permission Denied on backup VM

[Spyder13337]

everything i try to back up my VM i get this msg
1. i remove a cluster
2. i can see the cluster in the gui
3. if i use pvecm status i get

root@proxmox1:~# pvecm status
Cannot initialize CMAP service

PLease let me know if there is any information that you need to resolve this
i am on PVE 4.0

did i remove the cluster incorrectly i just need to back up my VM and re-install with 5.0

Task viewer: Backup


INFO: starting new backup job: vzdump 113 --storage backup --node proxmox1 --mode stop --remove 0
INFO: Starting Backup of VM 113 (qemu)
INFO: status = running
INFO: update VM 113: -lock backup
INFO: unable to open file '/etc/pve/nodes/proxmox1/qemu-server/113.conf.tmp.12356' - Permission denied
ERROR: Backup of VM 113 failed - command 'qm set 113 --lock backup' failed: exit code 2
INFO: Backup job finished with errors
TASK ERROR: job errors

 

[t.lamprecht]

did i remove the cluster incorrectly

Yes it seems so. How did you do it?

It seems that corosync doesn't runs anymore but pmxcfs still is in the state where it is in a cluster and needs corosync to communicate.

You should get a OK node again with the first steps from:
https://pve.proxmox.com/pve-docs/chapter-pvecm.html#pvecm_separate_node_without_reinstall
Read whole section first. The pvecm delnode won't be needed by you, I assume.

2. i can see the cluster in the gui

You mean you can still see the deleted note? That normally means that you need to do a full reload of the page CTRL+R or SHIFT+F5 or something like that.
Else, if that does not help, there are still VM configs of it in the cluster file system.
I.e. in /etc/pve/nodes/REMOVED-NODE/qemu-server (and /lxc). If you're sure you got the right node, backed every config up you can delete the whole folder from the removed node.

 

[Spyder13337]

Yes it seems so. How did you do it?

It seems that corosync doesn't runs anymore but pmxcfs still is in the state where it is in a cluster and needs corosync to communicate.

You should get a OK node again with the first steps from:
https://pve.proxmox.com/pve-docs/chapter-pvecm.html#pvecm_separate_node_without_reinstall
Read whole section first. The pvecm delnode won't be needed by you, I assume.



You mean you can still see the deleted note? That normally means that you need to do a full reload of the page CTRL+R or SHIFT+F5 or something like that.
Else, if that does not help, there are still VM configs of it in the cluster file system.
I.e. in /etc/pve/nodes/REMOVED-NODE/qemu-server (and /lxc). If you're sure you got the right node, backed every config up you can delete the whole folder from the removed node.

if i try to delete it it said permission denied

 

[Spyder13337]

Yes it seems so. How did you do it?

It seems that corosync doesn't runs anymore but pmxcfs still is in the state where it is in a cluster and needs corosync to communicate.

You should get a OK node again with the first steps from:
https://pve.proxmox.com/pve-docs/chapter-pvecm.html#pvecm_separate_node_without_reinstall
Read whole section first. The pvecm delnode won't be needed by you, I assume.



You mean you can still see the deleted note? That normally means that you need to do a full reload of the page CTRL+R or SHIFT+F5 or something like that.
Else, if that does not help, there are still VM configs of it in the cluster file system.
I.e. in /etc/pve/nodes/REMOVED-NODE/qemu-server (and /lxc). If you're sure you got the right node, backed every config up you can delete the whole folder from the removed node.

i get this error

root@proxmox1:/etc/corosync# pmxcfs -l
[main] notice: unable to aquire pmxcfs lock - trying again
[main] crit: unable to aquire pmxcfs lock: Resource temporarily unavailable
[main] notice: exit proxmox configuration filesystem (-1)

 

[t.lamprecht]

[main] crit: unable to aquire pmxcfs lock: Resource temporarily unavailable
[main] notice: exit proxmox configuration filesystem (-1)

Did you stop pve-cluster as described in the docs?
If that does not works and you did not started pmxcfs manually previously the remaining lock means that the either the node or pmxcfs died unexpectedly.
Both do not speak for a gracefull and successful cluster separation...

so this question is still open:

How did you do it?

If you do not tell me I do not know whats the real problem is...

If you are sure that pmxcfs does not run, here's the lockfile path do delete: /var/lib/pve-cluster/.pmxcfs.lockfile

 

[Spyder13337]

pvecm delnode proxmox2
also i remove it from /etc/host

 

[Spyder13337]

i follow the instruction then a we loss power for a hour so we shutdown the servers once it came back online issue was fixed

so thanks

 

[t.lamprecht]

pvecm delnode proxmox2
also i remove it from /etc/host

Ok, what happened to the other node (proxmox2)? Did it stayed offline or did it get online, still meaning that it was in the cluster?
Anyways glad you fixed it...

a we loss power for a hour

that could explain the leftover look.

 

[Spyder13337]

even after the reboot it was still there proxmox 2 on proxmox 1 VM

proxmox 2 is fine never had any issue issue was on proxmox 1 with

 

[t.lamprecht]

even after the reboot it was still there proxmox 2 on proxmox 1 VM

You checked that:

Else, if that does not help, there are still VM configs of it in the cluster file system.
I.e. in /etc/pve/nodes/REMOVED-NODE/qemu-server (and /lxc). If you're sure you got the right node, backed every config up you can delete the whole folder from the removed node.

If you have the `tree` tool installed you could check with:

tree -P *.conf /etc/pve/nodes

Or simply with ls:

ls -l /etc/pve/nodes/*/*/*.conf

When there are still config files around from the node which you removed the web interface will still show this node, as it wants to map the config ids somewhere...