Mailinglist Archive: opensuse (1261 mails)

< Previous Next >
Re: [opensuse] Services not started after update/reboot
  • From: Peter Suetterlin <P.Suetterlin@xxxxxxxxxxxx>
  • Date: Thu, 29 Jun 2017 14:29:34 +0100
  • Message-id: <20170629132934.gdihdbg23vysgevt@woodstock.pitnet>
Andrei Borzenkov wrote:

Thanks for the suggestion Adrei!

One possible reason could be ordering dependency loop, in this case
systemd picks up some service to skip to break loop. I expect it to be
logged though.

So I went through the systemd messages during boot. I didn't expect a
dependency loop, as the system had booted fine before.

Instead I noticed two things:
- the pager journalctl uses by default is case-sensitive :o
(my less is always set to not be case sensitive on searches...)
So I overlooked some messages:

- The services (NFS server and Postfix) had been stopped before they even
would have been started. The issue was a problem with some directory:

----
Jun 28 22:59:18 royac6 systemd[1]: Started Purge old kernels.
Jun 28 22:59:30 royac6 systemd[1]: Stopped NFS server and services.
Jun 28 22:59:30 royac6 systemd[1]: Stopping NFSv4 ID-name mapping service...
Jun 28 22:59:30 royac6 systemd[1]: Stopped NFS Mount Daemon.
Jun 28 22:59:30 royac6 systemd[1]: Created slice
system-mdadm\x2dlast\x2dresort.slice.
Jun 28 22:59:30 royac6 systemd[1]: Starting Activate md array even though
degraded...
Jun 28 22:59:30 royac6 systemd[1]: Stopped Postfix Mail Transport Agent.
Jun 28 22:59:30 royac6 systemd[1]: Stopped NFSv4 ID-name mapping service.
Jun 28 22:59:30 royac6 systemd[1]: Started Activate md array even though
degraded.
Jun 28 22:59:30 royac6 systemd[1]: Stopped target Local File Systems.
Jun 28 22:59:30 royac6 systemd[1]: Unmounting /home...
Jun 28 22:59:30 royac6 systemd[1]: Stopped (with error) /dev/md1.
Jun 28 22:59:30 royac6 systemd[1]: Unmounted /home.
----

The home RAID had been fine though:
----
Jun 28 22:59:00 royac6 kernel: md: bind<sda1>
Jun 28 22:59:00 royac6 kernel: md: bind<sdb1>
Jun 28 22:59:00 royac6 kernel: md/raid1:md1: active with 2 out of 2 mirrors
Jun 28 22:59:00 royac6 kernel: md1: detected capacity change from 0 to
1024061145088
...
Jun 28 22:59:00 royac6 systemd[1]: Mounted /home.
Jun 28 22:59:00 royac6 kernel: EXT4-fs (md1): mounted filesystem with ordered
data mode. Opts: discard
------

So something in the 30s between the two snippets something must have marked
the RAID as failed. The only thing I can see in the logs is a huge block
where os-prober runs the full phalanx of /usr/lib/os-probes/mounted/* scripts
on /dev/md1. Not sure what had triggered this.

However, the boot now just continues, and 2s later md1 is obviously fine
again:
----
Jun 28 22:59:32 royac6 kernel: EXT4-fs (md1): mounted filesystem with ordered
data mode. Opts: discard
Jun 28 22:59:32 royac6 systemd[1]: Mounted /home.
----

The stopped services however are never (re)started.

I'd call this a bug - just not sure where and why....

--
To unsubscribe, e-mail: opensuse+unsubscribe@xxxxxxxxxxxx
To contact the owner, e-mail: opensuse+owner@xxxxxxxxxxxx

< Previous Next >
Follow Ups