Bug 652150 - Restart xend causes losing running guest after crash with "on_crash=rename-restart"
Summary: Restart xend causes losing running guest after crash with "on_crash=rename-re...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Enterprise Linux 5
Classification: Red Hat
Component: xen
Version: 5.6
Hardware: x86_64
OS: Linux
medium
high
Target Milestone: rc
: ---
Assignee: Miroslav Rezanina
QA Contact: Virtualization Bugs
URL:
Whiteboard:
Depends On:
Blocks: 514499
TreeView+ depends on / blocked
 
Reported: 2010-11-11 06:55 UTC by YangGuang
Modified: 2011-07-21 11:59 UTC (History)
6 users (show)

Fixed In Version: xen-3.0.3-126.el5
Doc Type: Bug Fix
Doc Text:
Previously, when the "rename-restart" option was set as a reaction to a guest termination and the xend daemon was restarted after the guest crashed, xend lost track of the new guest instance. This bug has been fixed, and all guests are now visible via xend after the daemon restart.
Clone Of:
Environment:
Last Closed: 2011-07-21 09:15:43 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)
configuration file of pv guest (387 bytes, text/plain)
2010-11-11 06:57 UTC, YangGuang
no flags Details
dmesg info (15.22 KB, text/plain)
2010-11-11 11:49 UTC, YangGuang
no flags Details
xend.log (177.81 KB, text/plain)
2010-11-11 11:50 UTC, YangGuang
no flags Details


Links
System ID Private Priority Status Summary Last Updated
Red Hat Product Errata RHBA-2011:1070 0 normal SHIPPED_LIVE xen bug fix and enhancement update 2011-07-21 09:12:56 UTC

Description YangGuang 2010-11-11 06:55:51 UTC
Description of problem:
I have add "on_crash = rename-restart" to the pv guest config file. After guest crashed, restart xend will causes losing running guest. 

Version-Release number of selected component (if applicable):
kernel-xen-2.6.18-231.el5
xen-3.0.3-117.el5

How reproducible:
100%

Steps to Reproduce:
1.create a PV guest with "on_crash = rename-restart"
   [host]#xm cr pv-32-rhel5u5.cfg 


2.using "xm list" to watch guest state:
   [host]# xm li
   Name                                      ID Mem(MiB) VCPUs State   Time(s)
   Domain-0                                   0    32768    32 r-----    859.7
   xen-pv-guest-rhel5                         4      511     4 -b----     22.0


3.vnc to guest, make it crash.
   [guest]# echo c > /proc/sysrq-trigger


4.Using "xm list" to watch guest state:
   [host]# xm li
   Name                                      ID Mem(MiB) VCPUs State   Time(s)
   Domain-0                                   0    32768    32 r-----    944.1
   xen-pv-guest-rhel5                         5      511     4 -b----     20.0
   xen-pv-guest-rhel5-1                       4      511     1 ----c-     41.0

5.Restart xend.
  [host]# service xend restart


6.Using "xm list" to watch guest state:
   [host]# xm li
   Name                                      ID Mem(MiB) VCPUs State   Time(s)
   Domain-0                                   0    32768    32 r-----   1013.0
   xen-pv-guest-rhel5                         4      511     1 ----c-     41.0

Actual results:
1. After step6, running guest is disappear.

Expected results:
1. After step6, running guest should work well.

Additional info:

Comment 1 YangGuang 2010-11-11 06:57:04 UTC
Created attachment 459665 [details]
configuration file of pv guest

Comment 2 YangGuang 2010-11-11 11:49:51 UTC
Created attachment 459729 [details]
dmesg info

Comment 3 YangGuang 2010-11-11 11:50:40 UTC
Created attachment 459730 [details]
xend.log

Comment 4 Miroslav Rezanina 2010-12-02 10:40:54 UTC
Testing shows that rename is not properly propagated to hv so after restart, all rename data are only in xenstore but xend reload old infos from hv.

Comment 7 Miroslav Rezanina 2011-03-17 09:54:45 UTC
Fix built into xen-3.0.3-126.el5

Comment 9 Yuyu Zhou 2011-03-18 06:22:56 UTC
reproduced the bug on xen-3.0.3-117.el5 and verified the bug on xen-3.0.3-126.el5.

on xen-3.0.3-126.el5
Steps:
1.create a PV guest with "on_crash = rename-restart"
2.vnc to guest, make it crash.
3. Using "xm list" to watch guest state:
# xm list
Name                                      ID Mem(MiB) VCPUs State   Time(s)
Domain-0                                   0     5929     4 r-----     41.0
RHEL6.1-64-PV                              2     1023     1 -b----     15.2
RHEL6.1-64-PV-1                            1     1023     1 ----c-     12.2
4.Restart xend.
5. check the guest state again:
# xm list
Name                                      ID Mem(MiB) VCPUs State   Time(s)
Domain-0                                   0     5929     4 r-----     41.7
RHEL6.1-64-PV                              2     1023     1 -b----     15.2
RHEL6.1-64-PV-1                            1     1023     1 ----c-     12.2

So Change the status to VERIFIED.

Comment 10 Tomas Capek 2011-07-13 13:20:52 UTC
    Technical note added. If any revisions are required, please edit the "Technical Notes" field
    accordingly. All revisions will be proofread by the Engineering Content Services team.
    
    New Contents:
Previously, when the "rename-restart" option was set as a reaction to a guest termination and the xend daemon was restarted after the guest crashed, xend lost track of the new guest instance. This bug has been fixed, and all guests are now visible via xend after the daemon restart.

Comment 11 errata-xmlrpc 2011-07-21 09:15:43 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/RHBA-2011-1070.html

Comment 12 errata-xmlrpc 2011-07-21 11:59:28 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/RHBA-2011-1070.html


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