VMware Cloud Community
sxnxr
Commander
Commander
Jump to solution

Monitering snapshots

Hi

We get requests for snapshots all the time at all hours of the day. We currently use the vcenter scheduled tasks function to run the snapshots at a particular time. We set up a email alert that notifies a couple of ppl when it runs. The problem is we get a mail when that task starts the snapshot process not if the snapshot was taken successfully.

Is there a way to have vrops alert when a snapshot fails and look at the scheduled tasks for the name of the VM that the task started and if both VM names match send an email.

0 Kudos
1 Solution

Accepted Solutions
bbambrick
Enthusiast
Enthusiast
Jump to solution

I'd be careful of snapshots and vROPs right now. We run a reasonably sized environment but our backups are done via Netbackup snapshot backups. By default (6.1 and below) vROPs captures information about these snapshots, and stores them as objects. Every time a backup runs (every day) each VM gets snapped, backed up, and then the snap is removed. This has caused a massive bloating in the number of metrics in our cluster, making it extremely slow and unusable, as each snapshot is a unique snowflake...

The fix from VMware is to stop monitoring snapshots. I believe this has been rolled into the 6.2 default policies until they can figure out a more elegant solution.

Aside from that nugget, here's how I did something similar to what you're looking for:

  • Setup a powercli script which connects to your vCenters once a day and looks for snapshot related failures in the event logs. If it finds an error, then tag the VM with a custom tag.
  • In vROPs, setup a symptom for a VM having that custom tag, and setup an alert based off that symptom.
  • You can now setup a custom notification for that alert Smiley Happy

View solution in original post

0 Kudos
2 Replies
bbambrick
Enthusiast
Enthusiast
Jump to solution

I'd be careful of snapshots and vROPs right now. We run a reasonably sized environment but our backups are done via Netbackup snapshot backups. By default (6.1 and below) vROPs captures information about these snapshots, and stores them as objects. Every time a backup runs (every day) each VM gets snapped, backed up, and then the snap is removed. This has caused a massive bloating in the number of metrics in our cluster, making it extremely slow and unusable, as each snapshot is a unique snowflake...

The fix from VMware is to stop monitoring snapshots. I believe this has been rolled into the 6.2 default policies until they can figure out a more elegant solution.

Aside from that nugget, here's how I did something similar to what you're looking for:

  • Setup a powercli script which connects to your vCenters once a day and looks for snapshot related failures in the event logs. If it finds an error, then tag the VM with a custom tag.
  • In vROPs, setup a symptom for a VM having that custom tag, and setup an alert based off that symptom.
  • You can now setup a custom notification for that alert Smiley Happy
0 Kudos
vishalchand
Enthusiast
Enthusiast
Jump to solution

Are you able to share the script, I would like something similar for reporting purpose. Thanks in Advance

0 Kudos