New exception in the log

[dcardin]

Latest version, 0.9.12.01

ERROR:web_admin UNEXPECTED EXCEPTION on GET='NoneType' object has no attribute 'instances' Traceback (most recent call last): File "/app/lib/clients/web_admin.py", line 105, in do_GET elif getrequest.call_url(self, self.content_path): File "/app/lib/common/decorators.py", line 310, in call_url self.url2func[_name](_webserver, *args, **kwargs) File "/app/lib/clients/channels/channels.py", line 42, in channels_m3u _webserver.do_mime_response(200, 'audio/x-mpegurl', get_channels_m3u( File "/app/lib/clients/channels/channels.py", line 91, in get_channels_m3u if not _plugins[sid_data['namespace']] \

[rocky4546]

I think we are going to need a little more info. What we know is...

• You requested a m3u file from cabernet. The plugin under request somehow was not initialized.

One possible suspect is that this occurred very early after the app was started. So, will need to know when in the sequence of starting that this happened. No other possibilities come to mind. If this is the case, I can make a change to fix it, but need to understand what caused it.

Another issue could be running a docker image within a Windows OS. Let me know if this is true... I assume this is a new issue...

[rocky4546]

After a few tests, it does not look like an initial startup issue. It does somehow look like a Windows OS issue when the operating system from python is not a Windows OS. This is because Windows does not fork processes correctly, so python has something called pickling to address the shortcoming. BUT... pickling only happens when the OS is Windows. Is it possible to run python command from inside the docker and do the following?

import platform
platform.system()

Note: you can also go to Settings > Internal > Main and find OS Type. that is what Cabernet sees as the OS.

[dcardin]

Yes this is running on a linux machine, in a docker container. So the platform reported (python and web interface) is Linux (obviously) version: #39~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Fri Mar 17 21:16:15 UTC 2 Thanks!

[rocky4546]

I run on Ubuntu 20. This looks like a case where Ubuntu did not start the processes correctly. The solution would be to restart the app. However, if you can indicate the steps needed to reproduce the error and have a debug log from time the app starts to when you send in the m3u request and it fails, that would be what is needed to resolve this issue. If indeed it is a 1 in a billion case that Ubuntu messed up and you cannot reproduce the error, then I will close this as non-discrepant and try a simple band-ade fix that won't really resolve the overall issue.

[dcardin]

I have restarted the container and it is working. I have changed the logging level to debug and will raise a new issue if I can provide more info. Thanks!

[rocky4546]

Question. Did you delete a plugin and then run the channels.m3u? Although unable to repo the exact issue, I have found a few things associated with that. Fixing those issues and hoping that may help yours on next release.

[dcardin]

I might have. I had the tvhebdo epg plugin but I was only getting channel info, not an actual list of shows, so I removed it. Think that might have been a cause ?