[rabbitmq-discuss] Bad cookie when clustering

[Peter Fitzgibbons]

Hello All,

I have servers : vmrabbita, vmrabbitb, vmrabbitc
RabbitMQ 1.7.1
All 3 have the same /var/lib/rabbitmq/.erlang.cookie

a and c are clustered successfully.

b gives me this

 rabbitmqctl cluster rabbit at vmrabbita
Clustering node rabbit at vmrabbitb with [rabbit at vmrabbita] ...
Error: {unable_to_join_cluster,
           [rabbit at vmrabbita],
           {merge_schema_failed,
               "Bad cookie in table definition rabbit_durable_queue:
rabbit at vmrabbitb = {cstruct,rabbit_durable_queue,set,[],[rabbit at vmrabbitb
],[],0,read_write,[],[],false,amqqueue,[name,durable,auto_delete,arguments,pid],[],[],{{1266,16863,365586},rabbit at vmrabbitb},{{2,0},[]}},
rabbit at vmrabbita = {cstruct,rabbit_durable_queue,set,[],[rabbit at vmrabbitc
,rabbit at vmrabbita
],[],0,read_write,[],[],false,amqqueue,[name,durable,auto_delete,arguments,pid],[],[],{{1266,14484,94839},rabbit at vmrabbita
},{{4,0},{rabbit at vmrabbitc,{1266,16151,144423}}}}\n"}}

The machines are VMs and actually clones of each other, so I have no belief
that there is any "corruption" going on.
I have uninstalled and reinstalled rabbitmq-server on b.

I have removed the Mnesia db at /var/lib/rabbitmq/mnesia/rabbit and
restarted rabbitmq-server

There is nothing important on the servers.  Any forceful-reset including
data-deletion is welcome.

Could someone direct me to the appropriate change to try to fix this?


Peter Fitzgibbons
(847) 859-9550
Email: peter.fitzgibbons at gmail.com
IM GTalk: peter.fitzgibbons
IM AOL: peter.fitzgibbons at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20100212/a5472082/attachment.htm