Proxmox cluster migrate vm or ct copies the storage twice

[xvlvx]

Hi All,

I'm migrating my proxmox server to a new bigger server (more storage and compute power).

As a precaution I've setup cluster and added a third 'temporary' server to help migrating if things go south!
The migration was performed from old -> temp -> new, all within the cluster.

When migrating from old -> temp everything went beautiful. No issues (other then some usb-device passthroughs I needed to remove before migrating).

When migrating from temp -> new things didn't went 'wrong', but all storage volumes where copied twice.
See the log output below:

root@proxmox-temp:~# pct migrate 205 proxmox-new
2022-10-27 08:01:13 starting migration of CT 205 to node 'proxmox-new' (10.x.x.xxx)
2022-10-27 08:01:14 found local volume 'LocalStorage-1:vm-205-disk-0' (in current VM config)
2022-10-27 08:01:14 found local volume 'localstorage1:vm-205-disk-0' (via storage)
2022-10-27 08:01:15   Wiping ext4 signature on /dev/localstorage1/vm-205-disk-0.
2022-10-27 08:01:15   Logical volume "vm-205-disk-0" created.
2022-10-27 08:05:10 131072+0 records in
2022-10-27 08:05:10 131072+0 records out
2022-10-27 08:05:10 8589934592 bytes (8.6 GB, 8.0 GiB) copied, 234.387 s, 36.6 MB/s
2022-10-27 08:05:11 47+499960 records in
2022-10-27 08:05:11 47+499960 records out
2022-10-27 08:05:11 8589934592 bytes (8.6 GB, 8.0 GiB) copied, 235.526 s, 36.5 MB/s
2022-10-27 08:05:11 successfully imported 'localstorage1:vm-205-disk-0'
2022-10-27 08:05:11 volume 'localstorage1:vm-205-disk-0' is 'localstorage1:vm-205-disk-0' on the target
2022-10-27 08:05:13 volume localstorage1/vm-205-disk-0 already exists - importing with a different name
2022-10-27 08:05:13   Wiping ext4 signature on /dev/localstorage1/vm-205-disk-1.
2022-10-27 08:05:13   Logical volume "vm-205-disk-1" created.
2022-10-27 08:09:07 131072+0 records in
2022-10-27 08:09:07 131072+0 records out
2022-10-27 08:09:07 8589934592 bytes (8.6 GB, 8.0 GiB) copied, 234.015 s, 36.7 MB/s
2022-10-27 08:09:09 51+502827 records in
2022-10-27 08:09:09 51+502827 records out
2022-10-27 08:09:09 8589934592 bytes (8.6 GB, 8.0 GiB) copied, 236.196 s, 36.4 MB/s
2022-10-27 08:09:09 successfully imported 'LocalStorage-1:vm-205-disk-1'
2022-10-27 08:09:09 volume 'LocalStorage-1:vm-205-disk-0' is 'LocalStorage-1:vm-205-disk-1' on the target
2022-10-27 08:09:09 # /usr/bin/ssh -e none -o 'BatchMode=yes' -o 'HostKeyAlias=proxmox-new' root@10.x.x.xxx pvesr set-state 205 \''{}'\'
  Logical volume "vm-205-disk-0" successfully removed
2022-10-27 08:09:11 ERROR: removing local copy of 'LocalStorage-1:vm-205-disk-0' failed - can't activate LV 'localstorage1/vm-205-disk-0' to zero-out its data:   Failed to find logical volume "localstorage1/vm-205-disk-0"
2022-10-27 08:09:11 start final cleanup
2022-10-27 08:09:12 ERROR: migration finished with problems (duration 00:07:59)

Did anyone see this behavior before? Does anyone know what is causing it?

I've tried both the web-gui as the cli to perform the migration. Both provided the same results.

Best regards,

LVX

 

[dcsapak]

it seems you have configured the same storage twice? 'LocalStorage-1' and 'localstorage1'. such a configuration can lead to these problems, best to only configure each storage exactly once
the log shows you :

2022-10-27 08:01:14 found local volume 'LocalStorage-1:vm-205-disk-0' (in current VM config)
2022-10-27 08:01:14 found local volume 'localstorage1:vm-205-disk-0' (via storage)

that it found the 'second' disk image via the storage
(it must always copy *all* related disks, even those not in the config, otherwise it'll be very easy to get stray disk images)

 

[xvlvx]

Hi dcsapak,

Wow, nicely spotted!!! I didn't see that!
And indeed, after checking the '/etc/pve/storage.cfg' they where both in there, pointing to the same VolumeGroup.

After deleting the duplicate lines everything seems okay now.

Best regards,
LVX