[python] Diagnostic snapshot doesn't contain '/var/log/docker-logs' directory'

[Artem Panchenko]

Fuel version info (7.0 build #14): http://paste.openstack.org/show/374013/

After environment deployment failed system tests created diagnostic snapshot, but it didn't contain '/var/log/docker-logs/' directory:

https://ci.fuel-infra.org/job/master.fuel-library.pkgs.ubuntu.ha_neutron_vlan/2120/

2015-07-14 07:40:26,822 - INFO fuel_web_client.py:1050 -- Task {u'status': u'ready', u'name': u'dump', u'cluster': None, u'result': {}, u'progress': 100, u'message': u'/api/dump/fuel-snapshot-2015-07-14_07-39-17.tar.xz', u'id': 9, u'uuid': u'df5e4f35-01c9-48b9-b55b-36c267ae2930'} finished. Took 70 second
...
2015-07-14 07:40:26,839 - INFO decorators.py:46 -- Saving logs to "/home/jenkins/workspace/master.fuel-library.pkgs.ubuntu.ha_neutron_vlan/logs/2120/fail_error_deploy_neutron_vlan_ha-2015_07_14__07_40_26.tar.xz" file

That snapshot ('fail_error_deploy_neutron_vlan_ha-2015_07_14__07_40_26.tar.xz') is attached to the bug.

I checked 'shotgun.log' but found no errors. After master node reset (it hanged after revert from snapshot) I tried to generate diagnostic snapshot manually and everything was fine, I'll attach it to the bug too.

[Alexander Kislitsky (akislitsky)]

There are no errors in snapshot generation. Seems that docker-logs was not mounted. Assigning to library-team.

[Oleksiy Molchanov (omolchanov)]

I have checked logs and found no problem with mounts, rest of logs are present in /var/log. Seems that the problem is not related to library. Also I have checked the latest snapshots on the CI and docker-logs are there.

I am marking this as invalid, no chance to reproduce it.