unbound not running after automatic update

Bug #1775833 reported by Dr. Jens Harbott on 2018-06-08
22
This bug affects 4 people
Affects Status Importance Assigned to Milestone
unbound (Ubuntu)
Undecided
Unassigned

Bug Description

We have several machines where unbound was updated tonight by automated security updates to 1.5.8-1ubuntu1.1. Afterwards there was no unbound process running, output from systemctl status says:

systemctl status unbound
● unbound.service
   Loaded: loaded (/etc/init.d/unbound; bad; vendor preset: enabled)
  Drop-In: /run/systemd/generator/unbound.service.d
           └─50-insserv.conf-$named.conf, 50-unbound-$named.conf
   Active: active (exited) since Fri 2018-06-08 06:39:21 UTC; 5h 17min ago
     Docs: man:systemd-sysv-generator(8)
  Process: 25819 ExecStop=/etc/init.d/unbound stop (code=exited, status=0/SUCCESS)
  Process: 25857 ExecStart=/etc/init.d/unbound start (code=exited, status=0/SUCCESS)
    Tasks: 0
   Memory: 0B
      CPU: 0

Jun 08 06:39:22 ze02 unbound[1681]: [1681:0] info: 0.065536 0.131072 12826
Jun 08 06:39:22 ze02 unbound[1681]: [1681:0] info: 0.131072 0.262144 15352
Jun 08 06:39:22 ze02 unbound[1681]: [1681:0] info: 0.262144 0.524288 5310
Jun 08 06:39:22 ze02 unbound[1681]: [1681:0] info: 0.524288 1.000000 1067
Jun 08 06:39:22 ze02 unbound[1681]: [1681:0] info: 1.000000 2.000000 357
Jun 08 06:39:22 ze02 unbound[1681]: [1681:0] info: 2.000000 4.000000 84
Jun 08 06:39:22 ze02 unbound[1681]: [1681:0] info: 4.000000 8.000000 12
Jun 08 06:39:22 ze02 unbound[1681]: [1681:0] info: 64.000000 128.000000 16
Jun 08 06:39:22 ze02 unbound[1681]: [1681:0] info: 128.000000 256.000000 12
Jun 08 06:39:22 ze02 unbound[1681]: [1681:0] info: 256.000000 512.000000 52

After a restart everything is fine again. Looking at the extend log with journalctl it looks like the old process was too slow with stopping and systemd was too eager starting the new one. But it is not clear to me why there is no error logged from the new process, then. See http://paste.openstack.org/show/722969/ for that log.

Dr. Jens Harbott (j-harbott) wrote :

This only seem to happen under load, so it really looks like a race condition during the restart. Guessing the daemon finishes processing requests before it terminates.

Chris Williams (cswilliams) wrote :

I noticed the same issue on some of our unbound servers as well on Ubuntu 16.04. I agree that it's some kind of race condition since it only seemed to happen on some of our servers and not all of them even though they are all configured identically.

Launchpad Janitor (janitor) wrote :

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in unbound (Ubuntu):
status: New → Confirmed
Robbert Muller (mjrider) wrote :

Same here, 4 unbound instances, all but one crashed during the update. The last one wasn't under load during the restart

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers