Consolidating Snapshot ESXi 3.5 (delete)

[rucky]

Hello pals.

I have acquited a virtual server with 3 disks, which was snapshotted around 8 months ago. and since then the server has been running from these delta disks and obviously they have grown. Now almost every month we have a corrupt redo log (msg.hbacommom ms) (loop) as some of you might have come across before/.

Because of which we have to reboot the host, as this is a esxi3.5 and we cant hack into the service console (well we can) but it wont be supported then i guess.

It seems that these delta disks are causing the server to crash almost every month. As far as iam aware its not a good practice to leave the delta disks ruunning this long.

Therefore i would like to consolidate (delete) these snapshots, previously i have done it using 'delete all' for esx3.5, if that dint work, got vmware support to do it through SC (which is what they recommded)

can i follow the same procedure for esxi 3.5, ,,, ,, ,,, do i need to be careful about any things. Is it advisable to take a backup before this or you guys think it will be a smooth sailing cosolidating a 8 month snapshot.

I would like to your views on this

Thanks.

[AntonVZhbankov]

Delete snapshots then one-by-one.

---

MCSA, MCTS Hyper-V, VCP 3/4, VMware vExpert '2009

http://blog.vadmin.ru

[philvirt]

In vCetner, Right Click on VM and select snapshot manager and then delete each one manually.

If this fails, you can delete them manually via the ESX console.

Thanks,

Phil Couto

[rucky]

this is esx3.5 i... dont want to hack in the service console,. as vmware wuldnt support it then in case we need support ever//

which order do i delete them in ?

in this case i have only got one though.

Thnanks

[NTurnbull]

Hi, if commiting/deleting large snapshots from your ESXi host manged via VC then sometimes you can commit/delete a snapshot and you'll get a timeout or error message and your snaps will not show up in the snapshot manager. The snap is actually still being commited, it's just an operation timeout in VC. If you connect directly to your ESXi host from your VI client and comit the snap you will not get the same timeout problem - or if after the timeout message, connect directly and you'll see the snap commit operation still going through. If you do get a problem whereby there are no snaps showing in the snapshot manager and you can see delta files in the filesystem then a little trick you can do is to create another snap and then commit that and ESX will actually commit all the snaps in sequence. Each nested snap has a chain ID that relates to it's parent, so when you create the first snap the parent ID is the base disk and then the second snap's parent ID would be the first snap's ID and so on - when you create a new snapit checks for these ID's in order to keep this chain ID health.

Thanks,

Neil

[a_p_]

which order do i delete them in ?

Like Anton said "Delete snapshots then one-by-one". If you have multiple snapshots and you are low on diskspace on your datastore, always delete the first snapshot following the base image and wait until the snapshot file is deleted (you can monitor this through the datastore browser) before starting to delete the next snapshot. (see Neil's post about the timeout message).

André