Hide Forgot
Created attachment 756229 [details] engine.log, vdsm.log Description of problem: We have following scenarion: * Run a VM stateless (and open its console) * Stop vdsmd * Shutdown guest's OS from the guest itself (from console) * Start vdsmd When the guest is down and then vdsmd is started again, the stateless snapshot does still exists - it is not deleted. This state impacts immediate next start of the VM, which fails just because of the existence of the stateless snapshot. Failed to start VM jb-wxp, because exist snapshot for stateless state. Snapshot will be deleted. This failure cleans the snapshot and then everything is OK. Version-Release number of selected component (if applicable): sf17.2 How reproducible: 100% Steps to Reproduce: 1. Run a VM in stateless mode and open its console 2. stop vdsmd 3. shutdown guest's OS from the guest itself (console) 4. start vdsmd 5. check snapshosts list in admin portal Actual results: stateless snapshot does still exists for non-running VM Expected results: stateless snapshot should not exists for non-started VMs Additional info:
closing this as not a bug. following bug: https://bugzilla.redhat.com/show_bug.cgi?id=735097 the agreed solution/behaviour is to try and delete the stateless snapshot when vm is started again and to fail run of vm until the stateless snapshot is removed. once the snapshot is removed we can start the vm properly.