Netapp Snapshot deletion does not work as expected

[sandsturm]

Hi
All our Veeam backup jobs have configured Netapp Snapshots as Secondary targets with a retention of three. The primary storage integration settings are set to "Limit processed VM count per storage snapshot to 10". The meaning with the retention of three is to have three days of Snapshots because the backupjob runs once a day. The problem is, that for all jobs with more then 10 VMs, more then one storage snapshot will be created and I lose the retention of the three "days". Example, a Backupjob with 34 VMs creates a Netapp snapshot for every ten VMs and therefore I have three Netapp snapshots after the job was running, this means that I have a retention of maximum one day and not three days. If I disable the "Limit processed VM count per storage snapshot to 10" only one Netapp snapshot is created with the run of the backupjob and I have my three days retention.
My question is now: Is this the expected behavior? Because the setting: "Limit processed VM count per storage snapshot" is meant for the primary storage integration and in my opinion only as a helper for the backupjob itself (to shorten the time where the VMware snapshot is required) and not really for the "secondary target" or am I wrong?

thx,
sandsturm

[foggy]

Yes, this is currently the expected behavior. Well, maybe not so expected by customers but at least this is how it should work according to the implemented logic. Given the limit of 10 VMs on a snapshot, the job prepares VMs in groups of 10 (4 in the last one) and creates 4 snapshots, from which one goes away according to your configured retention of 3 snapshots. This logic is based just on the number of snapshots and should be improved to account for the fact that each snapshot contains a valid state for the group of VMs only, so all 4 snapshots should be taken as a single restore point. We will consider this improvement for the future versions, thanks!

[sandsturm]

Okay, thanks for the clarification. Would be nice if you could improve this with a feature request to not only hold an amount of snapshots but a retention for days. Similar to the new option of restore points or days in the retention policy settings of a VM job in the upcoming VBR10, which is a among others, a really cool new feature by the way

[foggy]

Yes, spreading this option to storage snapshots is on our list of future improvements.

[sandsturm]

@foggy, how about the feature to have the retention in days instead of just retention points for Snapshots and Snapvaults? Any plans, if this will be implemented?

thx
sandsturm

[Andreas Neufert]

Please see above the answer from Alexander.

[foggy]

Hi sandsturm, we're working on it, stay tuned.

[sandsturm]

okay, thanks

[Matthias.Mehrtens]

@foggy wrote:
we're working on it, stay tuned.

Is there a chance to get it with v11? I've got another customer asking for it.

[foggy]

Hi Matthias, in v11 we've redesigned retention to count VM restore points rather than volumes (the initial request in this thread), however, retention in days is not in v11 scope.

[Matthias.Mehrtens]

Hi Alexander, thanks for your quick reply!
I'm not sure if I understand it fully: Could you explain what "counting VM restore points rather than volumes" would mean for the example given above?

[foggy]

For the example in the original post (with 34 VMs in a job) it would mean that all the snapshots created by the job on a given day will be considered as a single restore point - so each of the processed VMs will have a single restore point on one of the four snapshots created within the job run. The next job run will create 4 snapshots more and that would be a second restore point. Retention will come into action only when the fourth set of snapshots is created and every VM has four valid restore points avaiable.

[Matthias.Mehrtens]

Great, thanks for clarifying!

[EcoboostPerformance]

I think that there are a few more customers that would be very interested in specifying a number of Days in a future version.

USE CASE: So that You are able to take adhoc snapshots with out stomping on your original BACKUP SLA (Application teams will absolutely love this). This feature is availble in a few other backup products.

Also while we are at it how about the option to remove snapshots when you delete a job. As it will just cause orphans if you remove a job. As veeam will never go do autocleanup since its set to Number of snapshots rather than days..

[foggy]

Answered in the adjacent thread.