CLI Guide

Arguments
Required arguments
[-c|--cluster] cluster *The cluster on which to roll back and resume I/O.
[-g|--consistency-group]
consistency-group
* Context path of the consistency group on which to roll back and resume I/O.
Optional arguments
[-f|--force]
Do not prompt for confirmation. Allows this command to be run using a non-interactive
script.
* - argument is positional.
Description
Use the choose-winner command when:
● I/O must be resumed on a cluster during a link outage
● The selected cluster has not yet detached its peer
● The detach-rules require manual intervention
The selected cluster will detach its peer cluster and resume I/O.
CAUTION:
When the clusters cannot communicate, it is possible to use this command to select both clusters as
the winning cluster (conflicting detach). In a conflicting detach, both clusters resume I/O independently.
When the inter-cluster link heals in such a situation, manual intervention is required to pick a winning cluster.
The data image of the winning cluster will be used to make the clusters consistent again. Any changes at the
losing cluster during the link outage are discarded.
Do not use this command to specify more than one cluster as the winner.
Examples
Select cluster-2 as the winner for consistency group TestCG:
VPlexcli:/clusters/cluster-2/consistency-groups/TestCG> choose-winner --cluster cluster-2
WARNING: This can cause data divergence and lead to data loss. Ensure the other cluster
is not serving I/O for this consistency group before continuing. Continue? (Yes/No) Yes
In the following example:
● l The two ls commands show a consistency group my_cg1 when an inter-cluster link outage has occurred. The detach-rule
is no-automatic-winner, so I/O stops at both clusters, the status summary is suspended (showing that I/O has
stopped), and the status details contain cluster-departure, indicating that I/O has stopped because the clusters can
no longer communicate with one another.
● The choose winner command forces cluster-1 to detach cluster-2.
● The ls command displays the change at cluster-1.
○ Cluster-1 status is suspended.
○ Cluster-2, is still suspended, cluster-departure.
○ Cluster-1 is the winner, so it detached cluster-2.
● I/O at cluster-1 remains suspended, waiting for the administrator.
VPlexcli:/> ll /clusters/cluster-2/consistency-groups/
my_cg1/
/clusters/cluster-2/consistency-groups/my_cg1:
98
Commands