Hp lefthand - multiple snapshot filling up

[frankive]

We have a Hp Lefthand 4330 San.
Is it at 76% capacity.
We are now receiving to get alert that 95% of the capcity is used during backupprocess iwth Veeam 8 patch 1.
When I log on on to see the snapshot I can see there is up tp 4-5 snapshots at the same time.
Is this due to that Veeam try to use the same snapshot on multiple jobs? We have 1 job for each customers. Abouth 15 jobs with appx. 3 VMs in each. All Vms reside on the same lefthand SAN.

If this is the case, shouldnt 1 snapshot suffice to be running for all the jobs since they are on the same SAN? What we are seeing in the Veeam log is that is creates a hardware snapshot for each VM.

[Vitaliy S.]

Hi Frank,

As far as I remember, every job should create its own snapshot. How many jobs do you run concurrently? Also do you possibly observe a similar behavior as reported in this thread Lefthand Snapshots (free space reporting)?

Thanks!

[frankive]

We only use thin volumes on out HP Lefthand.
But I can see that it takes some time to delete the snapshots (often there are 5-6 snapshots in "deleting").

And each VM is creating a snapshot, not each job. Is that something that can be adjusted?

[Gostev]

Create fewer jobs with more VMs in them, or at least populate each job with VMs from the same datastore. Our jobs will process all VMs located on the same datastore from the single storage snapshot.

[frankive]

I did try this now and I dont see this behavior.
I created a job with 2 VMs, both being located on the same LUN and CusterStorage Volume.
The HP Lefthand Console (CMC) shows that it creates 2 snapshots.

I am running Hyper-V in our datacenter and we are not using parallel processing. Can this cause this behaviour?

[Vitaliy S.]

Ah....you're running on Hyper-V infrastructure. In order to have a single snapshot for VMs in the job, your VMs should reside on the same host and on the same volume, both conditions should be always true. Also why don't you use parallel processing option?

[frankive]

I can confirm that this did not working in my environment.
Vms located on same host, vcdx located on same ClusterStorage. The 2 Vms create 1 snapshot each.

I think I have not enabled because while I tested it on v7 it somehow caused one of our 2 nodes to crash during backup. Since we removed it we have been running successfully.

[Vitaliy S.]

What about VMs config files? Do they also reside on a CSV? Additionally, please confirm you have this checkbox enabled?

[frankive]

The config files also resides on the sam CSV. But you were right about that checkbox, it was not checked.
The test job now runs fine, and the job created just one hardware snapshot.

That said; I am not sure I have earned the "expert"-title in the forum for this post..

I will remember to have it checked and also create jobs with VMs on the same CSV

[Vitaliy S.]

This checkbox is enabled by default for all new jobs created in v8. If there was an upgrade of Veeam backup server, then you will need to enable this functionality manually. Glad that we have sorted this out.

[frankive]

Its an upgrade from V7 so that could explain it from my point of view. In other case, I could have change the settings in an earlyjob and saved the settings as default.

Hopefully this can be to help for others who are not to often in the settings.

Veeam: juset set it up and it works

[frankive]

If i changed from on-host and to a dedicated offhost hyper-v proxy (with direct access to SAN).
Is it still necessary in hyper-v that the vm resides on same host and datastore to create a single snapshot?
Does still the 8 VM in same job matter?

Or should I know create 2 jobs (1 for each CSV) and the offhost proxy will handle more than 8 VMs at same time?

[foggy]

Yes, the requirements are the same, however the max number of VMs is limited by eight per snapshot in case of hardware VSS.