Red Hat Bugzilla – Bug 966597
NPE when stopping vm that uses local storage.
Last modified: 2015-09-22 09:09 EDT
Description of problem:
When trying to stop a vm that is running on an a local storage an npe is being thrown:
2013-05-23 16:49:48,373 ERROR [org.jboss.as.ejb3.invocation] (ajp-/127.0.0.1:8702-7) JBAS014134: EJB Invocation failed on component VdsEventListener for method public abstract void org.ovirt.engine.core.common.businessentities.IVdsEventListener.removeAsyncRunningCommand(org.ovirt.engine.core.compat.Guid): javax.ejb.EJBException: java.lang.NullPointerException
2013-05-23 16:49:48,380 ERROR [org.ovirt.engine.core.vdsbroker.DestroyVmVDSCommand] (ajp-/127.0.0.1:8702-7)  Command DestroyVmVDS execution failed. Exception: EJBException: java.lang.NullPointerException
server.log contains npes for run VM first and then stop vm every day from 14/05 until 23/05
From the vdsm log the flow here is:
user calls vmCreate - succeeds
user calls hibernateVM - succeeds
user calls vmDestroy - fails since the VM was already hibernated.
Thread-105::DEBUG::2013-05-23 15:40:55,850::vm::675::vm.Vm::(_startUnderlyingVm) vmId=`582b5afd-e719-4cb7-bf2c-93234bf104cf`::Start
Thread-175::DEBUG::2013-05-23 15:42:11,867::vm::183::vm.Vm::(_prepareGuest) vmId=`582b5afd-e719-4cb7-bf2c-93234bf104cf`::Save State begins
Thread-175::DEBUG::2013-05-23 15:42:13,541::vm::1110::vm.Vm::(setDownStatus) vmId=`582b5afd-e719-4cb7-bf2c-93234bf104cf`::Changed state to Down: SaveState succeeded
Thread-181::DEBUG::2013-05-23 15:42:15,633::libvirtvm::2691::vm.Vm::(destroy) vmId=`582b5afd-e719-4cb7-bf2c-93234bf104cf`::destroy Called
vdsm.log states that destroy failed since there is no such VM (obviously):
Thread-181::DEBUG::2013-05-23 15:42:15,633::libvirtvm::2691::vm.Vm::(destroy) ...
vmId=`582b5afd-e719-4cb7-bf2c-93234bf104cf`::trying to set state to Powering down when already Down
So I'd guess the test is faulty and since this has been happening since the server started up (14/5) but only reported on the 23/5 that this is not a regression?
In any event, this has nothing to do with storage.
no need to track this twice...
*** This bug has been marked as a duplicate of bug 975558 ***