You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug
The action runner stopped on its own. When I check syslog I see this:
Nov 4 02:30:47 ip-172-31-21-174 run.sh: √ Connected to GitHub
Nov 4 02:30:47 ip-172-31-21-174 run.sh: 2020-11-04 02:30:47Z: Listening for Jobs
Nov 4 02:30:49 ip-172-31-21-174 run.sh: Runner update in progress, do not shutdown runner.
Nov 4 02:30:50 ip-172-31-21-174 run.sh: Downloading 2.273.6 runner
Nov 4 02:30:56 ip-172-31-21-174 run.sh: Waiting for current job finish running.
Nov 4 02:30:56 ip-172-31-21-174 run.sh: Generate and execute update script.
Nov 4 02:30:56 ip-172-31-21-174 run.sh: Runner will exit shortly for update, should back online within 10 seconds.
Nov 4 02:30:56 ip-172-31-21-174 systemd: github_action_runner.service: main process exited, code=exited, status=3/NOTIMPLEMENTED
Nov 4 02:30:56 ip-172-31-21-174 systemd: Unit github_action_runner.service entered failed state.
Nov 4 02:30:56 ip-172-31-21-174 systemd: github_action_runner.service failed.
Restarting produces the same log
Reinstalling manually fixes it.
This has been happening to us in some form every couple weeks for several months now. It would be great if we could opt out of the updates or have a more stable updating paradigm. From my perspective the updates bring no new features or functionality that I even know about or care about, but they do cause us downtime for our build and deploy systems.
I think at this point where we have gone down maybe 7 or 8 times now? It would be nice to get some visibility into what is being done to make this paid offering more resilient. IE have you considered the updater doing a blue green deploy and leave us with a working "old version" if the new one has issues? How about just using a daemon watchdog to watch the process. If it dies like it does now, start it back up for me and allow the previous version to continue working vs getting stuck in an update failure loop
The text was updated successfully, but these errors were encountered:
We don't see regular problems with the updates - so if you're seeing problems, we'd need to see the diagnostic logs on the runner. Can you share those?
Describe the bug
The action runner stopped on its own. When I check syslog I see this:
Nov 4 02:30:47 ip-172-31-21-174 run.sh: √ Connected to GitHub
Nov 4 02:30:47 ip-172-31-21-174 run.sh: 2020-11-04 02:30:47Z: Listening for Jobs
Nov 4 02:30:49 ip-172-31-21-174 run.sh: Runner update in progress, do not shutdown runner.
Nov 4 02:30:50 ip-172-31-21-174 run.sh: Downloading 2.273.6 runner
Nov 4 02:30:56 ip-172-31-21-174 run.sh: Waiting for current job finish running.
Nov 4 02:30:56 ip-172-31-21-174 run.sh: Generate and execute update script.
Nov 4 02:30:56 ip-172-31-21-174 run.sh: Runner will exit shortly for update, should back online within 10 seconds.
Nov 4 02:30:56 ip-172-31-21-174 systemd: github_action_runner.service: main process exited, code=exited, status=3/NOTIMPLEMENTED
Nov 4 02:30:56 ip-172-31-21-174 systemd: Unit github_action_runner.service entered failed state.
Nov 4 02:30:56 ip-172-31-21-174 systemd: github_action_runner.service failed.
Restarting produces the same log
Reinstalling manually fixes it.
This has been happening to us in some form every couple weeks for several months now. It would be great if we could opt out of the updates or have a more stable updating paradigm. From my perspective the updates bring no new features or functionality that I even know about or care about, but they do cause us downtime for our build and deploy systems.
I think at this point where we have gone down maybe 7 or 8 times now? It would be nice to get some visibility into what is being done to make this paid offering more resilient. IE have you considered the updater doing a blue green deploy and leave us with a working "old version" if the new one has issues? How about just using a daemon watchdog to watch the process. If it dies like it does now, start it back up for me and allow the previous version to continue working vs getting stuck in an update failure loop
The text was updated successfully, but these errors were encountered: