Bug 919201

Summary:	Warning when migration is delayed/get stuck due to high guest memory writes.
Product:	Red Hat Enterprise Virtualization Manager	Reporter:	Amador Pahim <asegundo>
Component:	vdsm	Assignee:	Saveliev Peter <peet>
Status:	CLOSED ERRATA	QA Contact:	Tareq Alayan <talayan>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	3.1.0	CC:	acathrow, bazulay, danken, hateya, iheim, lpeer, michal.skrivanek, nicolas, ofrenkel, pstehlik, ykaul, zdover
Target Milestone:	---	Keywords:	Patch
Target Release:	3.2.0
Hardware:	All
OS:	Linux
Whiteboard:	virt
Fixed In Version:	vdsm-4.10.2-18.0.el6ev	Doc Type:	Bug Fix
Doc Text:	Previously, VDSM did not distinguish between migration failures caused by high guest-memory writes and migration failures caused by high network load. This meant that it wasn't clear why migrations failed. New error messages in the logs allow users to distinguish between migration failures caused by high guest-memory writes and migration failures caused by high network load.	Story Points:	---
Clone Of:		Environment:
Last Closed:	2013-06-10 20:43:06 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:	902971, 948448

Description Amador Pahim 2013-03-07 20:29:31 UTC

Description of problem:
Migration process consists in copy guest RAM to new Host. After the
copy, migration code loops back to the beginning and re-copies all pages
modified after the copy. The hope is that the list of pages which must
be copied shrinks with each pass over the memory range.
Guests with large number of memory writes can delay migration process
and even stuck it, if memory writes are faster than memory copy.
Currently, vdsm does not differentiate between migration not progressing due to high guest memory changes and due to memory copy hang.

Version-Release number of selected component (if applicable):
vdsm-4.10.2-1.6.el6.x86_64

Expected results:
Print a warning message when migration is not progressing due to high guest memory changes.

Comment 1 Amador Pahim 2013-03-07 20:31:11 UTC

Proposed patch:
http://gerrit.ovirt.org/12557

Comment 6 Dan Kenigsberg 2013-03-11 13:49:58 UTC

Why the current state, where the log has

 Migration Progress: %s seconds elapsed, %s%% of data processed ...

and the % of data processed is jumping/stalling, is not enough?

Comment 7 Amador Pahim 2013-03-11 14:27:36 UTC

No. Migration can be stalled for network load, destination host interruptions and so forth. In these cases, dataRemaining will be equal to smallest_dataRemaining. We should be able to identify when migration is not progressing due to high memory changes inside the guest (dataRemaining > smallest_dataRemaining). Btw, after some time trying to migrate in this scenario, the progress was used to go over 100% of data processed, but ffdc10c0 changed the progress to be always less than 100%. Now we can't identify the situation in any case.

Comment 8 Saveliev Peter 2013-03-20 15:23:33 UTC

Was merged as 95789edc988072202787729c3ff4e99ec95afcb6

Comment 15 errata-xmlrpc 2013-06-10 20:43:06 UTC

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

http://rhn.redhat.com/errata/RHSA-2013-0886.html