[Issues] Helm chart problems with adding volumes to gitlab-runner (kubernetes)

Hi Guys,

I have been using helm install of gitlab-runner for some time now and works wonderful, however I now need to add in a local host path to the containers which the controller spawns up.

Helm Install

Now if you look at https://docs.gitlab.com/runner/executors/kubernetes.html#host-path-volumes its using the config file which does not exist if you do it via helm. The best they offer is a custom values file which does not seem to have away to add the volumes in.

How can I add the equivalent to helm install of gitlab runner?

I am already using a custom version of values file which can be found here:

I have looked and played with distributed cache option, but the problem with this is that it downloads the cache on every stage and its only needed for one stage (vun scanning). As its 500+MB plus wasted bandwidth on AWS S3 per stage with about 8 stages, that bandwidth/cost adds up quickly. It would be much better to be local and faster for one.

While I understand that the host path will not necessary be present (filled/data) the script will fill from source if not present but will use if it is, so this is not an issue as any further runs which land on that k8s node will then have the files.

If anyone has a work around or away to hack the helm chart, please let me know.

Kind Regards

Looks like this is a common provlem mentioned by people on here. We should make an issue on the helm chart repo in gitlab

1 Like