Collector and OpsMgr agent stability.

[jeff.prusinski]

Hello.

I have a question about best practices. We have single collector connected to two VMware consoles managing about 27 hosts and 350 guests. What we are seeing is the collector and OpsMgr agent services constantly restarting. (btw we do have a case open with support, I'm just looking for information on what people have seen and done in the past. We are at V4.5). We've applied all the admin guide recommendations and have disabled disk properties and stats and network stats in the filters.

Any posts would be welcome.

Thanks,
Jeff.

[Gostev]

Jeff, I am moving this to correct forum - will ask OpsMgr PM to respond.

[Alec King]

Hi Jeff,

Do you have Ops Mgr R2? Because the overrides required to stop agent auto-restarts are different in R2.
There is also a recommended Microsoft hotfix when using R2, http://support.microsoft.com/kb/968760 .

This is detailed in the v5 Install Guide, pages 9 - 10 http://www.veeam.com/nworks_mp_for_vmwa ... ide_pg.pdf .

Hope that helps! If not, let me know your Veeam Support ticket number and I'll take a look in more detail.

Cheers,
Alec

[jeff.prusinski]

Yes thank you, I do have the R2 changes already in place. But I'm not yet prepared to move to ver5 MP just yet. I'd like for our environment to stabilize before that.

Am investigating the hotfix now.

Thanks.
Jeff.

[Alec King]

Thanks Jeff,

You probably need the hotfix. It's a Microsoft hotfix for .Net applications that handle large XML files....exactly what our Collector is
It will work when using both v4 and v5 nworks.

Without the hotfix you can see very high Handle Count for Ops Mgr agent processes (Health Service, Monitoring Host). This causes the Ops Mgr agent to auto-restart itself.

You can also override Handle Count threshold rules in Ops Mgr, but to get best system stability and a permanent fix I recommend the hotfix.

let me know how it goes!

Cheers
Alec