Speed up snapshot creation/deletion

[secretdk]

Hello,

We are using Nimble storage snapshot integration to do our backups. Currently it takes around 30 mins to create a VM snapshot of our VMware production cluster, create the storage snapshot and delete the VM snapshot. For backups this was not a big problem, but now we are trying to do replication as well. This also takes 30 mins every time we want to sync our replica nodes. Is there any way to make this go faster? Currently snapshot creation/deletion is the bottleneck in how long Veeam operations take.

[Andreas Neufert]

VMware Snapshot creation is a simple task that should take only some seconds.
Can you please explain a bit the timing?

Potentially post some screenshots of Job statistic. Or open a support case and let our support have a look at the logs?

[secretdk]

Creating the snapshot itself takes around 5-10 seconds. There are 121 VM's in a job and it only seems to do 2? at a time. That means it will take around 5-10 minutes just to create the snapshots. Next to that collecting disk files location takes 30 seconds on average and for some VM's even more than 2 minutes. Removing the VM snapshot after the storage snapshot has been taken takes another 10-15 seconds.

If there is no easy tweak to make it do snapshot creation/deletion with more VM's simultaneously or to speed up the disk collection, I will create a support ticket to deep dive into it with them. Just hoped that maybe I overlooked some setting that could resolve this.

[Andreas Neufert]

Usually we do arround 20 VM guest processings in parallel. So the question is why only 2 are processed on your side.

As well there is an option in the Job-Storage-Advanced-Storage/Integration area where you can create multiple snapshots per run. Bring 10VMs in consistent state, then create storage snapshot... then next 10 and so on and so on.
This should speed up things.

[secretdk]

Check. Making storage snapshots per X amount of VM's indeed makes sense in this case. That would speed up the processing of VM's instead of waiting for the entire batch. Will check with support on why the concurrent VM's being processed during snapshot creation/deletion is so low.

[Andreas Neufert]

Please check as well if you are on latest Veeam version v9.5 Update 4a. We always add performance enhancements within the storage area in any release.

[cmln]

Hello,
I'm having somewhat the same issue here.
I'm using Veeam 10.
When the job starts it takes around 10 min to snapshot the VMs (70 VMs), then it sits there for about 25 to 30 min doing nothing and then it starting to delete the snapshots 2 by 2.
The deletion 2 by 2 that doesn’t bother me, what’s bothering me is the time between taking the snapshot and deleting them.
Any idea?
Greetings

[Gostev]

Hello! Did you have our support engineer investigate through the debug logs what Veeam is doing during those 25 to 30 min? As there has to be either some activity or waiting for some infrastructure resource... an active job cannot just sit and do nothing. Thanks!