[erlang-questions] distribution trouble on Centos 5.2

Paul Fisher <>
Wed Apr 15 20:18:19 CEST 2009


Matt Handler wrote:
> Trying to get communication between nodes to work in erlang, but having 
> trouble (both nodes are running R12B5 with the patch).  I'm following 
> the steps from When Distribution Won't Work 
> <http://erlang.org/faq/problems.html#9.9>:
> 
>    1. start  <mailto:> ,
>        <mailto:> (both with
>       long names, both with same cookie, using -rsh ssh)
>    2. kill epmd and restart it with -d -d for debugging mode
>    3. from littleguy I call *net:ping('
>       <mailto:>').* and it returns *pang*
>    4. no messages go by on epmd on 
>       <mailto:>
>    5. running *tcpdump port 4369* on 
>       <mailto:> and pinging again from
>        <mailto:>, i see that
>       the messages are in fact coming in:
> 
> listening on eth0, link-type EN10MB (Ethernet), capture size 96 bytes
> 09:45:22.158227 IP server2.net.40141 > server1.net.epmd: S 
> 2310861565:2310861565(0) win 5840 <mss 1460,sackOK,timestamp 1087186457 
> 0,nop,wscale 7>
> 
> so it looks like the messages are making it between the servers, but not 
> making it to erlang.  any ideas?
> 

Try the following:
1) verify the cookie is the same on both nodes (erlang:get_cookie/0)
2) verify both vms are being started with either -name or -sname


--
paul



More information about the erlang-questions mailing list