Issue Summary

With Pgpool-II, users can configure a clustered PostgreSQL database that allows for High Availability. We have documentation that provides steps on how to configure this with Jira.

The general configuration according to documentation mentioned here and here is to configure Pgpool-II to use a primary / standby configuration (active-passive):

pgpool-II can use its own replication capabilities or those provided by other software, but it is often recommended to use PostgreSQL’s streaming replication. Streaming replication is a feature that replicates databases by shipping transaction logs (WALs) of PostgreSQL (primary) to multiple instances of PostgreSQL (standby).

Pgpool-II provides a high availability (HA) feature by using multiple PostgreSQL servers so that it automatically removes broken server from the server pool to continue the database task. This is called automatic failover. Also Pgpool-II offers an HA feature for Pgpool-II itself, called Watchdog (see Chapter 4 for more details). Moreover Pgpool-II hires sophisticated quorum algorithm to avoid false positive errors and split brain problem to make the whole HA system highly reliable. See Section 5.15.6 for more details.

In some customer configurations, there are deployments that use an active-active setup instead, which has led to instances where duplicate issues will be created, causing an issue to become corrupted and unable to be transitioned.

Suggested Solution

Provide a way for Jira to handle active-active configurations of Pgpool-II that allows for issues to be created without duplicates appearing.

Why This Is Important

There are some instances where a duplicate ticket may be created, causing one of them to become invalid and unable to be transitioned to another workflow status.

Workaround

Using either a single PostgreSQL database or an active-passive configuration with Pgpool-ii allows for valid issues to be created as expected.