[Linux-HA] Node reboots when starting heartbeat

[Alexander Födisch]

it looks very strange. the UUID of node1 is set correct (the same like is set in cib.xml).
is this ID stored in /var/lib/heartbeat/hb_uuid?


zentgpfsn01:~ # crm_uuid
07ca44ca-1bf5-4f12-8680-21f86c2e6bca

zentgpfsn01:~ # grep zentgpfsn01 /var/lib/heartbeat/crm/cib.xml
      <node uname="zentgpfsn01" type="normal" id="07ca44ca-1bf5-4f12-8680-21f86c2e6bca">



in /var/lib/heartbeat/hostcache also the right UUqIDs are set:

zentgpfsn01:~ # less /var/lib/heartbeat/hostcache
zentgpfsn01     07ca44ca-1bf5-4f12-8680-21f86c2e6bca    100
zentgpfsn02     f44cbb3e-fa3c-4f93-b433-0c9eb4bb5cba    100
zentgpfsn03     7aa4698a-a17a-4c5b-8cfe-f7226a21aee8    100


when I start heartbeat on node1, /var/lib/heartbeat/hostcache looks like following

zentgpfsn01:~ # less /var/lib/heartbeat/hostcache
zentgpfsn01	07ca44ca-1bf5-4f12-8680-21f86c2e6bca	100	
zentgpfsn02	f44cbb3e-fa3c-4f93-b433-0c9eb4bb5cba	100	
zentgpfsn03	7aa4698a-a17a-4c5b-8cfe-f7226a21aee8	100	
zentgpfsn01	00000000-0000-0000-0000-000000000000	100


it seems as node1 can not found its own UUID and start w/o one. that may the reason for entries in logfile and the reboot:

ccm[19837]: 2008/04/30_09:02:05 ERROR: llm_add: adding same node(zentgpfsn01) twice(?)
ccm[19837]: 2008/04/30_09:02:05 ERROR: set_llm_from_heartbeat: adding node zentgpfsn01 to llm failed
ccm[19837]: 2008/04/30_09:02:05 ERROR: Initialization failed. Exit
heartbeat[19822]: 2008/04/30_09:02:05 WARN: Managed /usr/lib64/heartbeat/ccm process 19837 exited with return code 1.
heartbeat[19822]: 2008/04/30_09:02:05 EMERG: Rebooting system.  Reason: /usr/lib64/heartbeat/ccm



but when the system is up again, the same UUID is still set:

zentgpfsn01:~ # crm_uuid
07ca44ca-1bf5-4f12-8680-21f86c2e6bca



anybody an idea? i'm a bit helpless....


Dominik Klein schrieb:
>> here is the cause:
>>
>>>  ccm[19751]: 2008/04/29_10:59:59 ERROR: llm_add: adding same
>>> node(zentgpfsn01) twice(?)
>>>  ccm[19751]: 2008/04/29_10:59:59 ERROR: set_llm_from_heartbeat: 
>>> adding node
>>> zentgpfsn01 to llm failed
>>>  ccm[19751]: 2008/04/29_10:59:59 ERROR: Initialization failed. Exit
>>>  heartbeat[19737]: 2008/04/29_10:59:59 WARN: Managed
>>> /usr/lib64/heartbeat/ccm process 19751 exited with return code 1.
>>
>> you don't have more than one machine with the same name by any chance?
> 
> I saw something like this when I recently re-installed my testcluster 
> and used an old (backup) configuration file. The uuid changed and so 
> every node was there twice which ended in quite a mess.
> 
> Regards
> Dominik
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
> 
> 

-- 
Mit besten Grüßen / Best Regards

Alexander Födisch

Max Planck Institute for Evolutionary Anthropology
-Central IT Department-
Deutscher Platz 6
D-04103 Leipzig

Phone:  +49 (0)341 3550-168
     	+49 (0)341 3550-154
Fax:    +49 (0)341 3550-119
Email:  foedisch at eva.mpg.de

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 5905 bytes
Desc: S/MIME Cryptographic Signature
Url : http://lists.community.tummy.com/pipermail/linux-ha/attachments/20080430/d9b4bc0e/smime.bin