Skip to content
You are reading Orion development version documentation and some displayed features may not be available in the stable release. You can switch to stable version using the version box at screen bottom.

High availability

Orion supports deploying more than one instance sharing the same database.

By placing the instances behind a load balancer, downtime can be limited during maintenance operations.

Orion HA

Storage

When configuring for high availability, both storage and known nodes storage must match so Orion provides up to date information. Known nodes are discovered Orion nodes.

Tip

Use the knownnodes client API methods to retrieve the known nodes.

Use a relational database to enable highly available scenarios. Orion supports relational databases PostgreSQL and Oracle.

Orion also supports LevelDB and MapDB stores. LevelDB and MapDB lock the local file system and are not suitable for highly available scenarios.

For high availability, ensure the storage and knownnodesstorage in the configuration file match for all Orion nodes.

Load balancer configuration

The load balancer must expose both client and node interfaces.

When configuring for high availability, configure the nodes in the Orion cluster (Orion A and Orion B in the above diagram) with the same set of keys and advertise the load balancer address.

Nginx configuration with two Orion nodes

 events { }

 http {
     upstream orion_8080 {
         server orion_1:8080 max_fails=3 fail_timeout=5s;
         server orion_2:8080 max_fails=3 fail_timeout=5s;
     }

     upstream orion_8888 {
         server orion_1:8888 max_fails=3 fail_timeout=5s;
         server orion_2:8888 max_fails=3 fail_timeout=5s;
     }

     server {
         listen 8080;

         location / {
             proxy_pass http://orion_8080;
         }
     }

     server {
         listen 8888;

         location / {
             proxy_pass http://orion_8888;
         }
     }
 }

The configuration defines two upstreams: orion_8080 and orion_8888.

Both upstreams define health checks: max_fails=3 fail_timeout=5s

The health checks help Nginx balance traffic among upstream servers.

ConsenSys has acquired Quorum from J.P. Morgan. Please read the FAQ.
Questions or feedback? You can discuss issues and obtain free support on Orion Discord channel.
For paid professional support by Consensys, contact us at quorum@consensys.net