Bug 460874

Summary: lost packets when live migrating (RHEL4 XEN)
Product: Red Hat Enterprise Linux 4 Reporter: Olivier Fourdan <ofourdan>
Component: kernel-xenAssignee: Don Dutile (Red Hat) <ddutile>
Status: CLOSED ERRATA QA Contact: Martin Jenner <mjenner>
Severity: urgent Docs Contact:
Priority: urgent    
Version: 4.8CC: bdevouge, ddutile, dhoward, dmair, fluo, herbert.xu, ijc, jim, jpirko, jplans, kem, mgahagan, mmatsuya, ofourdan, qcai, sputhenp, tao, vmayatsk, xen-maint
Target Milestone: rcKeywords: ZStream
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2009-05-18 19:06:17 UTC Type: ---
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: 453526    
Bug Blocks: 462680, 469891    
Attachments:
Description Flags
Fix race condition on interface bring-up
none
Use register_netdevice_notifier() instead of register_inetaddr_notifier()
none
Fix netdev_notify()
none
Fix race condition on interface bring-up
none
Proposed/posted patch for rhel4.8. none

Comment 1 Olivier Fourdan 2008-09-02 10:10:54 UTC
Created attachment 315515 [details]
Fix race condition on interface bring-up

This patch is from:

    https://bugzilla.redhat.com/show_bug.cgi?id=453526#c18
    https://bugzilla.redhat.com/attachment.cgi?id=311214

And was applied on the kernel 2.6.9-78

Comment 2 Olivier Fourdan 2008-09-02 10:13:49 UTC
Created attachment 315517 [details]
Use register_netdevice_notifier() instead of register_inetaddr_notifier()

This patch was created based on comment 32 in bug #453526 https://bugzilla.redhat.com/show_bug.cgi?id=453526#c32

It applies on top of a 2.6.9-78 kernel.

Comment 3 Olivier Fourdan 2008-09-02 10:17:18 UTC
Created attachment 315519 [details]
Fix netdev_notify()

This patch is based on comment 36 of bug #453526
https://bugzilla.redhat.com/show_bug.cgi?id=453526#c36

It applies on top of a kernel-2.6.9-78 source tree and complements previous attachment #315517 [details] (https://bugzilla.redhat.com/attachment.cgi?id=315517)

Comment 7 Olivier Fourdan 2008-09-23 08:14:56 UTC
Created attachment 317446 [details]
Fix race condition on interface bring-up

Thanks for noticing this, the attachement for comment #1 is not the correct file, this is the correct patch.

All three are required to fix the issue. Sorry for the mistake.

Comment 8 Herbert Xu 2008-09-23 08:24:23 UTC
Thanks Olivier.  This is the part of #458934 that we need for RHEL4.

Comment 10 Don Dutile (Red Hat) 2008-09-29 21:41:24 UTC
Created attachment 318010 [details]
Proposed/posted patch for rhel4.8.

May want to consider backporting to rhel4.7 z-stream if critical need before 4.8 release.

Comment 11 Vivek Goyal 2008-10-21 19:08:16 UTC
Committed in 78.15.EL . RPMS are available at http://people.redhat.com/vgoyal/rhel4/

Comment 13 Bill Burns 2008-10-28 12:08:43 UTC
*** Bug 468290 has been marked as a duplicate of this bug. ***

Comment 15 Don Dutile (Red Hat) 2008-12-01 17:08:54 UTC
*** Bug 473408 has been marked as a duplicate of this bug. ***

Comment 16 Issue Tracker 2009-01-28 20:51:48 UTC
Customer confirmed that 78.15.EL has fixed the issue.

Bill

Internal Status set to 'Waiting on Engineering'

This event sent from IssueTracker by bbraswel 
 issue 222899

Comment 19 Luo Fei 2009-04-22 10:48:34 UTC
I tested many times of live migration with 2.6.9-78.ELxenU or 2.6.9-88.ELxenU on the domU, RHEL5-Server-U3 on the two dom0s. 
with 2.6.9-78.ELxenU, many packets lost.
...
64 bytes from 10.16.71.121: icmp_seq=74 ttl=59 time=314 ms
64 bytes from 10.16.71.121: icmp_seq=75 ttl=59 time=312 ms
64 bytes from 10.16.71.121: icmp_seq=76 ttl=59 time=313 ms
64 bytes from 10.16.71.121: icmp_seq=380 ttl=59 time=312 ms
64 bytes from 10.16.71.121: icmp_seq=381 ttl=59 time=310 ms
64 bytes from 10.16.71.121: icmp_seq=382 ttl=59 time=321 ms
64 bytes from 10.16.71.121: icmp_seq=383 ttl=59 time=311 ms
...
with 2.6.9-88.ELxenU, shows no packet being lost.

Comment 21 errata-xmlrpc 2009-05-18 19:06:17 UTC
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

http://rhn.redhat.com/errata/RHSA-2009-1024.html