22.5. Setting up HAProxy as a load balancer

22.5.1. Installing HAProxy
22.5.2. Configuring HAProxy
22.5.3. Configuring separate sets for master and slaves
22.5.4. Cache-based sharding with HAProxy

In the Neo4j HA architecture, the cluster is typically fronted by a load balancer. In this section we will explore how to set up HAProxy to perform load balancing across the HA cluster.

22.5.1. Installing HAProxy

For this tutorial we will assume a Linux environment. We will also be installing HAProxy from source, and we’ll be using version 1.4.18. You need to ensure that your Linux server has a development environment set up. On Ubuntu/apt systems, simply do:

aptitude install build-essential

And on CentOS/yum systems do:

yum -y groupinstall 'Development Tools'

Then download the tarball from the HAProxy website. Once you’ve downloaded it, simply build and install HAProxy:

tar -zvxf haproxy-1.4.18.tar.gz
cd haproxy-1.4.18
make
cp haproxy /usr/sbin/haproxy

Or specify a target for make (TARGET=linux26 for linux kernel 2.6 or above or linux24 for 2.4 kernel)

tar -zvxf haproxy-1.4.18.tar.gz
cd haproxy-1.4.18
make TARGET=linux26
cp haproxy /usr/sbin/haproxy

22.5.2. Configuring HAProxy

HAProxy can be configured in many ways. The full documentation is available at their website.

For this example, we will configure HAProxy to load balance requests to three HA servers. Simply write the follow configuration to /etc/haproxy.cfg:

global
    daemon
    maxconn 256

defaults
    mode http
    timeout connect 5000ms
    timeout client 50000ms
    timeout server 50000ms

frontend http-in
    bind *:80
    default_backend neo4j

backend neo4j
    server s1 10.0.1.10:7474 maxconn 32
    server s2 10.0.1.11:7474 maxconn 32
    server s3 10.0.1.12:7474 maxconn 32

listen admin
    bind *:8080
    stats enable

HAProxy can now be started by running:

/usr/sbin/haproxy -f /etc/haproxy.cfg

You can connect to http://<ha-proxy-ip>:8080/haproxy?stats to view the status dashboard. This dashboard can be moved to run on port 80, and authentication can also be added. See the HAProxy documentation for details on this.

22.5.3. Configuring separate sets for master and slaves

It is possible to set HAProxy backends up to only include slaves or the master. For example, it may be desired to only write to slaves. To accomplish this, you need to have a small extension on the server than can report whether or not the machine is master via HTTP response codes. In this example, the extension exposes two URLs:

  • /hastatus/master, which returns 200 if the machine is the master, and 404 if the machine is a slave
  • /hastatus/slave, which returns 200 if the machine is a slave, and 404 if the machine is the master

The following example excludes the master from the set of machines. Request will only be sent to the slaves.

global
    daemon
    maxconn 256

defaults
    mode http
    timeout connect 5000ms
    timeout client 50000ms
    timeout server 50000ms

frontend http-in
    bind *:80
    default_backend neo4j-slaves

backend neo4j-slaves
    option httpchk GET /hastatus/slave
    server s1 10.0.1.10:7474 maxconn 32 check
    server s2 10.0.1.11:7474 maxconn 32 check
    server s3 10.0.1.12:7474 maxconn 32 check

listen admin
    bind *:8080
    stats enable

22.5.4. Cache-based sharding with HAProxy

Neo4j HA enables what is called cache-based sharding. If the dataset is too big to fit into the cache of any single machine, then by applying a consistent routing algorithm to requests, the caches on each machine will actually cache different parts of the graph. A typical routing key could be user ID.

In this example, the user ID is a query parameter in the URL being requested. This will route the same user to the same machine for each request.

global
    daemon
    maxconn 256

defaults
    mode http
    timeout connect 5000ms
    timeout client 50000ms
    timeout server 50000ms

frontend http-in
    bind *:80
    default_backend neo4j-slaves

backend neo4j-slaves
    balance url_param user_id
    server s1 10.0.1.10:7474 maxconn 32
    server s2 10.0.1.11:7474 maxconn 32
    server s3 10.0.1.12:7474 maxconn 32

listen admin
    bind *:8080
    stats enable

Naturally the health check and query parameter-based routing can be combined to only route requests to slaves by user ID. Other load balancing algorithms are also available, such as routing by source IP (source), the URI (uri) or HTTP headers(hdr()).