Hide Forgot
Created attachment 557182 [details] logs Description of problem: vm's created in rhevm have interface rhevm, currently we are importing the vm with interface rhevm which means that they cannot run. we should either automatically replace the interface during import or add a warning that the vm's interface should be changed manually so that the vm can run. we also do not get an error in engine log once vm failed to run - only in vdsm log once vm failed to run Version-Release number of selected component (if applicable): ovirt-engine-backend-3.0.0_0001-1.2.fc16.x86_64 How reproducible: 100% Steps to Reproduce: 1. create a vm in rhevm and export it to an export domain 2. import the vm to ovirt and try to run the vm 3. Actual results: vm still has rhevm interface and we cannot run the vm Expected results: we should either automatically replace the interface to ovirtmgmt or give warning to user to do so manually. Additional info: full logs attached from vdsm log: Traceback (most recent call last): File "/usr/share/vdsm/vm.py", line 518, in _startUnderlyingVm self._run() File "/usr/share/vdsm/libvirtvm.py", line 1169, in _run self._connection.createXML(domxml, flags), File "/usr/share/vdsm/libvirtconnection.py", line 79, in wrapper ret = f(*args, **kwargs) File "/usr/lib64/python2.7/site-packages/libvirt.py", line 2100, in createXML if ret is None:raise libvirtError('virDomainCreateXML() failed', conn=self) libvirtError: Failed to add tap interface to bridge 'rhevm': No such device engine log - only showing re-run: 2012-01-24 11:52:38,160 INFO [org.ovirt.engine.core.vdsbroker.CreateVmVDSCommand] (pool-5-thread-36) IncreasePendingVms::CreateVmIncreasing vds blond-vdsg pending vcpu count, n ow 1. Vm: XP-HA 2012-01-24 11:52:38,165 INFO [org.ovirt.engine.core.vdsbroker.CreateVmVDSCommand] (pool-5-thread-36) FINISH, CreateVmVDSCommand, return: WaitForLaunch, log id: 2f27789f 2012-01-24 11:52:43,013 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.DestroyVDSCommand] (QuartzScheduler_Worker-47) START, DestroyVDSCommand(vdsId = 785d6436-42ab-11e1-a421- 001a4a169741, vmId=ba681f03-5c6e-483d-9265-504c90042388, force=false, secondsToWait=0, gracefully=false), log id: 13698a65 2012-01-24 11:52:43,980 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.DestroyVDSCommand] (QuartzScheduler_Worker-47) FINISH, DestroyVDSCommand, log id: 13698a65 2012-01-24 11:52:44,006 INFO [org.ovirt.engine.core.vdsbroker.VdsUpdateRunTimeInfo] (QuartzScheduler_Worker-47) Running on vds during rerun failed vm: null 2012-01-24 11:52:44,010 INFO [org.ovirt.engine.core.vdsbroker.VdsUpdateRunTimeInfo] (QuartzScheduler_Worker-47) vm XP-HA running in db and not running in vds - add to rerun tre atment. vds blond-vdsg 2012-01-24 11:52:44,050 ERROR [org.ovirt.engine.core.vdsbroker.VdsUpdateRunTimeInfo] (QuartzScheduler_Worker-47) Rerun vm ba681f03-5c6e-483d-9265-504c90042388. Called from vds b lond-vdsg
I'm not sure how interesting the use case is of import/export between incompatible versions. but importing a VM with a differently named logical network and removing and warning is probably a valid use case
we will be having the same issue with templates imported. the rhevm interface will remain in template and all vm's created with it will have rhevm interface and not ovirtmgmt. so if we are fixing this issue we should also extend the fix to templates.
I don't see a reason to change the management network name, I don't think that keeping compatibility between rhevm and oVirt is needed.