Location for "snapshots with memory"
-
Where is the memory part stored when doing "snapshot with memory"?
It seems to take a very long time. The VM is stored in a local SR so it ought to go a lot faster than what it currently looks like:
EDIT: It looks like it might use the "Suspend SR" from the pool settings. This is unexpected if this is the case.
EDIT 2: the time said 8 hours so I restarted the toolstack, since the operation was not interruptable. The VM immediatedly crashed into offline state.
I found the suspend image on a remote SR:
Is there some automation to clear out old suspend images like this? I tried a SR rescan and rebooted the host (which is also master), but it remains.
-
Hi,
- The suspend SR is the place to store the virtual drive containing the memory.
- The VM is paused and all the RAM is copied to the storage. Depending on how much RAM you have it can take some time.
How much RAM your VM had?
-
@olivierlambert said in Location for "snapshots with memory":
Hi,
- The suspend SR is the place to store the virtual drive containing the memory.
- The VM is paused and all the RAM is copied to the storage. Depending on how much RAM you have it can take some time.
How much RAM your VM had?
It has 32GB RAM, 96GB disk. But as you usee the suspend image was 40GB when I aborted it. At that time it had been going for 20 minutes (with 8h estimated to go) trying to save it to our NFS SR (which is 25Gbit/s LACP).
We're using XCP-ng 8.3 and XOA on stable channel, both fully up to date.
-
It's indeed pretty long. I would maybe use a local SR as suspend SR to check if it's very different, that could help to pinpoint the issue.
