Bug 510510

Summary: cluster view inconsistent after "service cman stop; service cman start"
Product: Red Hat Enterprise Linux 5 Reporter: Benjamin Kahn <bkahn>
Component: cmanAssignee: Christine Caulfield <ccaulfie>
Status: CLOSED ERRATA QA Contact: Cluster QE <mspqa-list>
Severity: medium Docs Contact:
Priority: urgent    
Version: 5.3CC: cfeist, cluster-maint, cward, edamato, eva, jplans, matt, mrappa, pm-eus, tao
Target Milestone: rcKeywords: ZStream
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: cman-2.0.98-1.el5_3.7 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2009-08-03 07:40:41 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: 472786    
Bug Blocks:    

Description Benjamin Kahn 2009-07-09 15:06:55 UTC
This bug has been copied from bug #472786 and has been proposed
to be backported to 5.3 z-stream (EUS).

Comment 3 Christine Caulfield 2009-07-13 12:04:35 UTC
cherry-picked into the RHEL53 branch

commit 2bac6a6e1a24cd56f08bb89299296bf2a9a6ea58
Author: Christine Caulfield <ccaulfie>
Date:   Wed Dec 3 10:46:20 2008 +0000

    cman: Fix inconsistent state if a node leaves/joins quickly

Comment 5 Nate Straz 2009-07-24 15:20:17 UTC
I am still able to get the cluster in an inconsistent state after running "service cman restart" on one node.

[nstraz@z ~]$ for i in z{1,2,3,4}; do echo "$i:"; qarsh -l root $i cman_tool nodes; done
z1:
Node  Sts   Inc   Joined               Name
   1   M   1228   2009-07-24 08:58:41  z1
   2   M   1240   2009-07-24 08:59:11  z2
   3   M   1232   2009-07-24 08:58:41  z3
   4   M   1232   2009-07-24 08:58:41  z4
z2:
Node  Sts   Inc   Joined               Name
   1   M   1240   2009-07-24 08:59:12  z1
   2   M   1236   2009-07-24 08:59:12  z2
   3   M   1240   2009-07-24 08:59:12  z3
   4   M   1240   2009-07-24 08:59:12  z4
z3:
Node  Sts   Inc   Joined               Name
   1   M   1232   2009-07-24 08:58:41  z1
   2   M   1240   2009-07-24 08:59:11  z2
   3   M   1216   2009-07-24 08:32:16  z3
   4   M   1224   2009-07-24 08:32:17  z4
z4:
Node  Sts   Inc   Joined               Name
   1   X   1224                        z1
   2   X   1224                        z2
   3   M   1224   2009-07-24 08:32:18  z3
   4   M   1220   2009-07-24 08:32:17  z4

Comment 6 Christine Caulfield 2009-07-28 09:47:48 UTC
Given that bz#472786 has been verified as fixed, and it's the same bug, are we sure that the Z-stream package has that patch included ?

Comment 7 Chris Feist 2009-07-29 17:33:03 UTC
Yes, I talked with nate and he is using cman-2.0.98-1.el5_3.7 which I verified contains the patch:

-       if (node->state == NODESTATE_DEAD) {
+       if (node->state == NODESTATE_DEAD || node->state == NODESTATE_LEAVING) {

Comment 8 Nate Straz 2009-07-29 21:20:58 UTC
I reinstalled the systems and tried the test again and it passed this time around.  Last time I was working with 3, 4, and 5 node clusters, perhaps I didn't have all nodes updated correctly.  Calling this verified now.

Comment 10 errata-xmlrpc 2009-08-03 07:40:41 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-2009-1192.html