Comment 42 for bug 1012081

I now run lots of our autopkgtests (anything which is even remotely getting close to handling mounts, partitions, cgroups, etc.), including all cloud, LXC, deboostrap etc. packages. There were only two regressions:

 * apport's tests made specific assumptions about a conffile which now got moved. Fixed in https://launchpad.net/ubuntu/+source/apport/2.14.6-0ubuntu2

 * udisks2's tests seem to have become more racy with the new version (or the race condition more apparent due to running many tests in parallel). First attempt at fixing this was in https://launchpad.net/ubuntu/+source/udisks2/2.1.3-3 which helped in local QEMU, but not yet in CI's QEMU environment. I'll still track that separately, but it's not related to util-linux (that's ntfs-3g at most, but it's really just a race condition in the test).

Apart from those, everything is green again which was green before, so I consider this sufficiently regression tested.