Ghost VM after hardware node crash

[Joseph Maillardet]

Hello,

I am experiencing a problem after a disk failure of a server. On a cluster of 8 servers, the hard disk of one of them broke down.

I removed this machine from the cluster (pvecm delnode node_name). Since, this machine continues to appear in the web interface while it is no longer listed by: `pvecm nodes` or `pvecm status`.

$ sudo pvecm status
Quorum information
------------------
Date: Mon Jun 18 14:03:32 2018
Quorum provider: corosync_votequorum
Nodes: 6
Node ID: 0x00000008
Ring ID: 2/115076
Quorate: Yes

Votequorum information
----------------------
Expected votes: 7
Highest expected: 7
Total votes: 6
Quorum: 4
Flags: Quorate

Membership information
----------------------
Nodeid Votes Name
0x00000002 1 192.168.2.15
0x00000003 1 192.168.2.16
0x00000005 1 192.168.2.18
0x00000007 1 192.168.2.22
0x00000006 1 192.168.2.24
0x00000008 1 192.168.2.28 (local)​

$ sudo pvecm nodes

Membership information
----------------------
Nodeid Votes Name
2 1 192.168.2.15
3 1 192.168.2.16
5 1 192.168.2.18
7 1 debussy
6 1 192.168.2.24
8 1 berlioz (local)​

This ghost machine also displays a ghost VM that was on that node at the time of the crash :



How can I clean my cluster of this ghost ?

Regards,
JM

 

[devinacosta]

JM,

All you need to do is SSH into the host that it was running on and then delete it's configuration from this path:

If it's an LXC container, then you would just delete:
# rm /etc/pve/lxc/116.conf

If it's a QEMU image, then you would just delete:
# rm /etc/pve/qemu-server/116.conf

Then the GUI will update after a few seconds.

 

[Joseph Maillardet]

Thanks Devinacosta,

Finally, the problem was listed in /etc/pve/nodes, the « chopin » folder still exist with a lxc subfolder and a 116.conf file inside ! So I run :
# rm -r /etc/pve/nodes/chopin

And it seem to be all right.
Thanks again.