From: Jim Freeze <jim@...>
Date: 2005-05-16T05:47:25+09:00
Subject: Re: Why is this DRb code so slow?

* Dick Davies <rasputnik@hellooperator.net> [2005-05-16 02:45:29 +0900]:

> I'd guess DNS. check your forward and reverse for both the client
> and the server.

That is what I was thinking too, but then I ran it on a different
machine, and I get a no response error instead of a delayed result.

In the two cases below I use 

  DRb.start_service nil, obj

and

  DRb.start_service dri, obj

where dri is the value originally picked by DRb itself:

  "druby://rabbit:50827"

  # start the server - let drb decide
  % ruby drb.server
  URI: 'druby://rabbit:50827'
  I am server

  # run the client
  % ruby drb.client
  #<DRb::DRbObject:0x821d804 @uri="druby://rabbit:50827", @ref=nil>
  68217858
  No response from server
  druby://rabbit:50827 - #<Errno::ECONNREFUSED: Connection refused -
  connect(2)>

  # kill the old server, and manually do start_service at the previous dri:
  % ruby drb.server
  URI: 'druby://rabbit:50827'
  I am server 

  % ruby drb.client
  #<DRb::DRbObject:0x821d804 @uri="druby://rabbit:50827", @ref=nil>
  68217858
  ["test1", "test2"]
  # it works


So, why does the client fail if the server is running at the same
address. The only difference is that I set the address explicitly
or I let DRb choose the address.

-- 
Jim Freeze
Ruby: I can explain it to ya but I can't understand it fer ya.