Before we pickup all different stats and raw data of the dtrace, we need to have a look followings:
1. How many users are enabled in a EV server?
- We could see some few thousands users are running fine with a EV server however its not the definite answer because the performance is not only dependant on the number of users, but SQL availabiltity, Network, Index, Partition location, and etc.
- Once most of the user's items are settledown, archived as expected upto date, then only daily "aged" items need to be archived.
2. How many Archive tasks created on the EV server?
- Scattered tasks uses more process and resources, therefore this needs to be check if you see more than what it can handle.
3. Is the manual-archive from client allowed ? then we do not have restriction on abusing the system resource against End-users mistakes.
4. The recommended way to analyse the issue is
- Monitor event log. Any queue error, any mapi error ?
- Monitor MSMQ.
- Monitor the system resource usage in SQL. The SQL might have performance difficulty if there are too many items to indigest at certain point of time.
5. If you could disable the manual archive from user for while you can test followings, then you can find the most possbile cause of the issue.
- Disable manual archive
- Run Archive in report mode, then check the log and see how many items are expected to be archived in next schedule.
- Run report again in next day while compare the performance difference.
If you still see there is something needs to be addressed, then contact Symantec @.@ b
Joon
