Red Hat Bugzilla – Bug 1201020
[RFE] Audit load balancing migrations for more visibility
Last modified: 2017-04-24 20:55:40 EDT
Description of problem: Version-Release number of selected component (if applicable): RHEV 3.5 RHEL 6.6 host How reproducible: Every time (for me). Steps to Reproduce: 1. Start a VM on a host. 2. Force cpu usage on the host above the 'HighUtilization' mark defined in the cluster policy. 3. VM will/should migrate to another host. 4. Neither the engine log nor the Event Log indicate the reason. Actual results: Automatic migration occurs without the end user knowing why. Expected results: A message indicating the reason would be helpful. Additional info:
Gordon, your cluster is using even distribution load balancing, so this is working exactly as expected (and may even saved your VM from troubles). If you do not want the VM to be migrated automatically in such a case simply set the cluster scheduling policy to 'None' (or you can pin a VM to a specific host). See policy details in section 5.2.2.4 in [1]. Load balancing is not generating events to avoid flooding system and events log. If there's a strong case for it please let us know. Otherwise we should close this issue. [1] https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Virtualization/3.5/html/Administration_Guide/sect-Cluster_Tasks.html#sect-Explanation_of_Settings_and_Controls_in_the_New_Cluster_and_Edit_Cluster_Windows
We believe this is already working as expected as all migrations are reported to the audit log together with the reason for it.
Verified on rhevm-4.1.1-0.1.el7.noarch Now, for each migration under the engine we have the explicit reason, for example: Migration initiated by system (VM: golden_env_mixed_virtio_2, Source: host_mixed_2, Destination: host_mixed_1, Reason: Load balancing).