Thank you! The logs did help but I had to add log_to_file (as true) in the ceph.conf in
order to see it.
The plugin was successfully installed and did run as expected and could produce output as
desired and implemented in the code. However, I tried to write a new function, but
unfortunately it produced an error, but when I restored the working code I could not
enable the plugin again. Now it says: " Error ENOENT: all mgr daemons do not support
module 'newplugin', pass --force to force enablement".
I tried restarting the ceph-mgr, and made sure to do all the steps I did before, but seems
that it's still stuck on this error.
Do you have any ideas of what might help? Could this be a bug in ceph?
BTW I am using nautilus
-----Original Message-----
From: Willem Jan Withagen <wjw(a)digiware.nl>
Sent: Thursday, January 9, 2020 1:41 PM
To: Juan Miguel Olmo Martinez <jolmomar(a)redhat.com>om>; Khoury, Rima (Nokia - IL/Kfar
Sava) <rima.khoury(a)nokia.com>
Cc: dev(a)ceph.io
Subject: Re: New Ceph Plugin - assistance
On 9-1-2020 09:01, Juan Miguel Olmo Martinez wrote:
Hi,
Depending of the Ceph version you can have the files in different
places, but in any case is easy to search the file "module.py" in any
of your manager hosts, the results will show you the location of the
manager plugins.
In Nautilus they are in "/usr/share/ceph/mgr/"
You can see the active manager using the command "ceph -s"
A good source of information to see what is happening is the manager
log file (in var/log/ceph) in the manager host,
You can change the log level to debug using:
ceph daemon <your-mgr> config set debug_mgr 20/5 restore the log
level:
ceph daemon <your-mgr> config set debug_mgr 1/5
I had a lot of trouble getting plugins to work, since they needed py-libs that I did not
(yet) have.
Python tracebacks are logged in the ceph-mgr log file.
And there is was usually quite fast clear what the problem was.
Big chance that ceph-mgr is not loading you plugin due to loading errors but the traceback
in the log should give you more info.
--WjW