Failover for Jira Data Center

As Jira becomes a mission critical resource in your organization, maximizing application uptime becomes a prime consideration. Data Center is your solution for High Availability and performance at scale via clustering technology.

Data Center achieves High Availability via a multi-node cluster where-by all nodes are active, there are no passive nodes. For example if you have a three node cluster all serving requests when one node fails the remaining two nodes will handle all requests. This gives the administrator the opportunity to fix the failed node without impacting end users. Users on the failed node will be automatically redirected to one of the other nodes upon their next request and will be required to login again to the new node.

We recommend that your users check the "Remember me" checkbox at login so that they can login faster if their node goes down. Please also note that session based tasks such as 'bulk edit' will need to be restarted if a node goes down during the task. 

Below is the basic architectural diagram for Data Center:

Failover in the event of an outage is handled by the load balancer referenced in the system diagram above. The Data Center application is load balancer agnostic, meaning any type of hardware or software based load balancer will work, given that it has been configured with cookie based session affinity. You can administer the load balancer as you would with any other application such that traffic is directed to nodes in a round-robin fashion or based on concurrent load, whatever works best for your organization. Furthermore the load balancer can be used to direct particular types of traffic to specific nodes.

Data Center also provides disaster recovery failover in the event of a system-wide outage. An offsite disaster recovery system can be deployed in "disaster recovery mode." The Jira application will share application indexes from production to this disaster recovery site such that in the event of a system wide outage the disaster recovery system can come online much more quickly. To activate the disaster recovery system, first the system must be turned on, then traffic will need to be directed away from the production environment and to the disaster recovery system, this is typically done through DNS rules. Provided that the database and shared file system is regularly backed up from production to the DR site, downtime will be dramatically reduced even in the event of a disaster.

For more information on configuring the load balancer see the Data Center Load Balancer examples.