System failure workflow with an active/active link

The table below outlines what happens when one of the systems involved in an active/active link fails, where the system that fails is system A and the system that remains healthy is system B.

StepWhat you doWhat happens
System A fails
1On system B, fail over the linkIf DNS failover is enabled, system B broadcasts new DNS configuration
2If DNS failover is disabled, direct clients to write only to system B
System A comes back online
3

If system A has been rebuilt:

  • On system A, upload the replication SSL server certificate from system B
  • On system B, upload the replication SSL server certificate from system A
4On system B, update the link configuration as needed
5If the link is broken, on system B, send a request to restore the linkReplication link is recreated
6On system B, fail back the linkSystem A and system B broadcast original DNS configurations; replication restarts in both directions on the link