SSL Connection Timeout When Cloning Repo with Runner Autoscaler for Kubernetes in GKE

[Hafeez Ghanchi]

We are utilizing the Runner Autoscaler for Kubernetes in GKE to execute our pipelines. However, we encounter an issue when the runners attempt to clone the repository:

fatal: unable to access 'https://bitbucket.org/xxx/xxx/': SSL connection timeout

When I access the DinD container running on the runner-xxx pod, I can connect to bitbucket.org. However, running docker ps reveals several containers active for pipelines.

Interestingly, when I enter one of these containers from within the DinD and try to access bitbucket.org, the connection fails, even though other public domains are accessible.

Has anyone experienced similar issues or have suggestions on how to resolve the SSL connection timeout?

[Hafeez Ghanchi · Accepted solution]

After some serious sleuthing (and a few cups of coffee), I realized the culprit was the node's image type. As soon as I switched the cluster from a Container-Optimized OS image to an Ubuntu image, it was like magic—everything started working smoothly!

[Syahrul]

G'day, @Hafeez Ghanchi 

Welcome to the community.

When you mean the runner failing to clone, does this happen during the build setup, or was it during your build steps? 

Have you checked the runner logs to see if any specific error might've given you more insight? You can check our troubleshooting runner page below to locate your runner logs:

Troubleshooting runners 

Regards,
Syahrul