ROS Resources: Documentation | Support | Discussion Forum | Service Status | Q&A

Help with Initial Setup (404 reprepro issue)

I did a straight clone of the default configs (rewriting the server addresses with my instances) and managed to get jenkins functioning for the most part, but I’m getting a 404 issue on reprepro that I also saw here ( I checked that it is not authentication related. I clearly see root and jenkins-slave logging into the slave in the auth.log. Also, syslog and puppet.log look clean. The creation of docker images is failing due to the 404 issue since the reprepro repo directory is empty. Do you have any advice on what I should check out next. Do I need to do an initial prerelease to populate the repository?

Did you run the import_upstream job? It’s usually the first thing required to bootstrap the repo:

Almost any operation on the repo will bootstrap it. But the import_upstream has no dependencies so we just recommend running that instead of creating a single use job that would setup the repositories.

Not sure whether this is what caused it, but for my repo host I had to remove the ‘second part’ of the signing_key value in buildfarm_deployment_config/repo/common.yaml (here and in the other places where the key is referenced) or reprepro would complain about not being able to find the referenced key.

This stopped being a problem when I updated my config to use my own key, so I didn’t investigate this any further.

If you still have it around, you could check your puppet.log (in /var/log) and search for any reprepro lines. For me it complained that it couldn’t find the key, even though it was successfully imported in an earlier step.

@tfoote @gavanderhoorn thanks for the suggestions, I’ve been trying to get as far as I can with building locally. Once I get that working I will test these out. I’ve added https private git repo support to ros_buildfarm, ros_distro, and bloom