Hi,
My JupyterHub systemd service sometimes stops accepting connections (“Connection was reset”) and has to be restarted. The weird part is that systemctl status
still reports the service as running and no errors show up in journalctl
(in fact, beyond the point where the problem occurs nothing else shows in the log, JupyterHub seemingly cuts itself off from the world).
For obvious reasons, I didn’t want to make this a GitHub issue, since I don’t know how to reproduce and there are no error messages to guide the debugging. But I was wondering if maybe someone had a similar experience and managed to solve it nonetheless…? Or perhaps some suggestions on what diagnostics to try next time it happens, because frankly I’m at a loss here :)
Thanks for any pointers,
David
Thanks for the tip, but while I do have a few cronjobs, this problem doesn’t happen on any kind of regular schedule. But I’ll keep my eyes open and try to correlate it with the cron tasks next time it happens :)