TCP load balancing in HAProxy with persistent TCP connections

haproxyload balancingpersistent-connectiontcp

Currently, I have a single client (component) having a single persistent TCP connection with the server(another component), and messages are exchanged asynchronously. I wanted to have a load balancer (HAProxy preferably) where the connection b/w client and load balancer as well as b/w load balancer and multiple servers as persistent TCP connection.

I know HAProxy can easily be set up for the TCP load balancing, but I wanted to know does it support persistent connections out of the box. Would be great help if someone points me in the right direction. Thanks.

Best Answer

Yes it supports persistent TCP connections right out of the box. A simple implementation in the haproxy looks like this

listen tcpProxy 0.0.0.0:7000 
    mode tcp 
    balance source
    server tcp1 ip1:port check maxconn 10000
    server tcp2 ip2:port check maxconn 10000

Hope this helps

Related Solutions

The theoretical maximum number of open TCP connections that a modern Linux box can have

A single listening port can accept more than one connection simultaneously.

There is a '64K' limit that is often cited, but that is per client per server port, and needs clarifying.

Each TCP/IP packet has basically four fields for addressing. These are:

source_ip source_port destination_ip destination_port
<----- client ------> <--------- server ------------>

Inside the TCP stack, these four fields are used as a compound key to match up packets to connections (e.g. file descriptors).

If a client has many connections to the same port on the same destination, then three of those fields will be the same - only source_port varies to differentiate the different connections. Ports are 16-bit numbers, therefore the maximum number of connections any given client can have to any given host port is 64K.

However, multiple clients can each have up to 64K connections to some server's port, and if the server has multiple ports or either is multi-homed then you can multiply that further.

So the real limit is file descriptors. Each individual socket connection is given a file descriptor, so the limit is really the number of file descriptors that the system has been configured to allow and resources to handle. The maximum limit is typically up over 300K, but is configurable e.g. with sysctl.

The realistic limits being boasted about for normal boxes are around 80K for example single threaded Jabber messaging servers.

Linux – How to Add/Remove New Server behind Haproxy with Minimal Disruption

For the first question: how to add a third server. Changing the haproxy configuration to add C will require a restart of the load balancer. But, http://www.mgoff.in/2010/04/18/haproxy-reloading-your-config-with-minimal-service-impact/ says that a combination of -st and -sf command line arguments will minimize the disruption by allowing the haproxy instances to hand off the ports they listen to. Implementing this may require that you change your haproxy init.d script.

For the second: If you use a firewall to block A then traffic bound for A will go to B and C or just to C, depending on the haproxy configuration. This may disrupt a bit of customer traffic if the sessions have stickiness. If the application's sessions are stateless then the changes will be invisible to clients. If you shut down A after blocking its traffic then your customers will not notice.

Best Answer

Related Solutions

The theoretical maximum number of open TCP connections that a modern Linux box can have

Linux – How to Add/Remove New Server behind Haproxy with Minimal Disruption

Related Topic