|
Still
doesn’t tell me that Apache is the problem. My experience tells me that when
there is a custom application involved, it’s usually the problem. Anyway,
I’ll throw some things out there for you to consider. Look at
your database connections and how Moodle is configured to use them (are they
being closed appropriately, are the being pooled, pool size, timeouts, etc.) If you
are using a prefork and your MaxClients is at
the default of 150 and you only have 45 processes then it could be that the OS
can’t allocate any more connections. More on
the MaxClients directive here including how to adjust
settings for threaded and hybrid servers: http://httpd.apache.org/docs/2.2/mod/mpm_common.html#maxclients I don’t
think having a lot of sockets in a TIME_WAIT state for
Apache is necessarily bad. But if
the site is becoming unusable because of the load, in the past I have completely
disabled KeepAlives to make a site more responsive while I
figured out the real problem. But that was for a site that was 90% static
content so I don’t know if it would have the desired effect on your system. Lastly,
here’s the performance doc on the Apache site: http://httpd.apache.org/docs/2.2/misc/perf-tuning.html Let us
know how things go. If you still have problems, I’m interested in answers
to these questions: what’s the status of the web site? Is it still
responsive or is it slow? Can you identify requests that are slower than others?
Can you test responsiveness of a requests for static content versus dynamic? Cheers, Jeff From:
Kory Wheatley [mailto:wheakory@xxxxxxxxx] Apache was fine, until school started when more
students started using the web server. There's just an over abundance of
Apache processes in a TIME_WAIT state when you do a "netstat
-na". What parameters in apache do you recommend I look into, the
keepalive settings, maxclient settings. The memory on the system not an
issue, its just the CPU usage. On 9/5/07, Jeff
Beard <jeffb@xxxxxxxxxxxxx>
wrote: Hi there, I disagree with your assessment
of the problem. You haven't provided any information that would clearly
indicate to me that there's an issue with Apache performance. I recommend
looking at the application and database first. Especially if, for example,
things were fine last week but this week there's a resource problem. That would
indicate to me that you should examine changes made to the application and
database in the interim. Sorry I can't help more but I
think there's more analysis to do before we can make a call on what the problem
is much less offer advice on fixing it. Cheers, Jeff From: Kory Wheatley
[mailto:wheakory@xxxxxxxxx]
Our
apache server is utilizing the CPU to almost a 100% with 45 httpd process
opened. What sort of tuning could be done to help this
situation. |
|