A failed resume action will destroy and recreate the VM on the next resume
|Assignee:||Carlos Martín||% Done:|
|Category:||Core & System|
|Target version:||Release 4.14|
|Affected Versions:||OpenNebula 4.10, OpenNebula 4.12, OpenNebula 4.6, OpenNebula 4.8|
After a resume, the VM goes through the states pending -> prolog_resume/_undeploy.
To decide that the VM should go into prolog_resume/_undeploy instead of the regular prolog, we look for the previous history action. But if the prolog fails for any reason, we move the VM back to stopped/undeployed with a new history entry, reason NONE.
The next time the VM is resumed, the core will move it to the regular prolog state, creating a new VM from scratch instead of using the existing disks and checkpoint.
#6 Updated by Ruben S. Montero over 5 years ago
- Status changed from New to Closed
- Target version set to Release 4.14
- Resolution set to fixed
- Affected Versions OpenNebula 4.10, OpenNebula 4.12, OpenNebula 4.6, OpenNebula 4.8 added
Now, a VM can go to the UNDEPLOYED/STOPPED state after a successful epilog. The prolog_failed transition (that added the new history record) has been removed.