Ceph Dashboard not working after update to proxmox 7 (from 6.4)

[fxandrei]

So ive just updated a cluster from proxmox 6.4 to 7.1 (along with ceph from 14 to 16).
Everything works fine except one thing, and that is the dashboard.

If i try to enable it i get:

ceph mgr module enable dashboard
Error ENOENT: all mgr daemons do not support module 'dashboard', pass --force to force enablement

If i try to disable it, i get:
ceph mgr module disable dashboard
module 'dashboard' is already disabled

What could it be ? What could i check ?

EDIT: Forgot to mention. It was working before the update. Im not sure what happened. The package is installed and on the last version available.

 

[fxandrei]

So it seems there was a problem with the managers. I restarted them, and the passed "--force" to the enable command, and now the dashboard is working.

 

[Mael Strom]

Same issue on 7.2.7
Does it work correct after forcing?

 

[kwslavens]

Same issue on 7.2.7
Does it work correct after forcing?

I needed to run "apt install ceph-mgr-dashboard" on all nodes in the cluster to get rid of the "Error ENOENT: all mgr daemons do not support module 'dashboard', pass --force to force enablement" error message.
After the install then I could enable with out any issue.

 

[Mael Strom]

Meaningful note! Thank you.
And very disguising message, but it's the way life is…

 

[Ting]

Hi, everyone;

I have a 8 nodes cluster running for a while, and I have ceph on the same cluster. Yesterday, I tried to install and enable ceph dashboard, then caused my ceph manager service failed to restart and all manager service stopped and can not start, here is what I did step by step, and also I attached veversion -v" and "ceph.conf" file:

1. apt install ceph-mgr-dashboard (on all service manager nodes, I have 3 manager services, installed on all 3 of them) = ok
2. then, on #2 node(has manager service), i did: ceph mgr module enable dashboard --force. (I have to use --force, otherwise, will no enable)
3. on #2 node, ceph config-key set mgr/dashboard/server_addr ::
4. on #2 node, ceph mgr module disable dashboard
5. on #2 node, ceph mgr module enable dashboard
6. on #2 node, systemctl restart ceph-mgr@[servername].service. Now, here is the my problem, manager service failed to restart, and all rest of manager service shows stopped, and I can not restart, give me error message and exit code 1.
7. I have to destory all ceph managers, then re-create them in proxmox gui. Ceph running ok at all time.

I was not sure what did I do wrong, Any help would be much appreciated.

https://forum.proxmox.com/attachments/pveversion-txt.44432/

https://forum.proxmox.com/attachments/cephconf-txt.44433/

 

[7thSon]

I just ran into the same prollem.
After doing a

systemctl reset-failed ceph-mgr@%YOUR-NODE-NAME-HERE%.service

i was able to start the managers again

I'm on Ceph pacific 16.2.9 & Proxmox 7.3-3

 

[Mael Strom]

Hello there.
Despite forcing dashboard works (and, on clean pve, it works without forcing along as you equip all mgr's with dashboard via apt), I'll ended up with total mess in pve's ceph and unability to create even OSD's.
So, I handed a complete removal again and used ceph-deploy without any difficulties.

But beware of making cleanup with ceph-deploy, it may ruin pve, because some ceph (client) packages are dependency and removal is done by `ceph*`!

 

[Deepen Dhulla]

Can not install -- apt install ceph-mgr-dashboard get error.
The following packages have unmet dependencies:
ceph-mgr-dashboard : Depends: ceph-mgr (= 16.2.11-pve1) but 17.2.5-pve1 is to be installed

pveversion
pve-manager/7.4-3/9002ab8a (running kernel: 5.15.102-1-pve)

ceph -v
ceph version 17.2.5 (e04241aa9b639588fa6c864845287d2824cb6b55) quincy (stable)


----
OK Got it fixed ..I dont know what changed the repo back to pacific.

Changed it.

sed -i 's/pacific/quincy/' /etc/apt/sources.list.d/ceph.list

apt-get update

now install works.