outage at H1 site
This section describes the steps that are taken in the case of a failure or planned outage at the local production H1 site.
In a two Metro Mirror configuration, a swap can be performed to either H2 or H3. It is the user’s choice to decide which of the sites is the preferred site for the swap. In this example, the swap is being performed to H2.
40.4.1 Recover at H2
The first step in the process is to move production to H2 so that the host application I/O can continue. Either HyperSwap or a traditional restart can be used.
The steps involved in this process follow.
Freeze H1:H2 and H1:H3
Freeze commands for H1:H2 and H1:H3 remove the PPRC paths and suspend all PPRC pairs for H1:H2 and H1:H3, as shown in Example 40-7. A separate freeze command is required for each LSS to LSS PPRC relationship. With the use of consistency groups, the freeze command creates consistent data at the remote H2 and H3 sites by using extended long busy or queue full to temporarily queue dependent writes.
Depending on the type of failure at H1, the freeze commands might not complete successfully, and some or all H1 pairs can remain full duplex. If the H1 site has completely failed, these commands might not have even been issued or executed.
Example 40-7 Freeze PPRC Group H1:H2 and H1:H3
dscli> freezepprc -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYM31 50:60
CMUC00161I freezepprc: Remote Mirror and Copy consistency group IBM.2107-75CZM21/50:IBM.2107-75CYM31/60 successfully created.
dscli> freezepprc -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYK71 50:70
CMUC00161I freezepprc: Remote Mirror and Copy consistency group IBM.2107-75CZM21/50:IBM.2107-75CYK71/70 successfully created.
Unfreeze and run H1:H2 and H1:H3.
When all freeze commands have completed, the secondary H2 and H3 sites contain a consistent copy of data. The unfreeze or run command (also known as a consistency group created command) removes the extended long busy or queue full condition at the local H1 site. An unfreeze command is required for each LSS to LSS PPRC relationship.
As with the freeze command, depending on the type of failure at H1, the unfreeze commands might not complete successfully.
Example 40-8 shows an example of a DS CLI unfreeze command.
Example 40-8 Consistency group created, H1:H2 and H1:H3
dscli> unfreezepprc -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYM31 50:60
CMUC00198I unfreezepprc: Remote Mirror and Copy pair IBM.2107-75CZM21/50:IBM.2107-75CYM31/60 successfully thawed.
dscli> unfreezepprc -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYK71 50:70
CMUC00198I unfreezepprc: Remote Mirror and Copy pair IBM.2107-75CZM21/50:IBM.2107-75CYK71/70 successfully thawed.
Failover H2:H1 and H3:H1
The failover command for H2:H1 coverts the H2 volumes to suspended Metro Mirror primary volumes whose secondary volumes are H1. Similarly, the failover H3:H1 converts the H3 volumes to suspended primary volumes where secondary volumes are H1.
Example 40-9 and Example 40-10 show DS CLI examples of the failover commands for H2:H1 and H3:H1, respectively.
Example 40-9 Failover H2:H1
dscli> failoverpprc -dev IBM.2107-75CYM31 -remotedev IBM.2107-75CZM21 -type mmir 6000-600f:5000-500f
CMUC00196I failoverpprc: Remote Mirror and Copy pair IBM.2107-75CYM31/6000:IBM.2107-75CZM21/5000 successfully reversed.
...
CMUC00196I failoverpprc: Remote Mirror and Copy pair IBM.2107-75CYM31/600F:IBM.2107-75CZM21/500F successfully reversed.
Example 40-10 Failover H3:H1
dscli> failoverpprc -dev IBM.2107-75CYK71 -remotedev IBM.2107-75CZM21 -type mmir 7000-700f:5000-500f
CMUC00196I failoverpprc: Remote Mirror and Copy pair IBM.2107-75CYK71/7000:IBM.2107-75CZM21/5000 successfully reversed.
...
CMUC00196I failoverpprc: Remote Mirror and Copy pair IBM.2107-75CYK71/700F:IBM.2107-75CZM21/500F successfully reversed.
At this point, each volume on H1, H2, and H3 is a suspended primary volume pointing to the other two sites, as shown in Figure 40-4 on page 460. In other words:
•H1 is a suspended primary to H2
•H1 is a suspended primary to H3
•H2 is a suspended primary to H1
•H2 is a suspended primary to H3 (internally created MTIR pair)
•H3 is a suspended primary to H1
•H3 is a suspended primary to H2 (internally created MTIR pair)
Example 40-11 shows that the H1:H2 and H1:H3 pairs are suspended due to the freeze commands. If H1 is inaccessible due to the failure, the freeze commands might not be possible and a query might not be possible.
Example 40-11 lspprc for H1 pairs
dscli> lspprc -dev IBM.2107-75CZM21 5000-500f
ID State Reason Type SourceLSS
==============================================================================================
IBM.2107-75CZM21/5000:IBM.2107-75CYK71/7000 Suspended Freeze Metro Mirror IBM.2107-75CZM21/50 ...
IBM.2107-75CZM21/5000:IBM.2107-75CYM31/6000 Suspended Freeze Metro Mirror IBM.2107-75CZM21/50 ...
...
IBM.2107-75CZM21/500F:IBM.2107-75CYK71/700F Suspended Freeze Metro Mirror IBM.2107-75CZM21/50 ...
IBM.2107-75CZM21/500F:IBM.2107-75CYM31/600F Suspended Freeze Metro Mirror IBM.2107-75CZM21/50 ...
Example 40-12 on page 460 shows that the H2:H1 pairs are suspended from the failover H2:H1 command. The -multtgt option is used to also include the H2:H3 MTIR pairs in the output. As previously noted, the example shows that the suspended states H2:H3 and H3:H2 are created internally, not as the result of a user command.
Example 40-12 lspprc for H2 pairs
dscli> lspprc -dev IBM.2107-75CYM31 -multtgt 6000-600f
ID State Reason Type SourceLSS
============================================================================================================
IBM.2107-75CYM31/6000:IBM.2107-75CYK71/7000 Suspended Multi-target Internal Global Copy IBM.2107-75CYM31/60 ...
IBM.2107-75CYM31/6000:IBM.2107-75CZM21/5000 Suspended Host Source Metro Mirror IBM.2107-75CYM31/60 ...
...
IBM.2107-75CYM31/600F:IBM.2107-75CYK71/700F Suspended Multi-target Internal Global Copy IBM.2107-75CYM31/60 ...
IBM.2107-75CYM31/600F:IBM.2107-75CZM21/500F Suspended Host Source Metro Mirror IBM.2107-75CYM31/60 ...
Example 40-13 shows that the H3:H1 pairs are suspended from the failover H3:H1 command. The -multtgt option is used to include the H3:H2 MTIR pairs in the output.
Example 40-13 lspprc for H3 pairs
dscli> lspprc -dev IBM.2107-75CYK71 -multtgt 7000-700f
ID State Reason Type SourceLSS
============================================================================================================
IBM.2107-75CYK71/7000:IBM.2107-75CYM31/6000 Suspended Multi-target Internal Global Copy IBM.2107-75CYK71/70 ...
IBM.2107-75CYK71/7000:IBM.2107-75CZM21/5000 Suspended Host Source Metro Mirror IBM.2107-75CYK71/70 ...
...
IBM.2107-75CYK71/700F:IBM.2107-75CYM31/600F Suspended Multi-target Internal Global Copy IBM.2107-75CYK71/70 ...
IBM.2107-75CYK71/700F:IBM.2107-75CZM21/500F Suspended Host Source Metro Mirror IBM.2107-75CYK71/70 ...
Resume host systems at H2
Host systems can be restarted at H2. If HyperSwap is used, the host I/O is directed to H2.
At this point, the host systems are running at H2, and the configuration appears as in Figure 40-4.
Figure 40-4 After recovery at H2
40.4.2 Start replication H2:H3
The Incremental Resynchronization capability is used to start replication H2:H3. With the change recording function of Multiple Target PPRC, only the tracks that are potentially out of synchronization between H2 and H3 are transferred and a full copy of data is not required.
Failback H2:H3
A failback command for H2:H3 merges the change recording bitmap files for H2 and H3 and initiates the transfer of the out-of-synchronization data.
Example 40-14 shows a sample DS CLI for the failback command.
Example 40-14 Failback H2:H3
dscli> failbackpprc -dev IBM.2107-75CYM31 -remotedev IBM.2107-75CYK71 -type mmir 6000-600f:7000-700f
CMUC00197I failbackpprc: Remote Mirror and Copy pair IBM.2107-75CYM31/6000:IBM.2107-75CYK71/7000 successfully failed back.
...
CMUC00197I failbackpprc: Remote Mirror and Copy pair IBM.2107-75CYM31/600F:IBM.2107-75CYK71/700F successfully failed back.
By transferring only the out-of-synchronization data, the pairs reach the full duplex state, as shown in Example 40-15.
Example 40-15 lspprc showing full duplex after incremental resync
dscli> lspprc -dev IBM.2107-75CYM31 -remotedev IBM.2107-75CYK71 6000-600f:7000-700f
ID State Reason Type SourceLSS Timeout (secs) Critical Mode First Pass Status
===========================================================================================================
IBM.2107-75CYM31/6000:IBM.2107-75CYK71/7000 Full Duplex - Metro Mirror IBM.2107-75CYM31/60 ...
...
IBM.2107-75CYM31/600F:IBM.2107-75CYK71/700F Full Duplex - Metro Mirror IBM.2107-75CYM31/60 ...
At this point in the process, the H2:H3 pairs have been resumed. The state of the configuration is that shown in Figure 40-5 on page 462.
Figure 40-5 After failback H2:H3 with Incremental Resynchronization
The data protection of Metro Mirror has now been restored after a failure of the primary site. Without Multiple Target PPRC, host access could have been restored at site H2, but there would have been no mirroring protection until the H1 site was recovered and the H2:H1 pairs resumed.
40.4.3 H1 recovered
When site H1 is recovered, mirroring can be resumed to H2 to restore the Multiple Target PPRC configuration.
Establish paths H1:H2 and H1:H3
The paths for H1:H2 and H1:H3 were removed by the freeze commands. After H1 is recovered, these paths need to be reestablished. The paths are not required until the time when mirroring is later resumed for H1:H2 and H1:H3, but establishing them as soon as H1 is recovered helps to ensure that they are available when needed.
The paths are reestablished by using the same command used to originally create the paths, as shown again in Example 40-16.
Example 40-16 Reestablish paths H1:H2 and H1:H3
dscli> mkpprcpath -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYM31 -remotewwnn 5005076305FFD71E -srclss 50 -tgtlss 60 -consistgrp I0201:I0232 I0234:I0332
CMUC00149I mkpprcpath: Remote Mirror and Copy path 50:60 successfully established.
dscli> mkpprcpath -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYK71 -remotewwnn 5005076305FFD71A -srclss 50 -tgtlss 70 -consistgrp I0201:I0032 I0234:I0306
CMUC00149I mkpprcpath: Remote Mirror and Copy path 50:70 successfully established.
Failback H2:H1
A failback command for H2:H1 resumes the H2:H1 Metro Mirror pairs, as shown in Example 40-17.
Example 40-17 Failback H2:H1
dscli> failbackpprc -dev IBM.2107-75CYM31 -remotedev IBM.2107-75CZM21 -type mmir 6000-600f:5000-500f
CMUC00197I failbackpprc: Remote Mirror and Copy pair IBM.2107-75CYM31/6000:IBM.2107-75CZM21/5000 successfully failed back.
...
CMUC00197I failbackpprc: Remote Mirror and Copy pair IBM.2107-75CYM31/600F:IBM.2107-75CZM21/500F successfully failed back.
When these pairs reach full duplex, the Multiple Target PPRC configuration has been restored, as shown in Figure 40-6.
Figure 40-6 Multiple Target PPRC topology restored
The suspended pairs between H1 and H3 are automatically converted by the DS8000 storage system to become MTIR pairs. The change recording mechanism is started to track the out-of-synchronization data between H1 and H3. This is now the same topology as the initial state, with two Metro Mirror pairs. However, the host systems are now running at H2 rather than H1.
Depending upon the specific circumstances, for example if H1 and H2 are local to each other, there might be no preference for running at one site versus the other. In other situations, there might be reasons to prefer that the host systems run at site H1. Section 40.5, “Return production to H1” on page 463 describes a method to return to the original H1 site.
40.5 Return production to H1
Returning to site H1 is similar to a move to H2 previously described in 40.4, “Outage at H1 site” on page 457. The main difference is that this is a planned move and the original move to H2 might have been the result of an error event at H1.
40.5.1 Move host systems to H1
The first step in the process is to move production back to H1. As with the move to H2, either HyperSwap or a traditional restart can be used. The steps involved in this process follow.
Query out of-synchronization tracks
Query the out-of-synchronization tracks from H1:H3 and H3:H1. As described in 39.3.3, “MTIR change recording” on page 445, after the H2:H1 pair reaches the full-duplex state, there can be a period of time before the out-of-sync tracks counts is reduced from the full volume amount. It is important to verify that these counts have dropped from the full volume amount to ensure that failback H1:H3 for the MTIR pairs does not perform a full copy of data.
Freeze H2:H1 and H2:H3
Freeze commands for H2:H1 and H2:H3 remove the PPRC paths and suspend the H1:H2 and H1:H3 PPRC pairs. A separate freeze command is required for each LSS to LSS PPRC relationship. The use of consistency groups ensures consistent data at H1 and H2.
Unfreeze and run H2:H1 and H2:H3
When all freeze commands have completed, the secondary H1 and H3 sites contain a consistent copy of data. The unfreeze or run command (also known as a consistency group created command) removes the extended long busy or queue-full condition at the primary H2 site. An unfreeze command is required for each LSS to LSS PPRC relationship.
Example 40-18 shows an example of the DS CLI commands.
Example 40-18 Freeze and unfreeze H2:H1 and H2:H3
dscli> freezepprc -dev IBM.2107-75CYM31 -remotedev IBM.2107-75CZM21 60:50
CMUC00161I freezepprc: Remote Mirror and Copy consistency group IBM.2107-75CYM31/60:IBM.2107-75CZM21/50 successfully created.
dscli> freezepprc -dev IBM.2107-75CYM31 -remotedev IBM.2107-75CYK71 60:70
CMUC00161I freezepprc: Remote Mirror and Copy consistency group IBM.2107-75CYM31/60:IBM.2107-75CYK71/70 successfully created.
dscli> unfreezepprc -dev IBM.2107-75CYM31 -remotedev IBM.2107-75CZM21 60:50
CMUC00198I unfreezepprc: Remote Mirror and Copy pair IBM.2107-75CYM31/60:IBM.2107-75CZM21/50 successfully thawed.
dscli> unfreezepprc -dev IBM.2107-75CYM31 -remotedev IBM.2107-75CYK71 60:70
CMUC00198I unfreezepprc: Remote Mirror and Copy pair IBM.2107-75CYM31/60:IBM.2107-75CYK71/70 successfully thawed.
Failover H1:H2 and H3:H2
The failover command for H1:H2 coverts the H1 volumes to suspended Metro Mirror primary volumes whose secondary volumes are H2. Similarly, the failover H3:H2 converts the H3 volumes to suspended primary volumes whose secondary volumes are H2.
Example 40-19 shows a DS CLI example of these failover commands.
Example 40-19 Failover H1:H2 and H1:H3
dscli> failoverpprc -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYM31 -type mmir 5000-500f:6000-600f
CMUC00196I failoverpprc: Remote Mirror and Copy pair IBM.2107-75CZM21/5000:IBM.2107-75CYM31/6000 successfully reversed.
...
CMUC00196I failoverpprc: Remote Mirror and Copy pair IBM.2107-75CZM21/500F:IBM.2107-75CYM31/600F successfully reversed.
dscli> failoverpprc -dev IBM.2107-75CYK71 -remotedev IBM.2107-75CYM31 -type mmir 7000-700f:6000-600f
CMUC00196I failoverpprc: Remote Mirror and Copy pair IBM.2107-75CYK71/7000:IBM.2107-75CYM31/6000 successfully reversed.
...
CMUC00196I failoverpprc: Remote Mirror and Copy pair IBM.2107-75CYK71/700F:IBM.2107-75CYM31/600F successfully reversed.
Resume host systems at H1
The host systems can now be restarted at H1. If HyperSwap is used, the host systems are switched to H1.
At this point, the host systems are running at H1. The configuration appears as in Figure 40-7.
Figure 40-7 Move production back to H1
Each volume on H1, H2, and H3 is a suspended primary volume pointing to the other two sites.
40.5.2 Start H1:H2 and H1:H3 replication
Because this is a planned swap, there is no need to wait for the secondary system to be recovered. The replication H1:H2 and H1:H3 can be immediately resumed.
Establish H2:H1 and H2:H3 paths
The paths from H2:H1 and H2:H3 were removed by the previous freeze commands. reestablishing these paths as soon as possible helps to ensure that they are available when needed.
Example 40-20 on page 466 shows sample DS CLI commands to establish the paths for H2:H1 and H2:H3.
Example 40-20 Reestablish paths for H2:H1 and H2:H3
dscli> mkpprcpath -dev IBM.2107-75CYM31 -remotedev IBM.2107-75CZM21 -remotewwnn 5005076305FFD75A -srclss 60 -tgtlss 50 -consistgrp I0232:I0201 I0332:I0234
CMUC00149I mkpprcpath: Remote Mirror and Copy path 60:50 successfully established.
dscli> mkpprcpath -dev IBM.2107-75CYM31 -remotedev IBM.2107-75CYK71 -remotewwnn 5005076305FFD71A -srclss 60 -tgtlss 70 -consistgrp I0300:I0300 I0330:I0330
CMUC00149I mkpprcpath: Remote Mirror and Copy path 60:70 successfully established.
Failback H1:H2 and H1:H3
A failback command for H1:H2 and H1:H3 starts replication for all pairs. Example 40-21 shows sample commands for the failback.
Example 40-21 Failback H1:H2 and H1:H3
dscli> failbackpprc -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYM31 -type mmir 5000-500f:6000-600f
CMUC00197I failbackpprc: Remote Mirror and Copy pair IBM.2107-75CZM21/5000:IBM.2107-75CYM31/6000 successfully failed back.
...
CMUC00197I failbackpprc: Remote Mirror and Copy pair IBM.2107-75CZM21/500F:IBM.2107-75CYM31/600F successfully failed back.
dscli> failbackpprc -dev IBM.2107-75CZM21 -remotedev IBM.2107-75CYK71 -type mmir 5000-500f:7000-700f
CMUC00197I failbackpprc: Remote Mirror and Copy pair IBM.2107-75CZM21/5000:IBM.2107-75CYK71/7000 successfully failed back.
...
CMUC00197I failbackpprc: Remote Mirror and Copy pair IBM.2107-75CZM21/500F:IBM.2107-75CYK71/700F successfully failed back.
The suspended pairs between H2 and H3 are changed to MTIR pairs.
The original configuration is now restored, as show in Figure 40-8.
Figure 40-8 Return to original configuration
A query of the H1 volumes shows that all pairs have returned to the full-duplex state, as shown in Example 40-22.
Example 40-22 lspprc showing H1:H2 and H1:H3 in full duplex state
dscli> lspprc -dev IBM.2107-75CZM21 5000-500f
ID State Reason Type SourceLSS
================================================================================================
IBM.2107-75CZM21/5000:IBM.2107-75CYK71/7000 Full Duplex - Metro Mirror IBM.2107-75CZM21/50 ...
IBM.2107-75CZM21/5000:IBM.2107-75CYM31/6000 Full Duplex - Metro Mirror IBM.2107-75CZM21/50 ...
...
IBM.2107-75CZM21/500F:IBM.2107-75CYK71/700F Full Duplex - Metro Mirror IBM.2107-75CZM21/50 ...
IBM.2107-75CZM21/500F:IBM.2107-75CYM31/600F Full Duplex - Metro Mirror IBM.2107-75CZM21/50 ...
40.6 Copy Services Manager
The IBM Copy Services Manager provides an easy and reliable user interface to manage large copy service configurations, including a Multiple Target PPRC configuration with two Metro Mirror pairs. It simplifies and automates the tasks for recovery after both planned and unplanned outages.
On a DS8000, the Copy Services Manager can now be ordered pre installed on the DS8000 HMCs. This makes it very easy to use the Copy Services Manager. If you want a graphical user interface you should use Copy Services Manager. The DS8000 GUI no longer supports copy services. However, you can still access the old GUI to do some basic copy services functions.
Copy Services Manager is licensed for the primary usable capacity.
This book does not attempt to describe the operation or use of Copy Services Manager but gives a few examples to show the power and simplicity of this product in a Multiple Target PPRC configuration.
Figure 40-9 on page 468 is a sample of the GUI interface showing Metro Mirror pairs H1:H2 and H1:H3, including icons indicating that both relationships are HyperSwap-capable.
The role pairs displayed at the bottom of the window show the internal MTIR pairs that are created between the H2 and H3 secondary volumes.
Figure 40-9 Copy Services Manager
Complex actions can be performed in a simple and intuitive manner. For example, a planned HyperSwap can be initiated by selecting Commands and then HyperSwap, as shown in Figure 40-10.
Figure 40-10 Selecting HyperSwap
With just a few simple and intuitive actions, Copy Services Manager can HyperSwap the activity to either of the remote sites. After moving production to one of the secondary sites, replication can be started to either or both of the other sites.
As an example, Figure 40-11 shows the results after a HyperSwap to H2 and restarting replication for both H2:H1 and H2:H3.
Figure 40-11 After HyperSwap to H2 and restarting H2:H1, H2:H3
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.