Okay great. Thanks for the reply. I have done the following.
I have changed the following line to track the 'serving' time for a page.
LogFormat "%h %l %u %t \"%r\" %>s %b \"%{Referer}i\" \"%{User-Agent}i\"
**%T/%D**" combined (added %T/%D)
For number of processes for httpd:
With a single device, after the pfdhcplistener is done, for the first ever
request to the captive portal, I saw 10 httpd processes. But for all the
subsequent requests, the number of processes were between 3-4. This was for a
single device on the VLAN.
I also couldn't find the 'worker' module configuration anywhere. So I think
httpd is running with default configuration as far as threads, processes,
connections etc.I think, I should be adding the following sample configuration
somewhere, can you tell me here would this go?
ServerLimit 16StartServers 2MaxClients 200MinSpareThreads 25MaxSpareThreads
75ThreadsPerChild 25
I have turned on the slow query log in mysql and no queries have showed up so
far, been running for past 3 hours.
Anything else that I can look at?
From: [email protected]
Date: Tue, 16 Jun 2015 13:49:42 -0400
To: [email protected]
Subject: Re: [PacketFence-users] Caching for Captive Portal
On Jun 16, 2015, at 9:35 , Andy A <[email protected]> wrote:Hello
What's the best way to speed-up the serving of captive portal pages to reduce
latency?Currently the latency of each page is quite high in 7-8 seconds bracket.
Hi Andy,Before anything you really need to know where the bottleneck is.E.g.
Improvements to the webserver (for instance) will not help if the database is
at fault.
Try to see if the httpd is hitting the maximum number of processes it can
handle concurrently for a start. You may be swamped by what we colloquially
call “parked devices”, i.e. devices that constantly query the portal even
though no one is actually going to register (mostly phones).Or you may have a
single device that is trying to open hundreds of connections to the portal.
If that’s not the case, try to see how long it takes to serve each query.
Apache can log the time per requests.
Enable the slow query log in mysql. Etc.
Your question unfortunately does not lend itself to a simple answer.Each server
is unhappy in it’s own way.
Regards,
--
Louis Munro
[email protected] :: www.inverse.ca
+1.514.447.4918 x125 :: +1 (866) 353-6153 x125
Inverse inc. :: Leaders behind SOGo (www.sogo.nu) and PacketFence
(www.packetfence.org)
------------------------------------------------------------------------------
_______________________________________________
PacketFence-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/packetfence-users
------------------------------------------------------------------------------
_______________________________________________
PacketFence-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/packetfence-users
