cluster node out and cannot be restarted. how to take posession of its assets

[alexskysilk]

I have a node that was shut down properly but will no longer boot. It will be a while until I can get to it to fix but in the meantime, how do I take over its assets (VMs and CTs) that are now showing as ? status in the defunct node?

 

[fabian]

I have a node that was shut down properly but will no longer boot. It will be a while until I can get to it to fix but in the meantime, how do I take over its assets (VMs and CTs) that are now showing as ? status in the defunct node?

if they don't reference any resources (local disks, physical hardware) that are specific to the node, and you are 100% sure that the node is completely down, you can just use mv to move the config file from the broken node's directory in /etc/pve/ to the target node's directory.

 

[alexskysilk]

Thanks @fabian. that did the trick.

I am curious though. this node came down after update but all its assets remain in ? state; why did that happen? its as if the node needs to be fenced but isnt... pvecm status shows its gone, but its assets are inaccessible by the cluster

 

[dcsapak]

I am curious though. this node came down after update but all its assets remain in ? state; why did that happen? its as if the node needs to be fenced but isnt... pvecm status shows its gone, but its assets are inaccessible by the cluster

are the vms declared as ha? if not, the behaviour is expected, as the semantic here is that only the node where the vms are have control over them, 'stealing' them can only be done by the ha-manager
under certain circumstances

as for fencing: this only means that the host to be fenced will be resetted, this alone does not do anything to the vms