[fuel-qa][fuel-devops] Couldn't revert ready snapshot after first extra compute test

[Victor Ryzhenkin]

This problem appears when we running rh test group from our commits.

Commits:
https://review.openstack.org/#/c/284677/
https://review.openstack.org/#/c/284681/

Now in stable/8.0 we have only one test. these two patches adds two more new tests.

Each test works good separately.
http://jenkins-product.srt.mirantis.net:8080/job/8.0.custom_system_test/1981/
http://jenkins-product.srt.mirantis.net:8080/job/8.0.custom_system_test/1971/

But if we run all tests in one job, we got this:
http://jenkins-product.srt.mirantis.net:8080/job/8.0.custom_system_test/1973/

First test passed, but subsequent tests failed due error while reverting ready snapshot:
libvirt: error : internal error: Child process (/usr/bin/qemu-img snapshot -a ready /var/lib/libvirt/images/8.0.custom_system_test.1973.2016-02-26_19-34-09_slave-02-system) unexpected exit status 1: qemu-img: Could not apply snapshot 'ready': -2 (No such file or directory)
libvirt: error : internal error: Child process (/usr/bin/qemu-img snapshot -a ready /var/lib/libvirt/images/8.0.custom_system_test.1973.2016-02-26_19-34-09_slave-02-system) unexpected exit status 1: qemu-img: Could not apply snapshot 'ready': -2 (No such file or directory)

The possible root cause of this is upload of custom image to system volume in test.

[Artem Panchenko (apanchenko-8)]

AFAIU you are overwriting 'system' disk data on one of slave nodes during the test:

https://github.com/openstack/fuel-qa/blob/master/fuelweb_test/tests/test_rh_compute.py#L567
https://github.com/openstack/fuel-qa/blob/master/fuelweb_test/tests/test_rh_compute.py#L110-L118

But snapshots are saved to the same qcow file as that data, so qemu fails when you try to revert old snapshot. I believe that is expected behavior, because you are 'hacking' virtual environment outside the fuel-devops. Closing as invalid.