Docker Pull (Client.Timeout exceeded while awaiting headers)


We are running Octopus Server 2020.3.3 running Linux tentacles in Pull mode on CentOS 7. I’m trying to deploy a Docker image using the standard step template and receiving the error below. I have confirmed that the feed credentials are correct as I can interactively log into the docker registry from the host and perform a Docker Pull of the same image. The problem only seems to be apparent when trying the same thing via the calamari package download function. Please can you help?

Error response from daemon: Get https://docker-deploy-***.***.****.local/v2/: Get https://docker-deploy-****.**.****.local:443/api/docker/docker-deploy/v2/token?account=s****&client_id=docker&offline_token=true&service=docker-deploy-sec01.dl.npres.local%3A443: net/http: request canceled (Client.Timeout exceeded while awaiting headers) (Client.Timeout exceeded while awaiting headers)

August 26th 2020 09:53:43


Login Failed

August 26th 2020 09:53:43


Failed to download package v0.8.9 from feed: 'https://docker-deploy-sec01.dl.npres.local'

August 26th 2020 09:53:43


Unable to pull Docker image

Hi James,

Welcome to the Octopus forums!

Thanks for reaching out.

When you set up the external feed and you click the test button, does it work there?

Are you using http auth or a token for the docker feed?

Does anything show up in the docker feed when the bad logins are happening?

Please let me know if you have any questions about the above.


Hi Jeremy,

Thanks for your reply. The test button does find a package in the new feed, it is slow but it does return the searched for package.

I am just using http auth for the login.

I have been testing further throughout today and, the problem seems to be intermittent. Sometimes it throws the error at login, other times it logs in but then times out during the download and then other times the docker pull completes.

I am wondering if there is any configuration setting I might use to try to increase the timeout period?

Hey James,

Thanks for the information. It feels like this is probably environmental since it’s intermittent.

Unfortunately, there is no configurable timeout for docker pull.

Would you be able to monitor the network traffic between the source and destination to see if it may be getting blocked/terminated in the instances where it doesn’t complete successfully? Is there anything in your antivirus/firewall logs that indicate it may be causing an issue sporadically?

Please let me know how it goes.


This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.