[MITgcm-devel] forge is slow: too many "httpd" proc ?

[Constantinos Evangelinos]

On Friday 07 May 2010 08:50:28 am marshall-admin at techsquare.com wrote:
> it seems that a long-term solution may be
> required. perhaps a 32bit box with 2G ram
> isn't quite enough for all the traffic, etc.

If we were running something really popular maybe but I doubt our regular 
traffic is that high! Do we know which originator is causing all that traffic 
after 11am? I have some testreport scripts that download the code tarball 
around 11:15am and they failed:

wget http://mitgcm.org/download/daily_snapshot/MITgcm_ss_20100506.tar.gz
--23:15:28--  
http://mitgcm.org:80/download/daily_snapshot/MITgcm_ss_20100506.tar.gz
           => `MITgcm_ss_20100506.tar.gz'
Connecting to mitgcm.org:80... connected!
HTTP request sent, awaiting response... 
Read error (Connection reset by peer) in headers.
Retrying.

> jmc, are you around today (friday) ?
> perhaps we can consider some optimization
> strategies...

Isn't there a way to limit the number of httpd threads?

Constantinos
PS> I removed JMC and Dimitri from the recepients list so that they don't get 
too many copies of this e-mail.
-- 
Dr. Constantinos Evangelinos
Department of Earth, Atmospheric and Planetary Sciences
Massachusetts Institute of Technology