How can I programmatically restore JIRA to a new machine?

[Chris Wright]

We want JIRA hosted in the cloud, in AWS.

Backup/restore of the Jira DB isn't an issue, but the problem is when I spin up a new instance (i.e. something went wrong with the old instance), I want a way to automatically set up everything with our old settings.

I couldn't find anything in Atlassian's support for this.

The best I could see was using the Linux .bin, installing that, using the browser to set up the database and admin and license, then copying over the server.xml and restoring from the Database backup... which seems pretty clunky.

The .bin didn't seem to support --help or passing any parameters to avoid needing to interact with it. The same thing goes for the "Jira Database Setup" done via the browser.

Does anyone have any better ideas / implementation? I can't imagine I'm the first person to try this.

Edit: I should mention that these don't support video, so the option mentioned by Atlassian about using their Setup Wizard and then using the created response.varfile isn't viable (since it doesn't create one).

[Nic Brough -Adaptavist-]

You really don't want to bother running through the whole setup again.

Back up

• The database
• The Jira-home directory
• The attachments directory

If you then restore those three on to the same systems, you won't have to change anything. If you use them to create new systems, then you will have to change the database connection settings, the base url and look at the licences.

If you install on a new server set, then you need to run the install, change the database settings, the base url, customisations, additional drivers and libraries, attachment settings, reindexing, ... and so-on. This is what backups are for - to get you running on the same, or similar, servers.

[Harry Chan]

Hi, I don't see why you would need Atlassian's help on this especially on AWS. Just setup AWS with all your data on the EBS. Snapshot it. Restore the snapshot. Store the snapshots on S3.

Is there something I'm missing here?

Otherwise you can always create an Image and use that to create a "template" when you deploy a new instance. Then restore the database.

[Steffen Opel]

@Harry - minor correction to avoid confusion: An EBS snapshot is by definition stored in S3, you cannot store it there yourself; in fact you cannot access it via S3 yourself at all, but only through the related EBS snapshot APIs (or tools using these), see my related answer to How do I save an AMI to a S3 bucket? for more on this.

[Renjith Pillai]

And an additional point. You can use a .varfile created during the installation process from another machine. Customize the paths as per your target instance data.

https://confluence.atlassian.com/display/JIRA/Installing+JIRA+on+Linux#InstallingJIRAonLinux-PerforminganUnattendedInstallation