timgreen7077
asked on
Exchange 2010 DAG Replay Queue
Just a quick question. Why at 2am in the morning will the replay log queue be in the high 100s or low 1000s at times when no one is currently sending an receiving emails at that time of morning. Even if emails are coming through its very very few. The strange part is this is really random an doesn't happen all the time, but when it does happen it just seems strange because there is virtually almost no email traffic. I checked the eventlogs and there are no issues going on with exchange I also checked the HA Exchange logs under event logs and everything is fine there also. The copy queue logs are never having this issue its only the replay, so I just the question is why does this happen when there is no email traffic, it it generally clears without issue. It may start increasing for about 10 minutes and then rapidly decrease and all is fine. Just curious, so any ideas I will be great. I also check with our storage guys and he says everything is fine on the storage side.
My guess will be to first check for any scheduled tasks and backups running on exchange server mailbox role.
I agree with Vikas, it sounds like an exchange-aware backup or snapshot tool that is quiescing the Exchange database temporarily so it can get an application-consistent copy of the mailbox database.
ASKER
There are no scheduled tasks or backups running at that time. We do have NetApp snapshots that start at 9pm and completes around 10pm, so it only lasts about an hour. There are no other tasks running at 2am in the morning. Again this is very random. During the day i can understand at times, but early mornings it just seems strange. Any other suggestions?
Is this the passive copy? Is the disk shared with any other workloads that could be affecting performance? (High I/O on disk) Can you check the CPU/RAM usage on the server at the time this is happening? (It could be the server is busy with Anti-Virus scans)
ASKER
Yes it is a passive copy, and the disk is shared with other Exchange DBs, the systems resources are fine, and not virus scan is running.
We have 5 passive copies running on that server and no active copies, and its generally only 1 or 2 passive DBs that this may happen to. This doesn't happen all the time, but seems strange at 2 in the morning.
We have 5 passive copies running on that server and no active copies, and its generally only 1 or 2 passive DBs that this may happen to. This doesn't happen all the time, but seems strange at 2 in the morning.
How is the CPU usage? When you see this issue? How much memory and CPU allocated to these server. I assume all are VM's Make sure all VM tools are updated. Next, what is current SP and RU on the server. Finally, you should not take snapshot for Exchange also never restore it too. You should only take traditional DB and log backups.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.