Red Hat Bugzilla – Bug 967523
engine: after LSM failure of vm based on template we fail to delete the template image in dst domain because volume is still marked as shared
Last modified: 2016-02-10 15:48:45 EST
Created attachment 753565 [details]
Description of problem:
I created a failure in LSM for a vm based on a template.
after the failure I deleted the vm and deleted the template.
the template was reported as removed but when we try to import the same template to the target domain we fail because the volume already exists.
looking at the logs we can see that when we deleted the template we failed to delete the volume on the target domain because the volume is shared.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. in iscsi storage with multiple domains import a template
2. copy the template to all domains
3. create a vm from the template (as thin copy) and run the vm n the hsm
4. live storage migrate the vm disk
5. when engine logs 'FINISH, CreateSnapshotVDSCommand' block connectivity to the storage from the hsm host
6. when the host becomes non-operational restore the connectivity to the storage from the hsm
7. after all is restored, remove the vm and the template
8. try to import the template back to the setup
we fail to remove the template volume from the dst domain because its marked as shared.
we cannot import the template back to the setup because it already exists.
we should be able to remove the template from all domains or at the very least report that we failed to remove the template from one of the domains so that user can contact gss.
Additional info: logs
Created attachment 753566 [details]
logs from failure of LSM
attaching also the logs from the LSM failure (before the delete).
*** This bug has been marked as a duplicate of bug 965936 ***
After BZ965936 will be resolved, The disks should reflect the real current status in the storage.
the user should see an extra illegal disk when move operation will fail.
When the user will try to delete the template which is related to the disk that has been failed on move, an appropriate CDA message should notify the user which disk in the engine is related to the volume of the template.