/var/log/upstart is missing

[Mike Conigliaro]

I'm not sure if this is the right place to submit this bug, but I created a new EC2 instance last night using the official Canonical AMI for Precise (ami-a29943cb), and I found that /var/log/upstart was missing. This means Upstart jobs won't be able to log anything, and results in the following daily error from logrotate:

/etc/cron.daily/logrotate:
error: error accessing /var/log/upstart: No such file or directory
error: upstart:1 glob failed for /var/log/upstart/*.log
error: found error in /var/log/upstart/*.log , skipping

The fix is simple:

mkdir /var/log/upstart

[Launchpad Janitor (janitor)]

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

[James Hunt (jamesodhunt)]

This is rather curious as /var/log/upstart is auto-created on install:

  dpkg -L upstart|grep /var/log/upstart

I suspect something else on those systems is removing /var/log/upstart/ which is causing subsequent logrotate runs to error.

[Janne (jannehson51)]

Is there anything I can do to help finding what is causing this error?

[James Hunt (jamesodhunt)]

This issue might be caused by a cron job that is clearing out /var/log/? Or some logrotate configuration issue possibly?

[Mike Conigliaro (mconigliaro)]

For what it's worth, I only created this directory once, and it hasn't disappeared yet. If some cron job or logrotate config was the culprit, I imagine the directory would eventually disappear again.

[glance (glance-acc)]

/var/log/upstart is missing in the precise-cloud-images allso.

Reinstalling the upstart package and a reboot later, everything is back to normal.

[markus (markus-bengts)]

I just launched a precise image from http://cloud.ubuntu.com/ami/ and checked the contents of /var/log (as user root). Everything in /var/log had been created after the server had been started. It seems that all content of /var/log is ignored when the server image is created.

[Steve Langasek (vorlon)]

I think this is a duplicate of a bug that was already fixed in the cloud images, but I don't remember where that fix happened. Reassigning to cloud-init.

[Ben Howard (utlemming)]

Yes, this is a duplicate bug. It was fixed in the latest AMI. You're using the release AMI. Please pick the current one http://cloud.ubuntu.com/ami.