Administrator Guide

Table 12. Consistency group field descriptions (continued)
Property Description
Advanced properties
auto-resume-at-loser
Determines whether I/O automatically resumes at the detached cluster for the
volumes in a consistency group when the cluster regains connectivity with its
peer cluster.
● Relevant only for multi-cluster consistency groups that contain distributed
volumes.
● Modifiable using the set command. Set this property to true to allow the
volumes to resume I/O without user intervention (using the resume-at-
loser command).
● true - I/O automatically resumes on the losing cluster after the inter-cluster
link has been restored.
● false (default) - I/O must be resumed manually after the inter-cluster link
has been restored.
● Leave this property set to false to give administrators time to restart the
application. Otherwise, dirty data in the host’s cache is not consistent with
the image on disk to which the winning cluster has been actively writing.
Setting this property to true can cause a spontaneous change of the view
of data presented to applications at the losing cluster. Most applications
cannot tolerate this data change. If the host flushes those dirty pages out
of sequence, the data image may be corrupted.
Display-only properties
active-clusters
For synchronous consistency groups, this property is always empty ([ ]).
operational status
Current status for this consistency group with respect to each cluster on which it
is visible.
● ok - I/O can be serviced on the volumes in the consistency group.
● suspended - I/O is suspended for the volumes in the consistency group.
The reasons are described in operational status: details.
● degraded - I/O is continuing, but there are other problems as described in
operational status: details.
● unknown - The status is unknown, mostly because of lost management
connectivity.
operational status: details
If operational status is ok, this field is empty: [ ]. Otherwise, it displays
additional information, which may be any of the following:
● cluster-departure - Not all the visible clusters are in communication.
● data-safe-failure - A single director has failed. The volumes are still
crash-consistent, and remain in this state, unless a second failure occurs
before the first is recovered.
● rebuilding-across-clusters - One or more distributed member
volumes is being rebuilt. At least one volume in the group is out of date at
that cluster and is re-syncing. If the link goes out at this time, the entire
consistency group is suspended. Use the rebuild status command to
display which volume is out of date at which cluster.
● rebuilding-within-cluster - One or more local rebuilds is in progress at this
cluster.
● requires-resolve-conflicting-detach - After the inter-cluster link
is restored, two clusters have discovered that they have detached from
one another and resumed I/O independently. The clusters are continuing to
service I/O on their independent versions of the data. The consistency-
group resolve-conflicting-detach command must be used to make
the view of data consistent again at the clusters.
● requires-resume-after-rollback - A cluster has detached its peer
cluster and rolled back the view of data, but is awaiting the consistency-
group resume-after-rollback command before resuming I/O.
Displayed:
80
Consistency Groups