Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1097298

Summary: The start time for 'migration_max_time_per_gib_mem' appears to be calculated too early.
Product: [Retired] oVirt Reporter: Vinzenz Feenstra [evilissimo] <vfeenstr>
Component: vdsmAssignee: Vinzenz Feenstra [evilissimo] <vfeenstr>
Status: CLOSED CURRENTRELEASE QA Contact: Gil Klein <gklein>
Severity: high Docs Contact:
Priority: urgent    
Version: 3.3CC: acathrow, bazulay, emarcian, flo_bugzilla, gklein, gwatson, iheim, jentrena, lpeer, mgoldboi, michal.skrivanek, mkalinin, sherold, yeylon
Target Milestone: ---Keywords: Regression
Target Release: 3.4.2   
Hardware: x86_64   
OS: Linux   
Whiteboard: virt
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: 1090109 Environment:
Last Closed: 2014-06-11 06:44:31 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 1090109, 1097332, 1114269    

Description Vinzenz Feenstra [evilissimo] 2014-05-13 13:55:42 UTC 
Description of problem:

With the introduction of 'migration_max_time_per_gib_mem' we have noticed some migrations failing due to an error similar to the following;

"The migration took 133 seconds which is exceeding the configured maximum time for migrations of 128 seconds. The migration will be aborted."


On closer inspection, it appears that the start time for this parameter may be captured too soon, i.e. in this specific case the actual data transfer had only been underway for 10 seconds before 'migration_max_time_per_gib_mem' was exceeded. The reason being that multiple migrations had been started simultaneously and the default value of 3 for 'max_outgoing_migrations' was limiting concurrent migrations to the point that some VMs took around 120 seconds just to acquire the migration semaphore.

Actual results:

The timer for 'migration_max_time_per_gib_mem' should probably be started once the semaphore has been acquired, as opposed to when the migration request was initiated.
Comment 1 Michal Skrivanek 2014-05-13 15:15:49 UTC 
merged to master, posted in 3.4
Comment 2 Vinzenz Feenstra [evilissimo] 2014-05-14 11:48:38 UTC 
Merged to ovirt-3.4 branch as http://gerrit.ovirt.org/gitweb?p=vdsm.git;a=commit;h=cc38b94d0eaa67d2b1811d4fdc84563c2e6d489e
Comment 3 Eldad Marciano 2014-05-25 11:46:09 UTC 
-build is av9.2 installed
-reducing the time out in order to reproduced the problem for 1gb ram.
-bug fixed.
-migration time not not expired, multiple migration pass (tested on 6 vms, as customer described).
-code fix in vm.py located.
Comment 4 Sandro Bonazzola 2014-06-11 06:44:31 UTC 
This is an automated message

oVirt 3.4.2 has been released:
 * should fix your issue
 * should be available at your local mirror within two days.

If problems still persist, please make note of it in this bug report.