vCenter alerts

Lexluethar

Has anyone worked with creating an alert for the scratch config location? I would like a vCenter alert setup to email out if the scratch config location is on non-persistent storage OR is blank.

We currently write out log files to an iSCSI storage array, we started doing this because we started ordering servers that don't have hard drives, only a flash drive for ESXi to install on. It was great at first because it was a centralized location to store everything so if you completely lost a Host you could still have access to the logs.

I'm finding though that monitoring this is a pain. If the Host loses connection to the iSCSI array it defaults back to a temp location. It's fine in one off situations where we remove a Host for troubleshooting but if it's unplanned going back and viewing what logs are being written has become a chore.

Any input would be appreciated,

iBrokeIT

Unfortunately the VMware alerts are somewhat limiting and do not cover this specific scenario even though an ESXi hosts notes that there is a configuration issue on the summary tab when your logs are stored on non-persistent storage.

Maybe try setting up a host profile that defines the location for your scratch config and then set an alert for when your host goes out of compliance from that profile.

Lexluethar

Thanks man, that's the end goal and I have a profile setup just haven't started applying it yet. I know in vCenter you can have an alert if a Host isn't compliant with a profile - that might be my only option.

It's not a huge deal, just would be nice to know if a Host had storage issues that it didn't wipe the scratch location. At this point any time there is an issue i just have to double check that location.

Thanks!

scott28tt

vSphere 6 includes a default "Host storage status" alarm that can trigger if a host loses access to a storage device - that might do the job...?

Lexluethar

5.5 has a similar alarm, we use something similar with our 3rd party monitoring tool which will alert if a Host losses storage. The issue here with the all storage alarm is it doesn't take into account the current scratch location. So yes you can assume that during a storage issues the scratch location may not be in use BUT not every storage issue would cause the scratch location to reset to the default which is temp.

So i'd like a mechanism to alert if that currentscratchlocation is either blank or in a temp folder (AKA non-persistent storage). The oddball thing imo is vCenter alerts you, you get a big yellow splat on the Host showing this issue but it doesn't consider it an alarm. Similar to if SSH is enabled, you get the yellow splat but no alarm definition along with that yellow splat.