New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Dockercfg registry endpoint format invalid in v0.8.1 #4267

Closed
jakedt opened this Issue Feb 20, 2014 · 5 comments

Comments

Projects
None yet
5 participants
@jakedt
Contributor

jakedt commented Feb 20, 2014

When a user does a login against a naked domain for a registry, that domain gets copied into the .dockercfg directly.

{"registry.com":{"auth":"<base64encodedbasicauth>","email":"somebody@gmail.com"}}

In previous versions of Docker, this was being stored as the fully qualified endpoint, including the schema:

{"https://registry.com/v1/":{"auth":"<base64encodedbasicauth>","email":"somebody@gmail.com"}}

This would be fine if further push and pull commands did the same resolution from the naked domain that the original login command did (i.e. choose between http and https, add the API version). As a workaround, users can use the login command with the fully qualified endpoint (e.g. docker login https://registry.com/v1/)

jakedt referenced this issue Feb 20, 2014

Do not ping registry from the cli
Docker-DCO-1.1-Signed-off-by: Michael Crosby <michael@crosbymichael.com> (github: crosbymichael)
@coderlol

This comment has been minimized.

Show comment
Hide comment
@coderlol

coderlol Feb 23, 2014

Wow...thank-you!!!! I spent 3 hrs trying to figure why docker pull failed authentication while docker login succeeded...about to file an issue and found your workaround.

coderlol commented Feb 23, 2014

Wow...thank-you!!!! I spent 3 hrs trying to figure why docker pull failed authentication while docker login succeeded...about to file an issue and found your workaround.

@mhennings

This comment has been minimized.

Show comment
Hide comment
@mhennings

mhennings Feb 27, 2014

Contributor

is there any plan / idea how we are going to fix this?

Contributor

mhennings commented Feb 27, 2014

is there any plan / idea how we are going to fix this?

@kencochrane

This comment has been minimized.

Show comment
Hide comment
@kencochrane

kencochrane Feb 28, 2014

Contributor

@shin- @vieux any ideas what might have broke this?

Contributor

kencochrane commented Feb 28, 2014

@shin- @vieux any ideas what might have broke this?

@jakedt

This comment has been minimized.

Show comment
Hide comment
@jakedt

jakedt Mar 1, 2014

Contributor

My pull request #4270 fixes it. The change that stopped doing the ping before finishing docker login (commit referenced from the pull request) because it's no longer resolving a raw domain name into the http(s):///v1/ format before storing it, but that format is still being used to retrieve it later, which fails.

Contributor

jakedt commented Mar 1, 2014

My pull request #4270 fixes it. The change that stopped doing the ping before finishing docker login (commit referenced from the pull request) because it's no longer resolving a raw domain name into the http(s):///v1/ format before storing it, but that format is still being used to retrieve it later, which fails.

@unclejack

This comment has been minimized.

Show comment
Hide comment
@unclejack

unclejack Mar 8, 2014

Contributor

This was fixed by PR #4270.

Contributor

unclejack commented Mar 8, 2014

This was fixed by PR #4270.

@unclejack unclejack closed this Mar 8, 2014

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment