HP StorageWorks SAN Virtualization Services Platform Administrator Guide (5697-0934, May 2011)

the mirror tasks are missing and that the mirror cannot decide which task to use. The window
offers these options:
• Resume—Use this option after you have fixed the problem that prevented the access to
some of the mirror tasks.
• Recover—Use this option after you decide that you cannot recover the failed task. This
option offers you a list of the surviving tasks and asks you to pick the one to use. Select
one of the tasks from the drop-down list, and click OK.
Figure 37 Recovery tab—VSM in Partial state
After clicking OK, the VSM application will restart itself and will come up as the active VSM.
After the VSM comes up, the VSM will use only the surviving setup virtual disk to synchronize
the mirrored tasks.
3. On the surviving site, for every failed synchronous mirror group, right-click the mirror group,
and select Manage > Recovery > Force Resume.
4. To run Force Resume on all of the failed synchronous mirror groups, right-click the synchronous
mirror entity in the left pane, and select Recovery > Force Resume All.
Force Resume assigns the local synchronous mirror task to be the new primary task for the
group, marks the other tasks as secondary, and brings the synchronous mirror virtual disk to
a state of Normal.
If the mirror mode of the synchronous mirror group is Automatic Failover, the host is able to
access the virtual disk.
If the mirror mode is Synchronous Always, the host cannot use the virtual disk because one
task has failed. In this case, you can either change the mirror mode of the synchronous mirror
group to Automatic Failover (the preferred action) or break all of the failed tasks.
Re-establishing the connection with the recovered site after Force Resume
1. Make sure that all of the components at the recovered site are online except for the DPM that
you deliberately powered off.
2. Make sure that the VSM servers at both sites are online and see each other.
3. Turn on the power to the DPM at the recovered site. The DPM at the recovered site will join
the system as a passive DPM for all virtual disks assigned to that DPM group.
4. Initiate resynchronization or resume for all failed synchronous mirror tasks. The DPMs use the
currently active task as the source for synchronizing the task at the recovered site. The DPM
might decide to fail this resynchronization request in case the failed task is a more recent
update than the surviving task that was recovered by Force Resume. In this case, you must
break this failed task and use it as a standalone virtual disk.
Site failures 91