Disaster Recovery Configuration for Jira and Confluence

Zach Newman

I am planning on setting up a cold standby DR server to mirror our Confluence and Jira server. The filesystem will be fully replicated as well as the databases. However, I have a few configuration questions.

The base url will not need to be changed, correct? As long as the DNS record points to the new DR server all links should redirect. I don't wan't to have to worry about configuration files pointing back to the production site somehow.

Does the server id need to be changed? Same with the license?

It looks like application inks should be fine as well. But I noticed when using Jira as the User Server for Confluence, Fisheye, etc. that they are configured by IP. Can a url be used instead? Otherwise these will fail.

Thanks!

Nic Brough -Adaptavist-

Correct on the base url. Check your hosts file on the replicated server though.

For the server ID, cheat. Install the development license you're entitled to on it instead of the real licence and it'll work fine with a different server ID. Disaster recovery/failover is one of the things your dev licence is intended for. This also makes failing back a lot more simple too, as there's no fooling around with server IDs!

Ah, no, application links explicitly rely on IP addresses, so you'll need to change them. In fact, they do work with urls, but not very well in my experience. I tried it on a test system, and noticed slowness (to the point of timeouts), gave up and went back to raw IPs. This was with an older version of Jira, but I don't think it will have changed much (Which reminds me, I need to test this stuff with IPV6 soon...)