After the node acting as spm was (accidentally) put into maintenance mode, the spm role is always contending. The original node doesnt take the role back either. About 10 instances were still running on the node when it was put into maintenance. I have ovirt 3.1.0-4 with a targetcli iscsi store. Here are the contents of the engine.log. Andy 2013-04-04 15:25:23,979 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.SpmStartVDSCommand] (QuartzScheduler_Worker-96) [5449512a] spmStart polling ended. spm status: Free 2013-04-04 15:25:23,993 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.HSMClearTaskVDSCommand] (QuartzScheduler_Worker-96) [5449512a] START, HSMClearTaskVDSCommand(vd sId = d265245e-3cc5-11e2-bce7-001018fc3b14, taskId=7c1cf7c6-21d4-4504-882d-3d465742b6ef), log id: d2da8ba 2013-04-04 15:25:24,011 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.HSMClearTaskVDSCommand] (QuartzScheduler_Worker-96) [5449512a] FINISH, HSMClearTaskVDSCommand, log id: d2da8ba 2013-04-04 15:25:24,012 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.SpmStartVDSCommand] (QuartzScheduler_Worker-96) [5449512a] FINISH, SpmStartVDSCommand, return: org.ovirt.engine.core.common.businessentities.SpmStatusResult@30a05218, log id: 5980d6f4 2013-04-04 15:25:24,014 INFO [org.ovirt.engine.core.bll.storage.SetStoragePoolStatusCommand] (QuartzScheduler_Worker-96) [6d6dbe98] Running command: SetStoragePoolStat usCommand internal: true. Entities affected : ID: 0ba15357-9b3b-4a76-8dac-b2b66b922174 Type: StoragePool 2013-04-04 15:25:24,843 ERROR [org.ovirt.engine.core.vdsbroker.irsbroker.IrsBrokerCommand] (QuartzScheduler_Worker-96) [6d6dbe98] IrsBroker::Failed::GetStoragePoolInfoV DS due to: IrsSpmStartFailedException: IRSGenericException: IRSErrorException: SpmStart failed 2013-04-04 15:25:24,900 INFO [org.ovirt.engine.core.vdsbroker.irsbroker.IrsBrokerCommand] (QuartzScheduler_Worker-96) [6d6dbe98] Irs placed on server null failed. Proc eed Failover 2013-04-04 15:25:34,977 INFO [org.ovirt.engine.core.vdsbroker.irsbroker.IrsBrokerCommand] (QuartzScheduler_Worker-96) [6d6dbe98] hostFromVds::selectedVds - moon-palace , spmStatus Free, storage pool Primary 2013-04-04 15:25:35,806 INFO [org.ovirt.engine.core.vdsbroker.irsbroker.IrsBrokerCommand] (QuartzScheduler_Worker-96) [6d6dbe98] starting spm on vds moon-palace, stora ge pool Primary, prevId -1, LVER 2520 2013-04-04 15:25:35,851 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.SpmStartVDSCommand] (QuartzScheduler_Worker-96) [6d6dbe98] START, SpmStartVDSCommand(vdsId = 3d 88c8b0-84bc-11e2-96b1-001018fc3b14, storagePoolId = 0ba15357-9b3b-4a76-8dac-b2b66b922174, prevId=-1, prevLVER=2520, storagePoolFormatType=V2, recoveryMode=Manual, SCSIF encing=false), log id: 30f4f64e 2013-04-04 15:25:45,901 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.SpmStartVDSCommand] (QuartzScheduler_Worker-96) [6d6dbe98] spmStart polling started: taskId = 6 8002cff-478d-4c2c-8839-b935d8a858b1 2013-04-04 15:26:11,903 INFO [org.ovirt.engine.core.bll.VdsLoadBalancer] (QuartzScheduler_Worker-81) [45db3aaa] VdsLoadBalancer: Starting load balance for cluster: BP_ Primary, algorithm: EvenlyDistribute. 2013-04-04 15:26:12,807 INFO [org.ovirt.engine.core.bll.VdsLoadBalancer] (QuartzScheduler_Worker-81) [45db3aaa] VdsLoadBalancer: high util: 51, low util: 0, duration: 2, threashold: 80 2013-04-04 15:26:12,845 INFO [org.ovirt.engine.core.bll.VdsLoadBalancingAlgorithm] (QuartzScheduler_Worker-81) [45db3aaa] VdsLoadBalancer: number of relevant vdss (no migration, no pending): 3. 2013-04-04 15:26:12,846 INFO [org.ovirt.engine.core.bll.VdsCpuVdsLoadBalancingAlgorithm] (QuartzScheduler_Worker-81) [45db3aaa] VdsLoadBalancer: number of over utilize d vdss found: 0. 2013-04-04 15:26:12,846 INFO [org.ovirt.engine.core.bll.VdsCpuVdsLoadBalancingAlgorithm] (QuartzScheduler_Worker-81) [45db3aaa] VdsLoadBalancer: max cpu limit: 40, num ber of ready to migration vdss: 3 2013-04-04 15:26:19,024 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.SpmStartVDSCommand] (QuartzScheduler_Worker-96) [6d6dbe98] spmStart polling ended: taskId = 680 02cff-478d-4c2c-8839-b935d8a858b1 task status = finished 2013-04-04 15:26:19,024 ERROR [org.ovirt.engine.core.vdsbroker.vdsbroker.SpmStartVDSCommand] (QuartzScheduler_Worker-96) [6d6dbe98] Start SPM Task failed - result: clea nSuccess, message: VDSGenericException: VDSErrorException: Failed in vdscommand to HSMGetTaskStatusVDS, error = BlockSD master file system FSCK error This log sequence repeats indefinitely