Created attachment 600750 [details] logs Description of problem: SPM host after setting SPM priority to -1 and subsequent vdsm restart still acts as SPM. Version-Release number of selected component (if applicable): Host: vdsm-4.9.6-21.0.el6_3.x86_64 libvirt-0.9.10-21.el6_3.3.x86_64 RHEVM: rhevm-3.1.0-6.el6ev.noarch Steps to Reproduce: 1. install 3 hosts 2. set SPM priority - (-1, -1, 2) 3. deactivate/activate hosts 4. wait until 3. host become SPM 5. set 3. host's SPM priority to -1 in DB (all hosts now have -1 SPM priority) 6. restart VDSM on 3. host Actual results: - 3. host is still SPM Expected results: - no host acts as SPM Additional info: attached logs from host: vdsm.log, spm-lock.log attached logs from RHEVM: engine.log spm-logck.log: ---------------- [2012-07-27 13:18:35] releasing lease spUUID=04875b86-a253-4c98-9053-c8d711c75847 id=3 lease_path=/rhev/data-center/mnt/10.34.63.199:_pd01/04875b86-a253-4c98-9053-c8d711c75847/dom_md/leases vdsm.log: --------------- MainThread::INFO::2012-07-27 13:18:35,488::vmChannels::135::vds::(stop) VM channels listener was stopped.$ MainThread::DEBUG::2012-07-27 13:18:35,503::task::588::TaskManager.Task::(_updateState) Task=`17285ebb-d33d-4638-8a7d-1829b8ed4fb9`::moving from state init -> state preparing$ MainThread::INFO::2012-07-27 13:18:35,503::logUtils::37::dispatcher::(wrapper) Run and protect: prepareForShutdown(options=None)$ Thread-11::DEBUG::2012-07-27 13:18:35,505::storageServer::617::ConnectionMonitor::(_monitorConnections) Monitoring stopped$ MainThread::WARNING::2012-07-27 13:18:35,654::hsm::2979::Storage.HSM::(__releaseLocks) Found lease locks, releasing$ MainThread::DEBUG::2012-07-27 13:18:36,720::taskManager::80::TaskManager::(prepareForShutdown) Request to stop all tasks$ MainThread::INFO::2012-07-27 13:18:36,725::logUtils::39::dispatcher::(wrapper) Run and protect: prepareForShutdown, Return response: None$ MainThread::DEBUG::2012-07-27 13:18:36,725::task::1172::TaskManager.Task::(prepare) Task=`17285ebb-d33d-4638-8a7d-1829b8ed4fb9`::finished: None$ MainThread::DEBUG::2012-07-27 13:18:36,726::task::588::TaskManager.Task::(_updateState) Task=`17285ebb-d33d-4638-8a7d-1829b8ed4fb9`::moving from state preparing -> state finished engine.log: ----2012-07-27 13:19:06,236 INFO [org.ovirt.engine.core.bll.LoginUserCommand] (ajp-/0.0.0.0:8009-4) Running command: LoginUserCommand internal: false.$ 2012-07-27 13:19:06,246 WARN [org.ovirt.engine.core.bll.GetConfigurationValueQuery] (ajp-/0.0.0.0:8009-4) calling GetConfigurationValueQuery (ApplicationMode) with null version, using default general for version$ 2012-07-27 13:19:08,514 INFO [org.ovirt.engine.core.bll.storage.SetStoragePoolStatusCommand] (QuartzScheduler_Worker-49) [1510efa9] Running command: SetStoragePoolStatusCommand internal: true. Entities affected : ID: 30a479e2-8dc6-4c98-9f2e-f254d4aa477b Type: StoragePool$ 2012-07-27 13:19:08,529 ERROR [org.ovirt.engine.core.vdsbroker.irsbroker.IrsBrokerCommand] (QuartzScheduler_Worker-49) [1510efa9] IrsBroker::Failed::GetStoragePoolInfoVDS due to: ConnectException: Connection refused$ 2012-07-27 13:19:09,382 WARN [org.ovirt.engine.core.vdsbroker.VdsManager] (QuartzScheduler_Worker-63) [760cc54d] ResourceManager::refreshVdsRunTimeInfo::Failed to refresh VDS , vds = 1b6a40d0-d7c7-11e1-a6ab-001a4a013f69 : 10.34.63.137, VDS Network Error, continuing.$ VDSNetworkException: $ 2012-07-27 13:19:11,392 WARN [org.ovirt.engine.core.vdsbroker.VdsManager] (QuartzScheduler_Worker-44) [182bc200] ResourceManager::refreshVdsRunTimeInfo::Failed to refresh VDS , vds = 1b6a40d0-d7c7-11e1-a6ab-001a4a013f69 : 10.34.63.137, VDS Network Error, continuing.$ VDSNetworkException:-----------
I assume the logic which detects the last SPM will resume with this one. I don't think its too much of an issue, and can be specified as a release note that when setting a host spm priority to -1, it should be moved to maint first for this to take effect
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: Hosts must be moved to maintenance mode before changing their Storage Pool Manager (SPM) priority to '-1', otherwise the change will not take effect. An SPM priority of '-1' indicates the host must never be considered for the SPM role.
Adding flag in line with Miki's request to use it for filtering release notes.