I want to run step “Run docker container”.
When i try to run i got the following error:
Error response from daemon: Get https://registry.harbor.com/v2/: unauthorized: authentication required
June 22nd 2022 11:50:09Info
Login Failed
June 22nd 2022 11:50:09Error
Failed to download package stg/test_package vstaging from feed: ‘https://registry.harbor.com/’
June 22nd 2022 11:50:09Error
Unable to pull Docker image
I’m using Harbor as External Feed and when i test it from the Feed configuration i’m able to see all the packages inside it.
As i see there is no credentials entries in “Run Docker Container” step and they based on what configured in External Feed on Library.
How can i solve this issue and download and run image using “Run Docker Container” step?
I noticed your feed is https:// so I am wondering if the same has happened to you, are you using a helm chart at all or does your config state http anywhere?
Sorry that didn’t work for you. Can I get a bit more information about your setup please so I can look into this further. We do have an issue at the moment where GitHub Container Registry’s (might not just be github itself) are causing errors in the UI when deploying and I am wondering if this is happening to your instance. Its not quite the same error message but it is similar. Are you getting version errors in the UI when trying to select a package like what is in the GitHub issue?
One thing to check, can you try running your script outside of Octopus please, if it gives you the same error you would need to check your harbor feed and make sure none of the accounts are locked out. We can safety say its not an Octopus Issue if you are getting the same errors when running it outside of Octopus, especially since the feed authenticates inside Octopus through the External feeds page (are you using the same account when running the ‘run docker container’ step)?
If it does allow you to pull the docker image then this is an Octopus Issue and we would need to investigate further. To do this are you able to answer some questions please:
What Octopus Version are you running (your email isn’t in our system so I cant tell if you are running Octopus Server or on our Cloud platform).
Has this only just started happening or is this the first time you are setting this project up? If this project worked before have you recently upgraded or changed any configurations to your Harbor feed or Octopus that may have impacted the step not working?
Are you able to logon to your Harbor feed with the account you are trying to use when pulling the docker image?
It seems like Harbor can only be installed on-prem. The website for it is here.
Are you able to clarify you are using the correct feed when you are setting up your Run Docker Container step and that the feed does connect and pull some packages.
It seems a little odd that your logs are pointing to an investment company that’s all so I wanted to make sure the feed you are using in the step is the correct one.
Hey Clare, I am currently having the same issue. I believe that AndreyB may have put https://registry.harbor.com as a possible replacement so their registry URL isn’t on a public forum. Octopus allows us to search the registry (When using the docker run process step) and the “Save and Test” external feed option works perfectly fine. It seems like the credentials are not getting passed over to the tentacle. I have even tried manually using “docker login” on both base and root user to have the docker daemon already logged into the registry with no luck. Any idea what could be the issue?
Thanks for the extra information. I have just passed this to our engineers to take a look at. I tried setting up my own feed and its a bit of a nightmare so it seems so I have asked around to see if anyone has one we can test with. If not ill have to try configure it myself.
The engineers will need to take a look regardless now as it seems like more of a global issue since two of you have reported the same issue occurring.
I have never heard of Harbor nor have some of our senior engineers so please excuse the ignorance on my part. Its also something we don’t usually see our customers use.
If it tests fine through our external feed its definitely something to do with how we are passing those credentials in the step as you alluded to. I will get back to both of you when our engineers respond.
Can I just ask, what Octopus Version are you running please as we need to get an idea if this is a recent bug in a new version or an older bug in an older version.
Thanks for taking a look, currently we are running Octopus Server docker version 2022.2.6996. I also swapped the registry to public (luckily our server is on-prem right now) so that no authentication was required, but I still got the authentication error. I am thinking that Octopus Tentacle is sending empty username/pass. Thanks again!
Brill thanks for that, I will get that to our engineers to, that’s a very recent Octopus version so hopefully we can replicate it and if it is a bug, get a fix out for it.
I will keep everyone in the thread updated with any news from the engineers.
I have managed to sort of replicate the issue you are having, I used Harbor’s demo feed and managed to get the logon failure when I deploy even though my feed tests fine and pulls packages.
I have created a GitHub issue for this and sent it to our engineers, if you subscribe to that you should be notified when a fix is out for it. I will get back to you if I need any further information.
I think my error is slightly different as I did not use an On-Prem feed but the logon failure is still there even though the credentials are clearly correct.
Hopefully the engineers can sort this out for you, reach out in the meantime if there is anything else you both need.
