Created attachment 1471800 [details] engine, vdsm, qemu, libvirt logs Description of problem: while VM is in Not-Responding status and the host is Non-Operational, the engine enters the migration flow which has no chance to succeed. Engine returns to the failed migration in a circle (see the log fragment below). Version-Release number of selected component (if applicable): rhv-release-4.2.5-5-001.noarch libvirt-3.9.0-14.el7_5.6.x86_64 vdsm-4.20.35-1.el7ev.x86_64 qemu-kvm-rhev-2.10.0-21.el7_5.4.x86_64 How reproducible: 100% Steps to Reproduce: 1. Run NFS VM. Block the storage on host by the command 'iptables -I INPUT -s yellow-vdsb.qa.lab.tlv.redhat.com -j DROP'. 2. After the VM is in Not-Responding state and the host is Not-Operational check the engine.log Actual results: engine returns to the failed migration in a circle (see the log fragment below). 2018-07-29 16:01:15,203+03 INFO [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] (EE-ManagedThreadFactory-engineScheduled-Thread-77) [] EVENT_ID: USER_RUN_VM(32), VM golden_env_mixed_virtio_1_1 started on Host host_mixed_2 2018-07-29 16:03:16,469+03 INFO [org.ovirt.engine.core.vdsbroker.monitoring.VmAnalyzer] (EE-ManagedThreadFactory-engineScheduled-Thread-87) [] VM 'cd15b8d4-29cf-41b4-860c-27a38f993fa1'(golden_env_mixed_virtio_1_1) moved from 'Up' --> 'NotResponding' 2018-07-29 16:03:16,606+03 WARN [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] (EE-ManagedThreadFactory-engineScheduled-Thread-87) [] EVENT_ID: VM_NOT_RESPONDING(126), VM golden_env_mixed_virtio_1_1 is not responding. 2018-07-29 16:07:36,727+03 ERROR [org.ovirt.engine.core.bll.SetNonOperationalVdsCommand] (EE-ManagedThreadFactory-engine-Thread-17887) [7c4836a2] Failed to migrate VM 'golden_env_mixed_virtio_1_1' 2018-07-29 16:16:15,525+03 ERROR [org.ovirt.engine.core.bll.SetNonOperationalVdsCommand] (EE-ManagedThreadFactory-engine-Thread-18032) [7fa0ad53] Failed to migrate VM 'golden_env_mixed_virtio_1_1' 2018-07-29 16:25:17,320+03 ERROR [org.ovirt.engine.core.bll.SetNonOperationalVdsCommand] (EE-ManagedThreadFactory-engine-Thread-18209) [23d38dca] Failed to migrate VM 'golden_env_mixed_virtio_1_1' 2018-07-29 16:35:18,053+03 ERROR [org.ovirt.engine.core.bll.SetNonOperationalVdsCommand] (EE-ManagedThreadFactory-engine-Thread-18413) [30931f20] Failed to migrate VM 'golden_env_mixed_virtio_1_1' 2018-07-29 16:45:20,125+03 ERROR [org.ovirt.engine.core.bll.SetNonOperationalVdsCommand] (EE-ManagedThreadFactory-engine-Thread-18614) [c648901] Failed to migrate VM 'golden_env_mixed_virtio_1_1' 2018-07-29 16:55:13,448+03 ERROR [org.ovirt.engine.core.bll.SetNonOperationalVdsCommand] (EE-ManagedThreadFactory-engine-Thread-18811) [6d8cbb62] Failed to migrate VM 'golden_env_mixed_virtio_1_1' Expected results: must not enter the migration flow in this case. Reason: it is not HA VM and such migration could never succeed since host is Non-Operqtional and the VM is Not responding . Additional info:
this is by design. Non-Operational may mean many other things, not just a storage problem (and also for storage problem it doen't necessarily mean a VM accesses that storage and cannot migrate)