Bug 173633

Summary: cman/sm nodeid lookup fails
Product: [Retired] Red Hat Cluster Suite Reporter: Scott Cannata <scott.cannata>
Component: cmanAssignee: David Teigland <teigland>
Status: CLOSED WORKSFORME QA Contact: Cluster QE <mspqa-list>
Severity: medium Docs Contact:
Priority: medium    
Version: 4CC: ccaulfie, cluster-maint
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2006-05-04 16:52:31 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:
Attachments:
Description Flags
ascii file output from kdb, see above description none

Description Scott Cannata 2005-11-18 19:14:07 UTC 
Description of problem:

System running 2.6.9-22 kernel sitting idle went into kdb with
Assertion failed on line 52 of file cluster/cman/sm_misc.c

Attached typescript file has kdb: bt, sr -t, ps, dmesg output.

Version-Release number of selected component (if applicable):

kernel = 2.6.9-22
cman-1.0.2-0
cman-devel-1.0.2-0






How reproducible:

Not sure. We've only seen it once to date and the system
was not doing anything. Testers came in and found it kdb.

Steps to Reproduce:
1.
2.
3.
  
Actual results:


Expected results:


Additional info:
Comment 1 Scott Cannata 2005-11-18 19:14:09 UTC 
Created attachment 121241 [details]
ascii file output from kdb, see above description
Comment 2 David Teigland 2005-11-22 18:43:22 UTC 
Could you verify that cluster.conf was the same on all nodes?,
and a copy of that may be helpful to see, along with
'cman_tool nodes' from one of the other nodes if they are
still running.

This assertion failure may indicate some sort of internal
consistency problem within cman: the sm portion is looking
up a nodeid that the cnxman portion doesn't know about, which
shouldn't be possible.  If the assignment of nodeid's to nodes
was changing while the cluster was running (different versions
of cluster.conf on the nodes possibly), that might lead to this
kind of error.  If untypically large nodeids are being used, that
may point toward the cnxman code that dynamically increases the
standard node arrays.

This assertion failure was reported once before to me in an
email (in May by Dan Phung) and he had been updating cluster.conf
on some nodes.

I'm adding a printk to the code to provide a little more
information in the assertion message if this happens again.
Comment 3 David Teigland 2006-01-04 16:19:02 UTC 
waiting for someone to see this again and report with the
additional info from the panic