dMp

At least the greenlet is no longer going in lalaland. So: - do you experience degradation and ultimately unresponsiveness after these errors? - are you running nginx in front of...

@mcadam that's likely this: https://github.com/benoitc/gunicorn/issues/818

@garo and others 1. as a workaround, I would suggest you delegate the layer delivery directly to S3 (see storage_redirect under https://github.com/docker/docker-registry/#general-options - doc sucks but it should be easy...

@all we recently removed ParallelKey fetch from S3 from the registry - it was triggering timeouts for large objects, and cluttering disk space with orphaned temporary files. It's hard to...

It's code removal (#961). You can test by using the latest master.

I would definitely recommend using the Redis LRU cache. I would also advise to delegate actual bits delivery to either S3 or a Cloudfront view of it (https://github.com/docker/docker-registry/blob/master/ADVANCED.md#s3-storage-with-cloudfront)

@dsw88 I now think you are not talking about the same issue that was initially described here (the recommended workarounds were/are here to address registry side EOF/issues communicating with S3)....

Your search index is somehow corrupted. I would recommend deleting the index.

This here (the open-source python docker registry) will not stop "working". Though, it is not actively maintained. The docker engine will still interact with it for the time being, though...

Can you report this to https://github.com/docker/docker ? Bad file descriptor points to an engine bug, or some other local filesystem corruption.