June 30, 2015 at 12:22 pm
We have 3 replica setup in a single datacenter with one primary replica and one sync secondary and one ASYNC secondary. We have just added in 3 more ASYNC secondaries in a different data center. Since doing this we have seen our HADR_SYNC_COMMIT waits increase in both number and duration.
My understanding is that these waits are only shown when waiting for the sync replica to send back the acknowledgement that the transaction has been hardened to stable media.
So wondering why the increase in this wait just from adding ASYNC secondaries?
Any insight would be appreciated.
July 1, 2015 at 7:52 am
Hi,
Are this wait increase going in parallel with any network related waits increase?
July 1, 2015 at 12:11 pm
No no other additions in network usage. We removed two of the three ASYNC nodes at the remote site today and HADR_SYNC_COMMIT waits dropped dramatically. So these waits are definitely part of ASYNC log block transfer. I just don't understand if this is normal or a bug.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply