When ceph-disk has failed, collect more data

It is possible that a race condition can cause ceph-disk
to fail when attempting to mount a device, as it has
already tried mounting it. When that, or other unforseen
events, happens, we want to collect more data.

Partial-Bug: 1746118
Change-Id: I9c09951dbeb0329a8b0fcc57e9351492ef20c42d
This commit is contained in:
Chris MacNaughton 2018-03-22 09:28:29 +01:00
parent 23f0c12c47
commit 6238df8a36
1 changed files with 9 additions and 0 deletions

View File

@ -1513,10 +1513,19 @@ def osdize_dev(dev, osd_format, osd_journal, reformat_osd=False,
log("osdize cmd: {}".format(cmd))
subprocess.check_call(cmd)
except subprocess.CalledProcessError:
try:
lsblk_output = subprocess.check_output(
['lsblk', '-P']).decode('UTF-8')
except subprocess.CalledProcessError as e:
log("Couldn't get lsblk output: {}".format(e), ERROR)
if ignore_errors:
log('Unable to initialize device: {}'.format(dev), WARNING)
if lsblk_output:
log('lsblk output: {}'.format(lsblk_output), DEBUG)
else:
log('Unable to initialize device: {}'.format(dev), ERROR)
if lsblk_output:
log('lsblk output: {}'.format(lsblk_output), WARNING)
raise