Video Screencast Help

Journaling w EV 9.0.3 and Exchange 2010

Created: 28 Aug 2013 • Updated: 25 Sep 2013 | 11 comments
Scott _Hastings's picture
This issue has been solved. See solution.

We are migrating from Exchange 2007 to 2010. The performance on the 2010 journal mailbox is slower than the 2007 journal MB. Is this expected?


Operating Systems:

Comments 11 CommentsJump to latest comment

JesusWept3's picture

Not really, but then again someone else was complaining about this the other day, that their 2007 servers were a lot faster.

How much slower is slower btw?
Do you have any custom filters set? such as Orchestria, Compliance Acceelerator etc
Did you implement any CAS Load Balancers in your environment or any CAS arrays?
What version of Outlook do you have installed on the EV Server? 

Rob.Wilcox's picture

How much slower is slower? And how are you determing it?

Scott _Hastings's picture

I don't know exacly.. I need to measure. When we come out of backup mode, there will be ~ 25K items in the Jounal MBX and 8 hours later there were still ~ 11K items in it.

Outlook 2007 SP2.

Yes, we have F5 load balancers for the CAS Arrays

We use CA. I have entries in the InternalSMTPDomains string that could be slowing things down.

Scott _Hastings's picture

Rob, the 2007 journal mailbox cleared out within an hour. After 8 hours, I've gone from 25K to 11K items

Rob.Wilcox's picture

What state are the items in? Pending? Or 'normal'?

Scott _Hastings's picture

Ok.. Honestly I haven't thought about this. Should the Journal Mailbox be on a database that is not a DAG Member?  The journal task is pointing to a server that does not house any databases. Would this cause performance issues?

Please don't laugh at my question.....  ;-)

JesusWept3's picture

It should only cause a small lag when it first connects to the mailbox, it has to determine its location

Will that exchange server ever get the databases back? I'm not sure about best practices for dags but the exchange teams I've worked with want the db's back on their original server ASAP

That being said, if your journal task is set to do 1000 messages per pass, and say it takes 1 minute to properly query exchange and determine where the DB is located, for 10,000 messages you're looking at 10 minutes of lookup times

You could change the messages per pass but again I'd be more interested in the hotfix that I posted about earlier on

Oh and no, keep the journal live at all times, so keep it assigned to the DAGs

Scott _Hastings's picture

I installed the hot fix and didn't see any appreciable performance improvement. The journal task is set at 1000 items per pass.

Scott _Hastings's picture

Just an update...

Last night the database that houses this mailbox dismounted due to disk space being taken by transaction logs. These logs had not been backup up and truncated since the 26th (3 days ago)

This might have been causing the perfromance issue.