VMware Cloud Community
vmproteau
Enthusiast
Enthusiast

Simultaneous Snaphot creation

We're about to increase permissions for several engineers to create their own snapshots. I'm looking to refresh my best practice recommendatons. Snapshot duration we have covered but, I recall a best practice recommendation about taking simultaneous snapshots on a datastore. We are running ESX 5.0 with an EMC VNX FC backend. Can someone provide the current thinking is on this?

4 Replies
a_p_
Leadership
Leadership

IMO you shouldn't have issues with simultaneous creation of snapshots. Think about backups or even VMware View, where multiple snapshots are created at the same time.

André

0 Kudos
vmproteau
Enthusiast
Enthusiast

I know there used to be recommended limits and I see most backup programs have places to limit simultaneous snapshots. Generally, 4 is the most common recommended max referenced in backup programs but wondered if that translates directly to manually created snapshots or if the simultaneous snapshots on a datstore guidance had changed.

Netbackup Resource Limits.jpg

a_p_
Leadership
Leadership

>>> Generally, 4 is the most common recommended max referenced in backup programs ...

The "4" in the above screenshot actually refers to simultaneous backup jobs.

There may certainly be an impact with creating or deleting snapshots, but I'm not aware of a recommended maximum, except there's one from the storage vendor. However, I'm willing to learn, if there's something I'm missing Smiley Wink

André

0 Kudos
cfizz34vmware
Enthusiast
Enthusiast

Can anyone from the Netbackup community share there best options/settings for VM backups so not to overload NB or VC or the SAN? The issue I ran into with NB is if I set the query to pull in all VM's it would try to start all the VM snapshots all at once and fail after a certain period of time of waiting.  But if I change it to vCenter = 5 and snapshot = 5, it completes with no errors but takes forever.

I also would like suggestions how to set the queries or policy jobs so that it does not try to snapshot my cluster nodes at the same time.

0 Kudos