Cleaning up build-dir on remote agents

I've noticed that my build-dir on my remote agents is getting quite large. When i started digging around in there I saw directories related to plans that haven't existed in quite some time.

I'm wondering two things.

First, is there a way in bamboo to have the remote agents clean up their build-dir after a certain amount of time. I know about the build exipry settings, but that's both wonky and is just related to build results. I want to clean up the actual directory the code is checked out to and built.

The second is, why do the build dirs for removed plans exist? Shouldn't the server send signals out to the agents to remove that stuff?

Thanks

4 answers

1 accepted

2 votes
Accepted answer

Actually, I found exactly what i was looking for.

https://confluence.atlassian.com/display/BAMBOO/Deleting+a+job%27s+current+working+files

Configure plan->jobs->misc-> make sure "Clean working directory after each build" is checked.

The problem with that setting is that if/when a build fails, you have nothing to look at to determine or troubleshoot the cause of failure.

I haven't included this one in particular but I already have a bamboo plan that runs every night to do certain clean up/utility functions, thought I'd keep all the scheduling inside Bamboo while I was at it. I already ask the APIs for quite a few stats, should be easy enough to make it delete directories for any plans/branches that don't exists.

I couldn't find any built-in solution for quite a few things already so I try to make all my work arounds able to run nightly.

That's an interesting approach. I was running this stuff through cron jobs but I like your approach better. It has a nice side effect of making sure the agent that you're deleting from isn't currently trying to build any plans at the same time.

We use 15 local agents, so I hadn't thought about that aspect. Personally we clean up after each job but I may switch to just delete that agent's working directory nightly.

Teams with larger farms and round-the-clock builds can use Monit and Agent APIs for Bamboo to monitor the disk space of agents and react as needed.

The plugin allows you to disable the agent, and monitor it until running builds finish. Monit acts as the monitoring tool that kicks off the task to cleanup.

Monit control file, and cleanup shell script example

Link to the Monit example is dead. It looks like exactly what I need - can you show us the new location?

So I just saw this: https://answers.atlassian.com/questions/199454/describe-exactly-when-bamboo-workspaces-are-cleaned-removed under the related questions section. It probably would have been nice if that came up before i submitted the question.

At anyrate, it looks like that this functionality isn't supported and I get to add some custom shell scripting glue to make my 'out of the box' solution work. That or buy bigger harddrives.

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...

453 views 2 5
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