Still receiving SCOM Alerts for a shutdown VM

[ncott]

Hello,

We have an odd situation that I cannot figure out.
We have received a few vDisk alerts from SCOM for low disk space, however the VMs referenced in the alert were shutdown a month before.
The only activity occurring with these VMs are snapshot backups.
How, or why, is the nworks Veeam Collector still gathering info for a shut down VM? I would understand a heartbeat failure, but the fact it is reporting on vDisk space has me scratching my head.
Can anyone help me understand what is going on? Have I missed a configuration setting somewhere?

Thanks in advance for your help!

Nathan

[Alec King]

Hi Nathan,

This is probably caused by stale data remaining in WMI. We publish our performance data into WMI as instances, for each VM, each VM vDisk, etc. And even though the VM is shut down; sometimes the last remaining instance can be 'cached' in WMI, and so SCOM keeps collecting it.
You can see this situation in SCOM, if you look at a performance chart for a shut down VM. All the metrics will keep collecting, and the chart will be just straight lines....
And of course if the performance data is still collected, then it is possible for a monitor to keep firing, if the last performance sample was a breach of theshold. In your case, the last sample for vDisk Space was below threshold, so the vDisk Free Space monitor will keep alerting.

In our next update, we have moved to .Net 4.0 and a new WMI method, which allows us to revoke old stale data. Also our data source providers in the MP are updated, to skip powered off VMs.
For the moment in your current version, you will clear the WMI cache if you restart our nworks Virtual Infrastructure Service.

Hope that helps!
Regards,
Alec

[ncott]

Ah, ok. I figured it was something along those lines. Thanks!
I'll try restarting the service and see if that helps clear things up.

Good to hear about the improvements in the next version! Is there a timeframe for it's release? We're just now planning to upgrade to SCOM 2012 and are reevaluating all our components, so I would like to include the new version/license if possible.
In the meantime, I'm trying suggestion in this forum post: http://forums.veeam.com/viewtopic.php?f=1&t=4808
We've set up a resource group with all shutdown VMs and I will attempt to set a custom attribute and create a group from that to disable monitoring/alerting for that group.

Again, thanks for the quick reply!

[Alec King]

Hi Nathan,

We'll be releasing our vNext MP early in the new year - but we will have a Beta program starting in just a couple of weeks. If you'll have SCOM 2012 in your lab, I think you'd enjoy all the new cool dashboard views we've implemented for 2012 users. Loads of other new features, such as new Monitor types and lots of new Reports (those work for both 2007 & 2012 users). If you'd like to join the Beta program, just let me know!

Meantime, creating your group and applying overrides to disable the performance monitors will work. You can create Veeam.segment Custom Attribute in vCenter and use that as the condition for VMs joining your new group. This will be fed through to MP as property 'Segment' of the VMGUEST class.

Any questions let me know - and please advise if you'd like to join the beta! I'd welcome your feedback

Cheers,
Alec