From: Robert Klemme <shortcutter@...>
Date: 2009-04-17T16:03:30+09:00
Subject: Re: Weird performance issue

2009/4/17 Ryan Davis <ryand-ruby@zenspider.com>:
>
> On Apr 16, 2009, at 16:25 , court3nay wrote:
>
>> Hey, I'm debugging a weird performance issue. I have two VMs on the
>> same server running the same OS (arch linux) and the same version of
>> "enterprise" Ruby (ruby 1.8.6 (2008-08-08 patchlevel 286) [x86_64-
>> linux]).
>>
>> On one machine, I get these numbers
>>
>> irb(main):004:0> Benchmark.realtime { require 'rubygems' }
>> => 0.0482780933380127
>> irb(main):005:0> Benchmark.realtime { require 'tinder' }
>> => 0.395686149597168
>>
>> On the other,
>>
>> irb(main):002:0> Benchmark.realtime { require 'rubygems' }
>> => 1.20596885681152
>> irb(main):003:0> Benchmark.realtime { require 'tinder' }
>> => 5.7661280632019
>
> same number of gems on both slices?
>
> % gem list | wc -l
>
> you can also do something really gross like:
>
> % ruby -rtracer -e 'require "rubygems"' > server1.txt
>
> I don't think you should necessarily diff the two, but looking at rough
> things like # of lines should start to explain stuff. Something maybe like:
>

If that does not work then "strace" might be an option as well.  That
usually also contains timestamps and you might be able to spot a
syscall which is taking a lot of time.

Kind regards

robert


-- 
remember.guy do |as, often| as.you_can - without end
http://blog.rubybestpractices.com/