07-12-2016 07:14 AM
Hi,
I'm looking for some advice.
I've been using EV11 chf2 very happily on Exchange 2010 for a while, now we've recently migrated to Exchange 2013 the archive rate has reduced to about a third fo the previous rate.
All things being equal what kind of things could I do to troubleshoot this?
I have:
Thanks
07-12-2016 07:35 AM
is the EV service account hitting a throttling policy in Exch13?
look at your archiving schedule. what else is going on during that time? (eg, exchange backups, ev backups, sql maintenance)
07-12-2016 08:26 AM
Thanks for your response Andrew
The archiving schedule is clear, there are no EV/Exchange backups or maintenance during archiving windows.
The throttling policy is a good point though.
We now have a mixed Exchange 2010/2013 environment but with 3 EV systems sharing a common VSA. The policy was applied on Exchange 2010 originally and from what I understand from https://www.veritas.com/support/en_US/article.TECH216009 (below) I don't need to do anything more once I introduced Exchange 2013
Note: Please do not follow this steps if Microsoft Exchange 2013 is being deployed in a mixed environment with throttling policy already configured for Exchange 2010.
When I run "Get-ThrottlingPolicy EN* |FT Name,RcaMaxBurst,RcaRechargeRate,RcaCutoffBalance,RcaMaxConcurrency" from 2010 I see $NULL values and when I run the same from 2013 I see values of 'Unlimited'. These are the correct values as the manually creating throttling policy documents https://www.veritas.com/support/en_US/article.TECH157927 and https://www.veritas.com/support/en_US/article.TECH216009)
Am I missing something?
07-12-2016 09:47 AM
one way to rule it out would be to check the exchange logs during archiving. i dont remember the event id off the top of my head but i think there is one for a client exceeding connection limits or something like that. the other thing to check would be Exchange AV or whatever use like Microsoft TMG
07-13-2016 07:22 AM
Thanks again Andrew.
AV is the same as Exchange 2010 with updated exclusions for Exchange 2013.
The throttling policy looks OK, I think the below from a DTrace log confirms it:
5643 16:55:47.630 [15428] (ArchiveTask) <7012> EV:H {CAgentTask::CheckThrottlingPolicyDetails:#4764} The throttling policy [Throttling Policy Name] applied to user [VSA DN] has the correct RCAMaxConcurrency value
I've removed an old "ClosestGC" registry value and ensured the correct "DS Server" value is in place.
I've also set the correct proxy server address in the Domain Properties in EV, removing the SSL check. We had previous entries from some testing.
This seems to have improved the archive rate slightly, from maybe 40% to 60% of the previous rate. Better but still down on the Exchange 2010 rates.
I don't see anything obvious in the Exchange server event log, any chance you could dig out the event ID or text?
07-13-2016 10:19 AM
Which version of outlook are you using? In the old days, I remember the archvie rate with outlook 2007 is much faster then outlook 2013 due the way it manages the connection and it was recommanded for exchange 2010 below.
If you upgraded to outlook 2013,which is required to support exchage 2013, I think a there should be a performance drop. Again, this is a bit of curve ball. Support might have more say on this.
Cheers
JQ
07-14-2016 04:05 AM
Thanks JQ
Yes, Outlook 2013 SP1 32-bit, latest (up to 1st July) updates. This has been on the server for months archiving the Exchange 2010 mailboxes. We noticed a small drop in archive rate when updating Outlook (2007/2013) but nothing like the >30% we're seeing since moving the mailboxes.
I'm going to test changing the archive task connections from default 5 to somthing lower. Also get full offline SQL maintenance this weekend.
If I get no improvement I'll raise it with support.
07-15-2016 08:26 AM
Update.
I have reduced the Exchange connections per task from 5 to 1 and the archive rate has dropped to 1/5 of the previous reduced rate.
Perhaps this reduced rate is just the new baseline for Exchange 2013 archiving? Maybe an update to 11.0.1 chf5 is required but I've read nothing about this problem in any of the update notes.
I'll get a support request raised Monday.
Thanks for the comments AndrewB and John Q