www-apache-bugdb mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marc Slemko <ma...@znep.com>
Subject Re: general/1459: Please shorten "server seems busy" log message -- patch included
Date Sun, 23 Nov 1997 07:10:00 GMT
The following reply was made to PR general/1459; it has been noted by GNATS.

From: Marc Slemko <marcs@znep.com>
To: "M. D. Parker" <mdpc@netcom.com>
Cc: Apache bugs database <apbugs@apache.org>
Subject: Re: general/1459: Please shorten "server seems busy" log message --        patch
included
Date: Sat, 22 Nov 1997 22:57:31 -0700 (MST)

 On Sat, 22 Nov 1997, M. D. Parker wrote:
 
 > Hmmm...Interesting observations.
 > 
 > First, I like the new spawning algorithm a lot better than 
 > the old one.  The one process per second was causing some bottlenecks
 > in a very active / load variable web site I manage.  The range
 > is between 100 and 300 processes and this site.
 > 
 > The logs under the current system flag the "server busy" message
 > generally when you go BELOW the MinSpare # of servers.  Maybe that
 > is not exactly what is needed.  The notice under the current system
 > should be sent when a load situation is presented, not as a result for
 > restoration of MinSpare servers which is what it looks like now.
 > 
 > Anyway to get how many requests are pending but not servicable NOW
 > due to lack of processes?  Maybe that might be an interesting statistic/
 
 No.  (well, not under the sockets API without hacking the kernel)
 
 > trigger for such a message.  With Min/Max servers parameters you
 > are trying to anticipate a number of concurrent requests.  Right now
 > that seems to be a difficult number to obtain/monitor/estimate.
 > Maybe the trigger could be configurable as a server run-time directive.
 

Mime
View raw message