Since 2 days ago, we have been getting "unknown blob" errors when pulling from jfrog. I am attaching a sample log:
Command ['ssh', '-o', 'StrictHostKeyChecking=no', '-o', 'LogLevel=ERROR', 'localhost', 'docker', 'pull', '<redacted>.jfrog.io/<redacted>:latest'] failed with exit code 1 and output 'latest: Pulling from <redacted>
f5d23c7fed46: Pulling fs layer
3f4aa1d1dde5: Pulling fs layer
52c4bf0b6229: Pulling fs layer
fe61f8f5a308: Pulling fs layer
ebeed9e8b27e: Pulling fs layer
89831686aa31: Pulling fs layer
2e2c5baec652: Pulling fs layer
b6fa760c79e4: Pulling fs layer
2e2c5baec652: Waiting
ebeed9e8b27e: Waiting
b6fa760c79e4: Waiting
fe61f8f5a308: Waiting
3f4aa1d1dde5: Verifying Checksum
3f4aa1d1dde5: Download complete
f5d23c7fed46: Verifying Checksum
f5d23c7fed46: Download complete
fe61f8f5a308: Download complete
ebeed9e8b27e: Download complete
89831686aa31: Download complete
f5d23c7fed46: Pull complete
3f4aa1d1dde5: Pull complete
2e2c5baec652: Verifying Checksum
2e2c5baec652: Download complete
b6fa760c79e4: Downloading
unknown blob
This seems to have started during the kinesis outage. We first noticed it while we were trying to deploy a workaround during the outage. However the problem still persists.
The image pulls fine from docker hub, so it's not corrupted. This is currently breaking out automated deploy/provisioning process, as we have manually pull failed imaged from dockerhub.
Thanks, -Caius