On Friday, February 17, 2012 at 1:03 AM, Torben Knerr wrote:
On Thu, Feb 16, 2012 at 9:58 PM, Bryan McLellan <email@example.com (mailto:firstname.lastname@example.org)> wrote:
On Thu, Feb 16, 2012 at 5:01 AM, Torben Knerr <email@example.com (mailto:firstname.lastname@example.org)> wrote:
has anyone tried to set up the whole Atlassian Suite via Chef? I know there
is a JIRA cookbook, but I would need to set up Bamboo, Fisheye, Crowd, etc.
as well. And git or SVN as well. And everything nicely configured.
I wrote these internally at Opscode about a year ago, but they only go
as far as you would expect. You’ve still got to load up a web browser
for all the configuration. Going any further would be a feat of
engineering and unfortunately probably not stable between upstream
Yeah, that’s what I expected. I could imagine to mechanize these steps, but not sure if it would be worth the effort given that the webui changes quite often between atlassian versions…
Btw: so these are the same cookbooks that Bryan Brandau refers to?
Having a mix of Chef and backups isn’t necessarily a bad thing when you run into tricky applications like this. I recently set up Jenkins so that Chef would install the software and dependencies, install plugins, and install the libraries and such we needed to run the builds. For the actual Jenkins config, I didn’t have the time to reverse engineer the job XML files and write an LWRP to generate them, so I decided to just back up the data to S3. It worked pretty well when I tested rebuild/restore. It does require one extra step beyond just running Chef, but I restore the build history when I do this, so I do get some value out of it.
I wouldn’t be happy with this if this was for my revenue generating infrastructure, but for one-offs I’m primarily automating to make DR work smoothly so it’s an acceptable trade off.