I posted this earlier in the week and have seen no replies so I figured I would 
post again.

Thanks.

Good afternoon,

I have a webfarm made up of Apache servers running a Joomla website - running 
on CentOS 5.5 servers. This also includes two separate data centers with 
haproxy running in each of them and a round robin dns with 15 minutes TTL's.

Here are snippets from my current haproxy.cfg file:

global
        maxconn 4096 # Total Max Connections. This is dependent on ulimit
        daemon
        nbproc 8 # Number of processing cores. Dual Dual-core Opteron is 4 core$
        log 127.0.0.1   local0
        log 127.0.0.1   local1 notice

defaults
        log     global
        option  httplog
        option  dontlognull
        retries 3
        option redispatch
        maxconn 10000
        contimeout 1800000
        clitimeout 1800000
        srvtimeout 1800000


listen webfarm xxx.xxx.xxx.xxx:80
        mode http
        balance roundrobin
        cookie SERVERID insert indirect nocache
        option forwardfor
        option httpchk HEAD /srvhealth.txt HTTP/1.0
        server web1.mysite.com xxx.xxx.xxx.221:80 cookie webfarm_web1 check
        server web2.mysite.com xxx.xxx.xxx.222:80 cookie webfarm_web2 check

These would be the same in each DC with the exception of the server name - 
which might be web10, etc. for the second DC.

I had reports today that some end users were getting the joomla inactivity 
screen during their session when navigating to different pages on the web site. 
It looked like they had jumped to the second data center which would explain 
why they got the screen. I have taken the second DC out of the mix for the time 
being.

I was still getting some reports of this inactivity screen when the user was on 
one or the other web servers in the single DC...and again it looks like they 
jumped from one server to the other. I know that the inactivity screen is not 
from haproxy but it seems that the end user is shifting web servers which is 
what is causing the issue.

Once the user connects we want them to continue to hit the same server until 
they log out of their session. Is this the correct configuration to accomplish 
it? Is there a way running multiple DC's to keep them from jumping to another 
DC after they already have a session open?

I thank you in advance for any guidance.

Please let me know if there is any additional data needed.


Scott Oyer
Network Administrator
FTC Family of Companies
[email protected]
309-664-1742






Reply via email to