Video Screencast Help

Enterprise Vault - Correct sequence of index, collection, admin tasks

Created: 02 Sep 2013 • Updated: 03 Sep 2013 | 7 comments
This issue has been solved. See solution.

Hi everyone,

After some pretty major performance degredation after upgrading from 9.0.1 to 10.0 (and then 10.0.3) Enterprise Vault, I am being forced to review the schedules of the various background processes in order to ensure we are not overlapping or causing unecessary load on the server.

We run EV on our MS Exchange 2010 platform, all o which is on Windows 2008 VM's with the reccomended level of vCPU & vRAM on SAS disks.

My understanding is that the each environment has the following processes, all of which are scheduled to run at some point during a 24 hour period;

Mailbox Archiving, Mailbox Sync, Mailbox Provisioning, Indexing, Vault Store Collection, Index Admin Tasks

We have been advised to alter our sequence of processes so they look like the following;

1. Index admin tasks are run from Sunday - Friday inclusive between 11:00-18:00

2. Mailbox Provisioning runs daily @ 18:00

3. Site Schedule is every day 19:00-02:00 and I understand includes Mailbox Archiving tasks, Mailbox Sync and Indexing Task

4. Vault Store Collection runs between 21:00-00:00 for our oldest 10 Vault Stores & 21:00-02:00 for our newest 10 Vault Stores (I could potentially remove the overlap here if it is felt that 2-3 hours is adequate per run - or less for the older set of Vault Stores which shouldnt change too much)

5. Backup via Backup Exec 2012 runs from 03:00 daily

Whilst this seems logical, I want to ensure that the sequence of events is correct as we previously had our Vault Store collections running during the day from 10:00-16:00 - I just want to be 100% certain that everything is happening in the right order going forward.

Thanks in advance,

Martin

Operating Systems:

Comments 7 CommentsJump to latest comment

GabeV's picture

Hello M-Hood,

This schedule looks fine. However, you didn't mention when or how often you are running the SQL maintenance. During the archiving schedule (19:00-02:00) you have collections running (21:00-00:00 and 21:00-02:00). I am assuming that you are collecting files older than XX amount of days, so EV is not collecting data that was just archived. During this time, your SQL server will be under heavy load and can get fragmented really quick. Just make sure you monitor the SQL fragmentation, thus EV performance won't be affected.

I hope this helps.

“Success is not final, failure is not fatal: it is the courage to continue that counts.”–Winston Churchill

GabeV's picture

Just in case, here is the maintenance plan for SQL:

http://www.symantec.com/docs/TECH74666

“Success is not final, failure is not fatal: it is the courage to continue that counts.”–Winston Churchill

M_Hood's picture

Hi,

Thanks for the feedback. Our DBA is not too keen on the Symantec method of DB maintenance as he states rebuilding tables etc on a regular basis is actually detrimental. I have to go with him on this one as he owns the DB side of things.

We do have a scheduled defragment of the indexes on all DB's on the relevant SQL server at 02:00 each morning which works fine. I have just checked and most of the indexes are showing a "Logical Scan Fragmentation" of zero percent or slightly more.

You are right - our collection is set to collect files older than 10 days with a max collection size of 10MB - these are the default settings which we have not felt the need to adjust (tell me if I am wrong though!).

Hopefully this adds some weight to our hope / expectation that we are setup in an efficient manner?

Thanks again

GabeV's picture

We never recommend rebuilding tables ... I hope you meant rebuild indexes, right? ... otherwise, you would be, basically, deleting data from the databases. Now you have confirmed the collection settings (you're not wrong, btw) and the defrag scheduled, I think the schedule is OK.

“Success is not final, failure is not fatal: it is the courage to continue that counts.”–Winston Churchill

SOLUTION
M_Hood's picture

Hi - Yes, sorry - I do mean rebuilding the indexes!

Thanks for the feedback - its always good to gather second opinions

GabeV's picture

That's what I thought ... I just wanted to clarify. If you don't need more assistance regarding this topic, please mark the post that best solves your problem as the answer to this thread.

“Success is not final, failure is not fatal: it is the courage to continue that counts.”–Winston Churchill