HAProxy and “sharding”

haproxyscaling

I am cross posting this from stack overflow since it was suggested I might get a better answer here …

I was wondering if anyone has used HAProxy for sharding. Specifically, I would like to be able use a cookie I define w/a identifier and have any client that presents that cookie/identifier combination be directed to the same server. It seems possible and my preliminary tests seem to work but I am not a sure.

My listen configuration looks like:

listen  appli1-rewrite 0.0.0.0:80
cookie  mysession= prefix
balance roundrobin
server  app1_1 192.168.0.1:80 cookie server01
server  app1_2 192.168.0.2:80 cookie server02
option  httpclose

However, when I look at the request headers using Live Http Headers I don't see the mysession cookie indicating which server it will be going to, which is what I expected based on the prefix operation. I am quite new to HAProxy so any help/guidance would be great!

Best Answer

first, remove the equal sign after your cookie name, otherwise it will never match. Second, please ensure that your server correctly sets the "mysession" cookie, as the cookie prefix mode relies on the server to set the cookie. If it does not, use the insert mode instead. Third, please add "option httpclose" in your config, because if there is any sort of keepalive, haproxy won't see the cookie (or will not be able to mangle it).

Hoping this helps,
Willy

Related Solutions

HAProxy configuration and modification

How it works Basically it's a load balancer. You put HAProxy in front of Web Servers, and it will split the load/traffic between them based on the balance algorithm you chose.
How do I configure it? It has a single config file /etc/haproxy/haproxy.cfg . For the most simple configuration you define one frontend with one or more backends (the web servers behind haproxy). Something like this:
```
listen frontend *:80
   mode http
   balance roundrobin
   server backend1 192.168.1.2:80 cookie A check
   server backend2 192.168.1.3:80 cookie B check
```
How do I modify(add/remove) the already configured file? You just need to edit the file, At this time doesn't exist the option to use include files for configurarion. To apply the changes you need to restart haproxy.
How do I start/restart the HA Proxy after modifying the configuration file? If you have installed haproxy from a ubuntu/centos package you need:

/etc/init.d/haproxy restart
Even I don't know how to check my current installed HA Proxy version? If you're using CentOS:

rpm -qi haproxy

Check this document: http://haproxy.1wt.eu/download/1.3/doc/architecture.txt . It will guide you through different architectures for differente needs.

Way to add more backend server to haproxy without restarting haproxy

I haven't tested this specific use case but haproxy does support a "hot reload":

2.4.1) Hot reconfiguration
--------------------------
The '-st' and '-sf' command line options are used to inform previously running
processes that a configuration is being reloaded. They will receive the SIGTTOU
signal to ask them to temporarily stop listening to the ports so that the new
process can grab them. If anything wrong happens, the new process will send
them a SIGTTIN to tell them to re-listen to the ports and continue their normal
work. Otherwise, it will either ask them to finish (-sf) their work then softly
exit, or immediately terminate (-st), breaking existing sessions. A typical use
of this allows a configuration reload without service interruption :

 # haproxy -p /var/run/haproxy.pid -sf $(cat /var/run/haproxy.pid)

If you have an init script to start and stop haproxy it likely supports the reload argument with a function like:

haproxy_reload()
{
    $HAPROXY -f "$CONFIG" -p $PIDFILE -D $EXTRAOPTS -sf $(cat $PIDFILE) \
        || return 2
    return 0
}

Best Answer

Related Solutions

HAProxy configuration and modification

Way to add more backend server to haproxy without restarting haproxy

Related Topic