Problems building singularity image: Authentication error. about chipseq

[ewels]

Have you tried nfcore-chipseq? (All lower caps). I think this should work. Annoyingly, docker assumes that if it can’t find a repo then it must be private and that it’s an authentication problem. So bit of a misleading error.

Images aren’t pushed to shub currently. Need to fix the docs here.

Phil

[orzechoj]

Well I got a different error message for nfcore-chipseq:

$ singularity pull --name nfcore-chipseq.img docker://nfcore-chipseq
WARNING: pull for Docker Hub is not guaranteed to produce the
WARNING: same image on repeated pull. Use Singularity Registry
WARNING: (shub://) to pull exactly equivalent images.
Docker image path: index.docker.io/library/nfcore-chipseq:latest
ERROR UNAUTHORIZED: authentication required
ERROR Check existence, naming, and permissions
Cleaning up...
ERROR: pulling container failed!

[ewels]

Sorry, should't reply on my phone! It's a slash, not a hyphen. No hyphens at all, all lower case:

$ singularity pull --name nfcore-chipseq.img docker://nfcore/chipseq

You can tell from the docker pull command in the box on the right on the image page which has the image path.

Phil

[orzechoj]

Hmm, I'm still having the same problem 🤔

$ singularity pull --name nfcore-chipseq.img docker://nf-core/chipseq
WARNING: pull for Docker Hub is not guaranteed to produce the
WARNING: same image on repeated pull. Use Singularity Registry
WARNING: (shub://) to pull exactly equivalent images.
Docker image path: index.docker.io/nf-core/chipseq:latest
ERROR UNAUTHORIZED: authentication required
ERROR Check existence

[ewels]

You put a hyphen in again! It's nfcore/chipseq not nf-core/chipseq 😆 However, I confess that I did not test the command that I wrote above when I wrote it - I just tried it now and I get a similar error.

On closer inspection the reason for this becomes (embarassingly) clear. There are no image builds. So there's nothing to pull.

The last one failed during the R installation step. I remember now that I did see this I think, but we're currently rebuilding the entire container, so I figured that it wasn't worth looking into it any more.

Ok, so in summary:

1. Sorry for making you test so many commands! 😆
2. We've started making a nf-core python helper tool, and I will soon add a subcommand that will automatically handle all of this crap by looking at the pipeline code. So you will never have to go through this again.
3. We will finish work on the new container soon I hope, at which point the nfcore/chipseq container will exist!
4. For now, running with the scilifelab/ngi-chipseq container should be fine. All of the required software is there, though it doesn't reach the level of reproducibility that we're aiming for (but this repo doesn't have a stable release yet, so we'll get there).

Again, apologies for the half-responses over the past few days (I've been travelling around teaching on various courses). Thanks for testing! I'll close this now that we have an explanation I think.

Phil

[orzechoj]

Haha, ok good to have it sorted out. My test with scilifelab/ngi-chipseq container is still running, so it seems to work ok.