Comment 12 for bug 1654025

Hello Iain,

your patch had a few tests/pycodestyle errors, and I squashed them together. However, when testing this I sometimes run into a timeout:

$ runner/autopkgtest tests/testpkg-reboot -- lxd autopkgtest/ubuntu/zesty/amd64
autopkgtest [10:43:57]: git checkout: 77cd1c3 doc/README.package-tests.rst: Document network access
autopkgtest [10:43:57]: host donald; command line: runner/autopkgtest tests/testpkg-reboot -- lxd autopkgtest/ubuntu/zesty/amd64
autopkgtest [10:44:10]: testbed dpkg architecture: amd64
autopkgtest [10:44:12]: testbed running kernel: Linux 4.8.0-29-generic #31~lp1626436ProposedWithTwoPatches SMP Fri Nov 18 18:21:38 UTC
autopkgtest [10:44:12]: @@@@@@@@@@@@@@@@@@@@ built-tree tests/testpkg-reboot
autopkgtest [10:44:12]: test p: preparing testbed
autopkgtest [10:44:14]: test p: [-----------------------
bash: line 1: 441 Killed /tmp/autopkgtest.Hoe2Rq/tree/debian/tests/p 2> >(tee -a /tmp/autopkgtest.Hoe2Rq/p-stderr >&2) > >(tee -a /tmp/autopkgtest.Hoe2Rq/p-stdout)
autopkgtest [10:44:15]: test process requested reboot with marker phase1
error: Container is not running.
<VirtSubproc>: failure: timed out waiting for container autopkgtest-lxd-ydlkyp to restart
autopkgtest [10:46:06]: ERROR: testbed failure: cannot send to testbed: ['BrokenPipeError: [Errno 32] Broken pipe\n']

so this doesn't seem to be very reliable.

How does it fare on the production infra?