Setting High Availability in JIRA Datacenter with PostgreSQL database

[Hamza REZGUI]

Hi,

We are setting High Availability in our JIRA PostgreSQL database for Jira datacenter.

We would like to know how to specify two postgresql datasources (one open in read-write, another one open in read-only).

Particularly is there is a way for JIRA to distinguish read-only and read-write PostgreSQL servors

(In case of database service issue, our PostgreSQL clusters will databases will failover (automatically).

Thank you

[Nic Brough -Adaptavist-]

You do not.

Jira uses a single database source.  You can replicate your database (or database cluster) over to another database if you want, but that will be a backup, you cannot wire a Jira into it until you've made it read-write and stopped replication from the main system.

Jira can distinguish between read-only and read-write.  If you start it up while pointed to a read-only database, it will shut down as soon as it realises it does not have full access to the database.

Could you explain what you are trying to achieve here?

[Hamza REZGUI]

Hello Nic,

First of all, Happy new year and thank you for the response.

We are trying to point Jira to two databases:
- primary database (read,write) (master)
- standby database (read only) (slave)

The goal is if the master database server down, the slave replaces it.

Regards.
Hamza

[Nic Brough -Adaptavist-]

Monika?

Please could you tell us what you are trying to achieve here?  Please don't repeat the same question again - your question was perfectly clear that you want some form of failover.

Why are you trying to do this?  What does the end-user get out of it?  Are you looking for high availability, some form of disaster recovery, performance, or something else for example?

[Hamza REZGUI]

sorry Nic :) we are looking for database high availability

[Nic Brough -Adaptavist-]

Why?  What problem would "database high availability" solve?

Do you mean general "high availability" of a Jira system?

[Hamza REZGUI]

yes it is, HA for Jira system for production grade

[Nic Brough -Adaptavist-]

Ok, thanks!  I know it's been a bit repetetive, but it's important that I don't answer the wrong question.

Database high-availability is a component of high-availability for a system, but you do not do it at the application level.  In fact, you can't do it at the application level if you've got a clustered application.

If you want a highly available database to support your HA Jira DC cluster, then you need an HA database service (almost certainly a cluster).  Replication and secondary sources are not highly available database solutions, they're backup and failover solutions.

[Hamza REZGUI]

Nic, thank you for your explanations