Hi,

We have a two site DAG comprising of 4 physical mailbox servers (two at each site). Each of the mailbox servers is configured using SAS disks in a JBOD configuration.

We have 18 databases - one on each of the JBOD disks (no RAID). The active copies are spread across the four mailbox servers, with passive copies on the other servers (essentially we have 1 active and 3 passive copies of each database).

All seems to be working well but recently I've noticed that for no reason, two of the databases seem to be switching themselves to their non-preferred server on a regular basis. I've run CollectOverMetrics and although this does confirm that the Automatic move took place, the action reason given is "FailureItem" which is pretty non-specific and doesn't give me too much to work on.

The only logs that I can see some cause for concern are in the Exchange>High Availability>BlockReplication logs - where I'm frequently getting Event 266 which suggests that BlockModeReplication is terminated on the active side because passive copy "x" is not consuming data fast enough. I've also had Event 267 - BlockModeReplication is terminated on the passive side for database "X" because the depth limit of 10485760 bytes was reached.

Is this something to worry about or is this normal? Can anyone suggest anything else I can look at to try and nail down why these databases are switching over?

We're running Exchange 2013 SP1 CU5.

Any suggestions gratefully received.

Thanks
Steve

could you please check any backup process is freezing those databases ?

Regards

John

Hi John,

Thanks for your reply. We're using DPM2012 - would this still freeze the database given that its using VSS?

I will have a look at the backup schedule though - to see if this coincides with the switchover.

Cheers

Steve