Network communications fails with R13B02-1 on OS X

[Dave Bryson]

Backing down to otp R 12B-5  the problem disappears and all works as  
expected. One thing I noticed:

Each registering node on 12B-5 (epmd) has a different number AFTER the  
colon  ":"

   epmd: Thu Oct  1 12:51:31 2009: registering 'n1:2', port 49370
   ...
   epmd: Thu Oct  1 12:51:31 2009: registering 'n2:3', port 49372

Notice 'n1:2' and 'n2:3'

Where as on otp R13 below, the number AFTER the colon ':' is the same:
  Notice below: 'n1:2' and 'n2:2'

Could this be the issue?

Dave

On Oct 1, 2009, at 1:00 PM, Dave Bryson wrote:

> After more debugging I'm still baffled as to why two nodes cannot  
> communicate on the same host.  I started epmd with the debug flag  
> and see the nodes are registering but the net:ping does not seem to  
> make the PORT request to epmd:
>
> epmd: Thu Oct  1 12:51:25 2009: epmd running - daemon = 0
> epmd: Thu Oct  1 12:51:31 2009: ** got ALIVE2_REQ
> epmd: Thu Oct  1 12:51:31 2009: registering 'n1:2', port 49353
> epmd: Thu Oct  1 12:51:31 2009: type 77 proto 0 highvsn 5 lowvsn 5
> epmd: Thu Oct  1 12:51:31 2009: ** sent ALIVE2_RESP for "n1"
> epmd: Thu Oct  1 12:51:37 2009: ** got ALIVE2_REQ
> epmd: Thu Oct  1 12:51:37 2009: registering 'n2:2', port 49355
> epmd: Thu Oct  1 12:51:37 2009: type 77 proto 0 highvsn 5 lowvsn 5
> epmd: Thu Oct  1 12:51:37 2009: ** sent ALIVE2_RESP for "n2"
>
> Note: epmd is NOT showing a PORT2_REQ  from the net:ping().  I know  
> the code I'm testing with is correct because it works on Ubuntu.
>
> Does anyone have ANY idea why this is happening??
>
> Thanks in advance!
> Dave
>
> Begin forwarded message:
>
>> From: Dave Bryson <daveb@REDACTED>
>> Date: October 1, 2009 9:17:41 AM CDT
>> To: erlang-questions@REDACTED
>> Subject: Network communications fails with R13B02-1 on OS X
>>
>> Yesterday I upgraded to otp_13B02-1 from R12-5 on OS X 10.5.8. I  
>> notice today that any form of network communication between nodes  
>> seems to fail. Trying to start slaves on the same host (which I was  
>> able to do before the upgrade) fails with an {error,timeout}.  Even  
>> doing a simple 'net:ping(node1@REDACTED)' on the same host between  
>> two nodes fails with a pang.   I can run the same exact tests on  
>> Ubuntu with R13B02-1 and they pass with flying colors.  Has anyone  
>> else noticed this problem on OS X?  Any ideas what the problem may  
>> be?
>>
>> Thanks!
>> Dave
>