live migration failure "Host key verification failed."

[joulester]

Hi! I made a new install with version 5.0 on 2 nodes I made att separate cluster using this wiki https://pve.proxmox.com/wiki/Separate_Cluster_Network

no issues connecting each node with ssh to each other

to my problem I cant migrate a vm to the other node

2017-07-09 10:24:51 # /usr/bin/ssh -o 'BatchMode=yes' -o 'HostKeyAlias=R210II' root@192.168.3.201 /bin/true
2017-07-09 10:24:51 Host key verification failed.
2017-07-09 10:24:51 ERROR: migration aborted (duration 00:00:00): Can't connect to destination address using public key
TASK ERROR: migration aborted​

I have tried updating the "pvecm updatecerts"
but with no luck. This worked at proxmox 4.4 last time

What should I do next?

 

[dietmar]

This bug triggers if you have host names using capital letters. We already found the bug (patches are on the developers list), and we will try to rollout a fix next week.

 

[tomtom13]

fyi I've got same problem and I've got NO capital leters in host names (capital leters in host names don't exist anyway so why somebody would put them there)

Edit:
in my case for some reason /etc/ssh/ssh_known_hosts was NOT a symlink to /etc/pve/priv/known_hosts as it should.

 

[joulester]

Hi I still have a problem with the live migration!

2017-08-03 16:51:00 starting migration of VM 100 to node 'R710' (192.168.3.200)
2017-08-03 16:51:00 found local disk 'local:100/vm-100-disk-1.qcow2' (in current VM config)
2017-08-03 16:51:00 can't migrate local disk 'local:100/vm-100-disk-1.qcow2': can't live migrate attached local disks without with-local-disks option
2017-08-03 16:51:00 ERROR: Failed to sync data - can't migrate VM - check log
2017-08-03 16:51:00 aborting phase 1 - cleanup resources
2017-08-03 16:51:00 ERROR: migration aborted (duration 00:00:00): Failed to sync data - can't migrate VM - check log
TASK ERROR: migration aborted

 

[GadgetPig]

See if this applies to your situation:

https://forum.proxmox.com/threads/live-migration-is-now-possible-with-local-zfs.35972/#post-176278

 

[commgdog]

This bug triggers if you have host names using capital letters. We already found the bug (patches are on the developers list), and we will try to rollout a fix next week.

Did this every get patched? My host names use capital letters and I am unable to migrate (Same error as OP).

 

[commgdog]

I figured out a work around until there is a fix for this.

• Log onto the node you want to transfer away from via SSH
• Execute: /usr/bin/ssh -o 'HostKeyAlias=NODE_NAME' root@x.x.x.x /bin/true
• Where x.x.x.x is the IP of the node you want to transfer to and NODE_NAME is your node name (upper cased)

You will get a popup asking for authenticity of the host. Say yes and now you will be able to migrate.

The above command is the same one sent via the GUI when it try's to migrate and errors out, minus the option 'BatchMode=yes'.