QA (public) » openQA Project (public)

Task 1. is actually tricky to implement because 'right before it goes down' it seems not possible anymore to broadcast one further message to all web socket clients.

Maybe it is better not to stop the command server process directly but to ask it to quit itself via an RPC call. That would allow the command server to extend its livetime to do the last broadcast. If the RPC call fails isotovideo can still stop the process forcefully.

There's the same problem when the job has just been started but the os-autoinst command server isn't available yet. Also in this case the error message shouldn't occur.

Since the error message is annoying and seems to confuse people, I raise the prio back to normal.

I've heard enough complaints about it so I'm adding it the the current sprint.

This PR provides a simpler solution than suggested in the initial ticket: https://github.com/os-autoinst/openQA/pull/1830

The PR just moves the disturbing messages into the developer panel and clears obsolete ones. This is not as nice as the initially proposed solution (which can still be implemented) but already a start.

Not sure whether it is good enough right now. So far I implemented didn't implement the initial idea but did some simpler improvements instead and at least solved the problem mentioned in my 2nd comment. That should at least lower the priority.

With https://progress.opensuse.org/issues/45191 we have our feedback. So this error message should be prevented.