Why does Bamboo wait to start a build when commanded manually?


Using the latest bamboo on Windows Server 2012 R2 - all working great.  5 agents setup, powerful server with multi-procs and 128 GB RAM.  When I click the play button to command a build, I see it show three dots ... and I have to wait somethings several minutes for the build to start.  Nothing else is in progress.  Why is there a delay and how can I have this fire immediately or more efficiently?

Thank you.

1 answer

0 votes

First thing that comes to mind - maybe you have big source repository (or many of them in one plan).

When starting manual build it gets to queue but Bamboo server is updating every repository to find out set of commits to build. If there is big repository/many changes/many repositories it can take a lot of time...

If this is your case there is not much that can be done. If you can switch to Linux on server side it could help. There are known performance issues with windows and for example git (I don't know what repository type you are using). (as a side node Bamboo best works with PostgreSql database - but it's irrelevant for this case)

If you must use windows you can try log in and manually clone repository to see how long does it take. Then experiment with different git implementations (assuming you are using git). This one is new and promising: https://git-for-windows.github.io/

If you have one big repo (let's say > 1GB) you could also consider splitting it.

Thank you for the reply @Pawel Skierczynski I am using Windows Server 2012 R2 against SQL Server 2014 STD with Git. No large repo's, about 20 repo's in total, all set to use shallow clones, etc.

I do have quiet periods on the linked repositories, but there was no activity in the past 5 minutes on them which is their quiet period. Even if no activity would a commanded build action still way a quiet period amount of time? That may be the culprit.

Quiet period is not connected to manual build in any way. Nevertheless I would strongly discourage using quiet period feature. It will have performance problems with anything bigger than very small instance. Shallow clones will not help with change detection much. SQL Server requires specific configuration. I hope you've seen this https://confluence.atlassian.com/bamboo/microsoft-sql-server-289276827.html. Try running git clone manually on server to see how long doest it take. Also check what git implementation does server use? If using native git (recommended) setting logs to debug on com.atlassian.bamboo.plugins.git could tell us more (like if git fetch command is called early and just very slow). Logs can be tuned runtime from system admin menu.

Suggest an answer

Log in or Sign up to answer
Community showcase
Asked Thursday in Jira Ops

I'm John Allspaw, Ask Me Anything about incident analysis and postmortems

I'm John Allspaw, co-founder of   Adaptive Capacity Labs, where we help teams use their incidents to learn and improve. We bring research-driven methods and approaches to drive effective inciden...

4,140 views 13 10
View question

Atlassian User Groups

Connect with like-minded Atlassian users at free events near you!

Find a group

Connect with like-minded Atlassian users at free events near you!

Find my local user group

Unfortunately there are no AUG chapters near you at the moment.

Start an AUG

You're one step closer to meeting fellow Atlassian users at your local meet up. Learn more about AUGs

Groups near you