Bug 1410040 - VM enter to the migration loop when the engine has both VM to host soft affinity group and load balancer
Summary: VM enter to the migration loop when the engine has both VM to host soft affin...
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: ovirt-engine
Classification: oVirt
Component: Backend.Core
Version: 4.1.0
Hardware: x86_64
OS: Linux
high
high
Target Milestone: ovirt-4.1.0-beta
: 4.1.0.2
Assignee: Yanir Quinn
QA Contact: Artyom
URL:
Whiteboard:
Depends On:
Blocks: 1404660
TreeView+ depends on / blocked
 
Reported: 2017-01-04 10:16 UTC by Artyom
Modified: 2017-02-01 14:43 UTC (History)
2 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2017-02-01 14:43:28 UTC
oVirt Team: SLA
Embargoed:
rule-engine: ovirt-4.1+


Attachments (Terms of Use)
engine log (7.41 MB, text/plain)
2017-01-04 10:16 UTC, Artyom
no flags Details


Links
System ID Private Priority Status Summary Last Updated
oVirt gerrit 69012 0 master MERGED bll: Prevent migration to the same host when balancing 2017-01-09 10:26:21 UTC
oVirt gerrit 69014 0 master MERGED scheduling: Allow balancing to the same host 2017-01-09 09:45:09 UTC
oVirt gerrit 69842 0 ovirt-engine-4.1 MERGED bll: Prevent migration to the same host when balancing 2017-01-09 16:43:12 UTC
oVirt gerrit 69843 0 ovirt-engine-4.1 MERGED scheduling: Allow balancing to the same host 2017-01-09 16:42:53 UTC
oVirt gerrit 69863 0 master MERGED scheduling: Modify migration policy behavior 2017-01-10 11:30:11 UTC
oVirt gerrit 69892 0 ovirt-engine-4.1 MERGED scheduling: Modify migration policy behavior 2017-01-10 10:19:31 UTC
oVirt gerrit 69920 0 master MERGED scheduling: Changed default factor for VM to host. 2017-01-10 14:04:50 UTC
oVirt gerrit 69924 0 ovirt-engine-4.1 MERGED scheduling: Changed default factor for VM to host. 2017-01-10 16:39:10 UTC

Description Artyom 2017-01-04 10:16:13 UTC
Created attachment 1237086 [details]
engine log

Description of problem:
If I define soft VM to host affinity group and update cluster scheduler policy to the policy with the load balancer, the engine starts to migrate VM all time between the hosts.

Version-Release number of selected component (if applicable):
rhevm-4.1.0-0.3.beta2.el7.noarch

How reproducible:
Always

Steps to Reproduce:
1. In the environment with at least two hosts
2. Create soft, positive VM to host affinity group(af_1)
3. Add vm_1 and host_1 to af_1
4. Create 50% CPU load on the host_2
5. Start vm_1
6. Change the cluster scheduling policy to power_saving
7. Wait 4-5 minutes

Actual results:
	
Jan 4, 2017 12:12:23 PM
Migration completed (VM: golden_env_mixed_virtio_0, Source: host_mixed_1, Destination: host_mixed_3, Duration: 5 seconds, Total: 5 seconds, Actual downtime: (N/A))

Jan 4, 2017 12:12:19 PM
Migration initiated by system (VM: golden_env_mixed_virtio_0, Source: host_mixed_1, Destination: host_mixed_3, Reason: Affinity rules enforcement).

Jan 4, 2017 12:12:03 PM
Migration completed (VM: golden_env_mixed_virtio_0, Source: host_mixed_3, Destination: host_mixed_1, Duration: 6 seconds, Total: 6 seconds, Actual downtime: (N/A))

Jan 4, 2017 12:11:59 PM
Migration initiated by system (VM: golden_env_mixed_virtio_0, Source: host_mixed_3, Destination: host_mixed_1, Reason: Load balancing).

Jan 4, 2017 12:11:21 PM
Migration completed (VM: golden_env_mixed_virtio_0, Source: host_mixed_1, Destination: host_mixed_3, Duration: 4 seconds, Total: 4 seconds, Actual downtime: (N/A))

Jan 4, 2017 12:11:18 PM
Migration initiated by system (VM: golden_env_mixed_virtio_0, Source: host_mixed_1, Destination: host_mixed_3, Reason: Affinity rules enforcement).

Jan 4, 2017 12:11:01 PM
Migration completed (VM: golden_env_mixed_virtio_0, Source: host_mixed_3, Destination: host_mixed_1, Duration: 4 seconds, Total: 4 seconds, Actual downtime: (N/A))

Jan 4, 2017 12:10:57 PM
Migration initiated by system (VM: golden_env_mixed_virtio_0, Source: host_mixed_3, Destination: host_mixed_1, Reason: Load balancing).

Jan 4, 2017 12:10:20 PM
Migration completed (VM: golden_env_mixed_virtio_0, Source: host_mixed_1, Destination: host_mixed_3, Duration: 4 seconds, Total: 4 seconds, Actual downtime: (N/A))

Jan 4, 2017 12:10:16 PM
Migration initiated by system (VM: golden_env_mixed_virtio_0, Source: host_mixed_1, Destination: host_mixed_3, Reason: Affinity rules enforcement).

Jan 4, 2017 12:10:00 PM
Migration completed (VM: golden_env_mixed_virtio_0, Source: host_mixed_3, Destination: host_mixed_1, Duration: 5 seconds, Total: 5 seconds, Actual downtime: (N/A))

Jan 4, 2017 12:09:56 PM
Migration initiated by system (VM: golden_env_mixed_virtio_0, Source: host_mixed_3, Destination: host_mixed_1, Reason: Load balancing).

Expected results:
Engine must prevent such behavior

Additional info:

Comment 1 Artyom 2017-01-26 10:55:39 UTC
Verified on rhevm-4.1.0.2-0.2.el7.noarch

Checked that no loop appears for power_saving, even_distribution and even_vm_distribution


Note You need to log in before you can comment on or make changes to this bug.