Concourse CI hosted in N. Virginia & pipeline building in Oregon


Morning, I have a unique problem and am interested in hearing if anyone else has stumbled into this behavior as well as if they’ve found a solution.

I have a multi-stage environment for availability and disaster recover purposes. I currently host Concourse CI (v4.2.2) in N. Virginia cluster and when the N. Virginia cluster builds it’s Docker images and uploads to their repository, thing’s seem to work pretty smoothly and pretty timely.

Now when our Oregon cluster deploys a pipeline and builds in Concourse which is hosted in N. Virginia it seems that we stumble into a long delay and at times corruption of the pipeline building. I’ve stumbled into the following messages, “Waiting for docker to come up” “dial tcp X.X.X.X:4030: connect: connection refused” “failed to create volume” that simply go away once I simply trigger a re-build and perhaps it’s network latency, but is there anyway to solve this so I don’t stumble into such things?

I appreciate all feedback.



That is indeed quite an interesting use case.

I was thinking maybe you can host a worker in Oregon that reports to the web instance in N. Virginia to handle the Oregon tasks?
I think you can target specific workers using tags or assign them to a specific team intended for Oregon tasks.