Can I use Amazon RDS multi-zones with Jira SERVER?

Gerald_Blondel

Hi Everyone,

I know that Jira SERVER (not talking about Data Center here) is compatible with Amazon RDS PostgreSQL. This will be my setup.

But ... can I activate multi-zones to at least have a replica of my database available in another zone. Then I could start a stand-by Jira server and connect to it?

Thanks!

Andy Heinzer

Hi Gerald,

I understand that you are looking to use Jira Server with an Amazon RDS Postgresql database, and then also use this database in a manner where the database will have replicas of itself in different zones.

This might be possible to do, but this is not something that Atlassian would support for Jira Server. Please see our Jira Server Supported platforms documentation.  Technically, only Jira Data Center has support for Amazon RDS Postgresql.  I would not be surprised if you were able to get Jira Server working with an Amazon RDS Postgresql database. Because it's not too deviated from the 'vanilla-flavored' Postgresql we expect for Jira Server, but it's not exactly the same. This is not something we have extensively tested for in Jira Server.  Namely because we see this scenario as being something that is better managed in Data Center.

Only the Data Center line of Jira technically has support for the Amazon RDS postgresl database type.

That said, I am interested to learn more about your use case here to see if we can help here.  Your purposed configuration here has some potential problems as I can seem them for a Jira Server installation.  Namely, what happens if you have another Jira install that tries to start up on that database when your main Jira is also using the main database?  This is not an expected configuration for Jira Server.  You might get a startup failure error on the 2nd machine due to a database lock of sorts, but if you didn't I could see this be a potentially corrupting aspect of the data there.  Or, what happens if your main Jira site goes down, your standup starts up, and then the main site restarts itself?  This scenario of having two Jira Servers trying to read/write changes to what is essentially the same database (even if located in a different address/location), is something I would expect to cause all kinds of data corruption and unexpected behavior.  This kind of scenario is not expected to be something Jira Server can handle.  This is something that Jira Data Center has been built to be able to manage.

Does the multi-zone aspects here replicate data in both directions?  Or is this a one way sync?

Aside from the database aspects here, this method of won't be able to sync all the data of Jira.  For the Jira database does not contain file attachments or avatars that Jira uses.  These are actually stored on the $JIRAHOME/data/ folder of the Jira Server itself. 

I hope this helps.

Andy