Development Installing Jira and Confluence on Azure (Part 1)
I’ve only found one post on anyone installing Atlassian’s products (specifically Jira and Confluence) on Microsoft Azure, so since I was going through the process, I thought I would do a little field report on it.
It’s not difficult by any means, but there are a couple of tidbits to keep note of. Also, there is more out-of-date information on installing these than you can shake a stick at, so it took me a while to sift through the old stuff.
Since I expect to have 10 or fewer users for the foreseeable future, I’ve purchased a stand-alone license for $10 per product from Atlassian (but you need to do your own hosting, versus their cloud hosted solutions).
Update: December 2014
Much of this post is old and useless. I’ve just written a new post with much easier instructions and it is MUCH faster to install (mostly because both JIRA and Confluence have improved their installation).
Set up an Azure Virtual Machine
This is a straight-forward step. Simply go to Azure’s control panel and instantiate a new VM. I decided to create a ‘medium’ Linux VM running Ubuntu 12.04 LTS… You could probably get away with a ‘small’ instance, but you might run into out-of-memory errors when running both Jira and Confluence (although scaling up and down is just a click and a reboot anyways).
That’s basically it! One thing you might want to do is to enter your VM’s control panel (from the manage.windowsazure.com location) and change your SSH endpoint to port 22. If you don’t, just keep note of the SSH port so that you can explicitly enter it when you SSH to your VM.
Getting ready to install Jira and Confluence
The most helpful wiki pages I found on installing Jira and Confluence were this one and this one. They seemed up-to-date and had a few helpful notes (which I’ll mention in this post).
First thing is to SSH into the server and download Jira and Confluence (I’ve helpfully put the command line arguments for the ones I downloaded, and they go straight into my home directory):
wget https://www.atlassian.com/software/jira/downloads/binary/atlassian-jira-6.0.1-x64.bin wget https://www.atlassian.com/software/confluence/downloads/binary/atlassian-confluence-5.1.3-x64.bin
If you try to run these installers now, they will probably bail out, because they’re not executable. Save yourself some time and just do the ol’ chmod 777.
chmod 777 atlassian-jira-6.0.1-x64.bin chmod 777 atlassian-confluence-5.1.3-x64.bin
Installing Jira and Confluence
The installation process is about 99% the same for these two products, so I will do a detailed explanation for one of them (installation order doesn’t matter).
sudo sh atlassian-jira-6.0.1-x64.bin
If there is a java error running this command, then there is a good chance you downloaded a file for the incorrect architecture, try downloading the other one (originally I tried x32, but it didn’t work, so I tried x64 and it did work).
To simplify my life, I ran the express install, but take note of installation location and port (for me, default settings were /opt/atlassian/jira/conf and port 8080 for Jira).
Because I was worried there would be a port conflict with Jira and Confluence, I jumped the gun and manually changed Jira’s port. To do this, go to /opt/atlassian/jira/conf and edit server.xml… Change the ‘connector port’ from 8080 to something else (let’s say 10000). Now, to make sure those changes worked, go to jira/bin and run ‘sh stop-jira’, and then ‘sh start-jira’ (of note, a lot of installation docs talk about needing to manually start-jira, but I never ran into this problem… It seems to automatically start on its own).
Connecting to Jira
So, from an installation point of view, that should be it, right? Well… Not exactly. It’s running, but you probably can’t tell it’s running because you can’t access it. To make Jira accessible to the outside world, you need to set up another Azure Endpoint.
Much like with the SSH port, just open up the VM’s endpoints from Windows Azure and add a new one. Give it any name, then as the private port, use the one you manually set in server.xml (10000 in my case), and as the public port, pick one that you like (maybe even 10000 again).
Finally, open up a browser and head over to Jira… [Website].cloudapp.net:10000
It should be successful, and now you can continue with the database setup for Jira.
The installation process for Confluence is identical, except the port it automatically configured to was 8090.
That’s all for right now, and in my next post, I’ll cover the database setup and SMTP options.