| 2015-04-16 05:28:33 |
Deepak C Shetty |
description |
This bug is in continuation of https://bugs.launchpad.net/cinder/+bug/1441050 (Part 1 of the problem)
This bug is opened to track the 2nd issue that GlusterFS (and in future any FS backed driver that uses hyp assisted snapshot) is hitting.
Refer to comment nos: #5 (No. 2) , #6, #8, #10 of https://bugs.launchpad.net/cinder/+bug/1441050 for more details
In short, the testcase tries to delete the snapshot after the VM is OFF, which libvirt yet doesn't support for external snapshots, thus the excp seen as below :
2015-04-08 07:22:44.376 32701 TRACE nova.virt.libvirt.driver File "/usr/local/lib/python2.7/dist-packages/libvirt.py", line 792, in blockRebase
2015-04-08 07:22:44.376 32701 TRACE nova.virt.libvirt.driver if ret == -1: raise libvirtError ('virDomainBlockRebase() failed', dom=self)
2015-04-08 07:22:44.376 32701 TRACE nova.virt.libvirt.driver libvirtError: Requested operation is not valid: domain is not running
2015-04-08 07:22:44.376 32701 TRACE nova.virt.libvirt.driver
Some mailing list discussion around this issue is available at:
http://lists.openstack.org/pipermail/openstack-dev/2015-April/060983.html
This bug is opened to track this issue to figure whether we wait for libvirt support or change Nova-Cinder interactions
to better handle this scenario better ? |
This bug is in continuation of https://bugs.launchpad.net/cinder/+bug/1441050 (Part 1 of the problem)
This bug is opened to track the 2nd issue that GlusterFS (and in future any FS backed driver that uses hyp assisted snapshot) is hitting.
Refer to comment nos: #5 (No. 2) , #6, #8, #10 of https://bugs.launchpad.net/cinder/+bug/1441050 for more details
In short, the testcase tries to delete the snapshot after the VM is OFF, which libvirt yet doesn't support for external snapshots, thus the excp seen as below :
2015-04-08 07:22:44.376 32701 TRACE nova.virt.libvirt.driver File "/usr/local/lib/python2.7/dist-packages/libvirt.py", line 792, in blockRebase
2015-04-08 07:22:44.376 32701 TRACE nova.virt.libvirt.driver if ret == -1: raise libvirtError ('virDomainBlockRebase() failed', dom=self)
2015-04-08 07:22:44.376 32701 TRACE nova.virt.libvirt.driver libvirtError: Requested operation is not valid: domain is not running
2015-04-08 07:22:44.376 32701 TRACE nova.virt.libvirt.driver
Some mailing list discussion around this issue is available at:
http://lists.openstack.org/pipermail/openstack-dev/2015-April/060983.html
This bug is opened to track this issue to figure whether we wait for libvirt support or change Nova-Cinder interactions to better handle this scenario ?
Due to this, the Cinder-GlusterFS CI job (check-tempest-dsvm-full-glusterfs-nv) is failing, which was otherwise running successfully for several weeks. |
|
