Canceling a cold datastore migration

[SyApps]

Hello,

i have a powered down vm right now. I'm migration it to another data store. The drives around about 700GB total. If I reach my window for maintenance and need to bring the server back up, how long does it normally take once I cancel the move? Is it instant, will it take hours? Thanks in advance.

[abhilashhb]

I would suggest you to calculate and move it in the window than cancelling it in between the move.

I tried moving a 50GB machine from one datastore to another and it took around 5 sec to cancel the move. But i can't really tell you how long it will take in your case. If its a cold migration 700Gb would take less than an hour. But that again depends on the kind of storage you have and the connectivity.

[SyApps]

I cancelled it and it threw the error that it was cancelled by an admin instantly. I was 26% through it. There's nothing on the destination datastore and nothing held up. The only thing odd as of now is that VM is incredibly slow at the moment, although I did turn it on immediately afterwards. I see nothing wrong though, I'm going to reboot the VM for good measure and call it a night.

[abhilashhb]

Yes because you cancelled it there will be nothing that will be moved.

Yes may be a reboot will help Good Night

[SyApps]

Right, I wasn't sure if it would abandon the file's it created or what. Thanks for your help.

[grasshopper]

Hi SyApps,

IMHO, one should never cancel an action such as svMotion.  While it has gotten better since 4.x, any number of things can still go wrong.  If there is a risk of time over-run consider using an alternative method such as VMware Converter where you can seed the copy ahead of time and then do a final cut-over during your maintenance window.

Also, in general one should not move too fast when making big changes.  For example, don't cancel the task and then immediately power on the VM.  This could be asking for trouble.  Personally, I like to run /sbin/auto-backup.sh on any hosts involved to ensure that the ESXi config files have been properly updated after making significant changes (this runs by default once an hour anyway).  It sounds like you're okay this time, but If things had gotten really hosed up you may have needed to restart the vCenter or host services (both of which are absolute last resort tactics).

With that said I have also seen vCenter pick right up after doing some amazingly 'unwise' things like rebooting the VC while multiple svMotions are running (don't do this! lol).  I really should have run a 'Get-Task' from PowerCLI to ensure there was nothing running before rebooting VC but it happens.  Anyway, that's unrelated to your case but the point is VMware is surprisingly resilient, but in general try not to tempt fate :smileysilly:

Best of luck!