I have a workflow with a scattered task that scatters to approximately 50 instances. It appears that in three of those instances, the images failed to pull from Docker Hub. They are all using the same image identified by a specific sha256 instead of a tag that may change, so in theory, they shouldn't need to be pulled 50 times, but it seems something might be getting rate limited somewhere.
The tasks that error out say:
Task assoc_agg.assoc_aggregate:28:1 failed. The job was stopped before the command finished. PAPI error code 14. Execution failed: generic::unavailable: pulling image: docker login: retry budget exhausted (10 attempts): running ["docker" "login" "-u" "firecloud" "--password-stdin"]: exec: already started (standard error: "Error response from daemon: Get https://registry-1.docker.io/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)\n")
Please sign in to leave a comment.