ceph warning post upgrade to v8

[Just2Bad]

Hello tonight I upgraded all my hosts to V8 wthout any issue except this warning on Ceph

"Module 'dashboard' has failed dependency: PyO3 modules may only be initialized once per interpreter process"

 

[Just2Bad]

I restarted all host after upgrade, also restarted all manager and monitor services on all hosts but warning still there, haven't tried to destroy and reinstall

any idea ?

 

[jsterr]

I restarted all host after upgrade, also restarted all manager and monitor services on all hosts but warning still there, haven't tried to destroy and reinstall

any idea ?

Nope. Dashboard is nothing shipped by default, so this is something that might not get fixed that soon. What are you reasons for using ceph dashboard, as it does not really work with proxmox?

 

[Maximiliano]

Hello, did you follow the 7to8 guide [1]? specifically upgrading Ceph before PVE?

[1] https://pve.proxmox.com/wiki/Upgrade_from_7_to_8

 

[jsterr]

Hello, did you follow the 7to8 guide [1]? specifically upgrading Ceph before PVE?

[1] https://pve.proxmox.com/wiki/Upgrade_from_7_to_8

I also did and had the same error so not related to the upgrade process (regarding doing something wrong)

 

[martin_fryc]

Hello tonight I upgraded all my hosts to V8 wthout any issue except this warning on Ceph

"Module 'dashboard' has failed dependency: PyO3 modules may only be initialized once per interpreter process"

View attachment 52038

Hi,

After the update I have the same problem on the second and third node of our three node cluster. The problem only manifests itself when the Manager becomes active on these nodes. I tried deleting and recreating the Manager on all nodes, but it didn't help. When I activate the Manager on the first node, everything works fine.

 

[martin_fryc]

Hi,

After the update I have the same problem on the second and third node of our three node cluster. The problem only manifests itself when the Manager becomes active on these nodes. I tried deleting and recreating the Manager on all nodes, but it didn't help. When I activate the Manager on the first node, everything works fine.

I just found out that in my case it is not the "Dashboard" module, but the "restful" module. I apologize for the inaccuracy.

 

[jsterr]

I just found out that in my case it is not the "Dashboard" module, but the "restful" module. I apologize for the inaccuracy.

View attachment 52073

Nope your right, the errors switches from time to time, depending on which module is currently initalized I guess.

 

[spirit]

I can confirm this, seem to be a change in new python py03 0.17 lib, and dashboard code does not support it yet.

https://github.com/PyO3/pyo3/blob/7...-can-now-only-be-initialized-once-per-process

Each #[pymodule] can now only be initialized once per process​

To make PyO3 modules sound in the presence of Python sub-interpreters, for now it has been necessary to explicitly disable the ability to initialize a #[pymodule] more than once in the same process. Attempting to do this will now raise an ImportError.

 

[boomam]

Same issue for me, Ceph dashboard on :8443 errors with "No active ceph-mgr instance is currently running the dashboard. A failover may be in progress. Retrying in 5 seconds..." for me too.

 

[spirit]

I have send a report to ceph-users mailing list

 

[boomam]

I have send a report to ceph-users mailing list

Great!
Do you mind linking, so we can keep an eye on progress directly?

 

[Platup]

+1 I'm seeing the same "restful" module failing after the upgrade.

 

[lordjakoby]

+1 I'm seeing the same "restful" module failing after the upgrade.

+1 here as well

 

[phastier]

+1 same warning.

Does seem fully functional though.

 

[kafisatz]

Nope. Dashboard is nothing shipped by default, so this is something that might not get fixed that soon. What are you reasons for using ceph dashboard, as it does not really work with proxmox?

Hi
I am running into the same issue. I don't recall when/how I installed the dashboard. Do you know how I can uninstall it? Maybe this will get rid of the warning message.

 

[yogisha]

Hi
I am running into the same issue. I don't recall when/how I installed the dashboard. Do you know how I can uninstall it? Maybe this will get rid of the warning message.

You can simply disable the dashboard and it should get rid of the warning sign.

 

[kafisatz]

You can simply disable the dashboard and it should get rid of the warning sign.

But how can I disable it? I know nothing at all about ceph (except how to set it up via gui)

 

[yogisha]

But how can I disable it? I know nothing at all about ceph (except how to set it up via gui)

Run this on the node where you installed the dashboard:

ceph mgr module disable dashboard

and wait for a bit. The warning should disappear

 

[spirit]

I have report to ceph-users mailing, still no response

https://lists.ceph.io/hyperkitty/list/ceph-users@ceph.io/message/FB7XE6WYDK3EBJYPABSPX5B2LEILWWJA/