Hello dear colleagues!
Our customer tested failover for synchronous replication for two Unity 400. Customer simulated planned switchover between sites. Customer made "Failover" command and, after this he made "set -pause no" for start of replication in backward direction. Failover was finished correctly but pair status is Out-of-sync and Syncing process has been started as FULL. For group LUN with size 1TB per LUN this process took more than one hour of time.
Could somebody explain to me is this "works as designed" or I should open an SR?
You'll need to provide some more information - what is the distance between the two arrays? Is this FILE Replication or BLOCK Replication?
Yes, you are right. This is a block FC replication. Distance between sites is about 10 kilometers. Customer already tested this process with small LUNs (10-20GB) and he was satisfied with results.
I am facing the same issue. After failover using SRM and Reprotect. It take long time as initial syncronization. Is there a way to avoid full sync?
Unfortunately this works as designed. If you issue failover command from source box you can set an option -sync yes and system will make switchover a replication (swap source and target personality and establish incremental replication). You can use this option for planned failover. But if you issue command failover from target box and use an option set paused no, the box will swap source and target personality and establish FULL replication. And you has no any other option. I use AIX HACMP and I wrote special script for verifying states of source/target resources, replication status and selecting of failover options.