* Paul Gevers <elb...@debian.org> [210125 22:18]: > I don't have easy control over on which worker a job gets run. Unless I > start it manually on the worker, but then there are no logs. > > I went ahead and did that anyways, find the data attached.
Thanks for doing that. The log repeats these messages over and over: Jan 25 20:05:03 ci-025-b5185c19 systemd[1]: pdns-recursor.service: Main process exited, code=exited, status=226/NAMESPACE Jan 25 20:05:03 ci-025-b5185c19 systemd[1]: pdns-recursor.service: Failed with result 'exit-code'. TTBOMK this means some namespacing and/or protection feature implemented by systemd is broken on this worker. I don't know if this is a kernel, systemd, libc or some other problem, but it's not a pdns(-recursor) problem. Maybe it's best to start diffing the system configuration of the ppc64el workers, given the other one works? Chris