Replication issue

[itsathejoey]

I have been banging my head against my desk on this one lol. My LXCs will not replicate to one of the nodes in my cluster and all I get is this from syslog.

00-2: got unexpected replication job error - command '/usr/bin/ssh -e none -o 'BatchMode=yes' -o 'HostKeyAlias=proxnode4' root@192.168.3.4 -- pvesr prepare-local-job 100-2 ZFSNode1:subvol-100-disk-1 --last_sync 0' failed: exit code 255



I had this same issue on a different node and I was able to reload the OS on it to resolve and re format the ZFS drive, but in this instance however it isn't resolving. Have verified all nodes are on the same version.

 

[fiona]

Hi,
does ssh with HostKeyAlias work properly? You can check by running

/usr/bin/ssh -e none -o 'BatchMode=yes' -o 'HostKeyAlias=proxnode4' root@192.168.3.4 -- echo "test"

on a different node. Please also check the log for the replication job (go to the node's Replication tab, select the job and click on Log), which might contain more information.

 

[itsathejoey]

Yea it was SSH, but now when I cleared out the ssh hosts all my nodes are asking for password and now replication won't work to any of them.

 

[fiona]

You can use the command pvecm updatecerts -f to regenerate the host keys in the cluster. Needs to be run only on one node as long as you didn't change the fact that /root/.ssh/authorized_keys is a symlink to /etc/pve/priv/authorized_keys.

 

[itsathejoey]

So that helped a bit, but now going to one node I am getting this

 

[fiona]

Is it only a specific node? I'd try regenerating the certificate there by running pvecm updatecerts -f on that node.

 

[itsathejoey]

Well the error seemed to clear it self out because everything is working as normal now. Thanks for the help.