17 days ago
My frontend service builds fail immediately on Metal builder production-builderv3-us-west1-g754 with: ResourceExhausted: failed to read dockerfile: failed to create lease: write /var/lib/buildkit/runc-overlayfs/containerdmeta.db: no space left on device. Retried over several hours, same result. Fails before build steps run. Aligns with your May 20 Metal build queue incident. Please confirm if my builds are hitting full/unhealthy builders and help unblock.
2 Replies
Status changed to Awaiting Railway Response Railway • 17 days ago
17 days ago
The "no space left on device" error here is fallout from yesterday's incident, where GCP brought our builders back up in an incorrect state. That's been fixed on our end, so canceling and retrying the failed builds should get you unblocked. Sorry for the disruption this caused.
Status changed to Awaiting User Response Railway • 17 days ago
17 days ago
Hey there,
We had a builder disk pressure issue across our V3 fleet (us-west1 region) following the post-outage recovery on May 20. The build queue is fully healthy again now; end-to-end deploys across the platform are completing in under a minute.
I just kicked a fresh redeploy on your service from my side. It should land on a clean builder.
If you're still hitting the same "no space left on device" error after this, reply with the latest deployment ID and I'll route it for direct attention.
Sorry for the disruption,
Angelo
Status changed to Solved Railway • 10 days ago