Remote test session paused indefinitely when session was disconnected
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Next Generation Checkbox (CLI) |
Expired
|
Low
|
Maciej Kisielewski |
Bug Description
Honestly, I have no idea how to reproduce this. I discovered it while looking into this bug against testflinger:
https:/
In that bug, note the second disconnect example for Drapion. In that case, testflinger showed that there was a disconnection and then it was never able to reconnect, and eventually the testflinger instance exited.
Now in the course of investigating that, I got the idea to just try connecting manually to the SUT via checkbox remote myself. I had expected to find that the test session had completed silently in the background. What I found instead, was that the test session (which is fully automated, including submission to C3) had paused and only resumed testing after I manually connected to the session from my own desktop.
As the remote sessions are supposed to continue running headless, I'm a bit baffled as to why this one paused execution and waited for me to manually reconnect.
Also note, despite what the error messages in the above bug say, I did NOT, at any time, connect to the checkbox session from anywhere else. I merely fired off testflinger and walked away to wait for them to complete. I only connected to the session a couple days after the fact, expecting to find it already completed.
tags: | added: hwcert-server |
Changed in checkbox-ng: | |
assignee: | nobody → Maciej Kisielewski (kissiel) |
importance: | Undecided → Medium |
Changed in checkbox-ng: | |
status: | New → Triaged |
importance: | Medium → Low |
Bug was migrated to GitHub: https:/ /github. com/canonical/ checkbox/ issues/ 150.
Bug is no more monitored here.