[rabbitmq-discuss] Unable to get automatic clustering to work

Matthias Radestock matthias at rabbitmq.com
Wed Jan 19 08:55:44 GMT 2011


Mark J. Reed wrote:
> OK, I'm having better luck with 2.2.   But I still have a few
> instances where one node can't talk to another, even though TCP
> connectivity is there.  Any help diagnosing those?
> [...]
> rabbit2$ rabbitmqctl -n rabbit at rabbit1  # but 2 can't connect to 1
> Status of node slurp at rabbit1 ...
> Error: unable to connect to node slurp at slurpp1refq1: nodedown
> diagnostics:
> - nodes and their ports on slurpp1refq1: [{rabbit,12635}]
> - current node: rabbitmqctl7085 at slurpp1refw1
> - current node home dir: /var/lib/rabbitmq
> - current node cookie hash: aEuWObzDves/lfh1+WJREQ==

That doesn't look like a valid rabbitmqctl command line. It's missing 
the command, i.e. "status". Also, how come the output mentions 
slurp at rabbit1 and slurp at slurpp1refq1?

The "nodes and their ports" line tells you that there is a node called 
'rabbit' running on the slurpp1refq1 machine. So I suspect
   rabbitmqctl -n rabbit at slurpp1refq1 status
would work, except ...

Erlang node names are oblivious to host name aliasing. So if you have a 
machine that can be addressed as both 'slurpp1refq1' and 'rabbit1', and 
rabbit on the machine has a node name of 'rabbit at rabbit1' (check 
rabbit's startup log) then you must refer to that node exactly like 
that; trying to contact it as 'rabbit at slurpp1refq1' will not work.



More information about the rabbitmq-discuss mailing list