Large and Old Snapshots - Need to eliminate it

[Vaeryl]

I have a complicated problem here and I would like your help.

I started to manage an esxi 5.5 server. The machine has a HD of 1TB, and 2 critical VMs running on it (file server and the company's ERP).

The problem is that the old admin left snapshots on the machines for more than 2 years. Some have 8 snapshots, some of which already have 40 GB. I will need to delete this to avoid future problems, but I am afraid of having problems deleting them.

My question: What is the best practice to delete these snapshots? The server currently has 250GB of free space. I am afraid that by deleting such large snapshots, it can generate huge downtimes on critical servers.

Can someone help me?

[a_p_]

Welcome to the Community,

From what you explained, I actually don't see a real issue with deleting the snapshots.

The important part is to ensure that there's sufficient free disk space available for the snapshots to be merged with their base virtual disks. This is important if thin provisioned virtual disks are in use. With ESXi 5.5, open the datastore browser to see whether the base .vmdk file (the one without -00000x.vmdk) has a provisioned size that differs from the current size. In doubt, post a screenshot of the datastore browser screen which shows all the details.

If space is not an issue, then I'd suggest that you delete the snapshots out of business hours, and aout of backup windows. The best way to get rid of the snapshots is to use "Delete All" from the Snapshot Manager. If you do this with one VM after the other, this shouldn't impact work thhat much.

André

[berndweyand]

and be aware that vsphere 5.5 creates a helper snapshot to consolidate the others. so minimize traffic to the vms so the helper snapshot cannot grow to fast during consolidation - or better shutdown the vm if possible

[Vaeryl]

Andre, the issue is that I have already had major downtimes deleting considerably smaller snapshots on other servers. I fear that it will take many hours to erase everyone. My only window to stop this server is on Sundays. I would have no problem also to delete slowly, if it is safer.

Just to be clear, I have a basic understanding of how ESXi works. But this type of maintenance, on critical servers, still leaves me a little insecure.

I'm also sending a screenshot of the backup I took from one of the VMs. Note the size of the delta files, and the number of snapshots.

[scott28tt]

Consider also that 5.5 has been out of support for a while now, so it would not be advisable to continue to run such critical workloads on that platform.

[TheBobkin]

Hello Vaeryl​,

"I have already had major downtimes deleting considerably smaller snapshots on other servers"

Why would consolidation of snapshots require downtime? You know that you can do this while the VM is powered-on, right?

Fair enough, it will likely take less time to complete while powered-off (as the VM is doing nothing else), and you *should* aim to do this (powered-on consolidate) during non-peak hours if the storage isn't blazing fast and/or the VM is very busy.

To the folks mentioning needing extra space: I think you may be thinking of how this worked in way earlier versions - they should just require the space needed for the base-disk to expand to hold the snapshot data, though this isn't necessarily going to be the cumulative size of all the snapshots as they may have non-unique data.

Bob