Hi @RA2DREAMChallengeParticipants, As we've been working to finalize the challenge infrastructure, we've put together a sample Docker image and [Github repository](https://github.com/allaway/ra2-docker-demo). This Github repository recapitulates a lot of the information on the Docker Submission wiki page, but also has a companion Dockerfile, mock model.R file, and runscript (run.sh) to help contextualize the submission requirements and to provide a jumping-off point for those less familiar with Docker. Please let me know if you have any questions! Have a great weekend, Robert

Created by Robert Allaway allawayr
Great. Thanks.
Hi @allawayr, It finally worked, it was apparently a transient connectivity problem.
Weird - this is not something I have ever seen. I don't run my push command from a shell script though. My suggestions would be: -Run the docker push command directly without a shell script. -Delete the repository on Synapse (docker tools -> delete) and then try pushing again. Here are some other ideas: https://forums.docker.com/t/docker-pull-results-in-request-canceled-while-waiting-for-connection-client-timeout-exceeded-while-awaiting-headers/73064/9
I understand the queue is not active yet. The problem is with the push command, see output: ariel@Ivory:~/datasets/ra_prod$ sh push_command.sh The push refers to repository [docker.synapse.org/syn21499370/ra_model] b252e3b24cc4: Pushed f5fd72cb9a51: Retrying in 1 second c8a8a7d160e7: Pushing [==================================================>] 980.8kB/980.8kB 52159897c4d5: Pushing [==================================================>] 11.78kB 129f9449b6b4: Pushed 79bfd58631af: Layer already exists 79eeda22554c: Layer already exists 0c5b5c1cb7a3: Layer already exists 47ac0adb54a2: Layer already exists 50758505a30d: Pushed 78abf2f38575: Layer already exists b96c8d630793: Pushed 2399eeb9620d: Layer already exists b965ff14ad70: Layer already exists 7a5145032a85: Layer already exists 93cf512b0325: Retrying in 1 second 596513edafa6: Layer already exists 6bb0972b46b5: Layer already exists 95698a906106: Layer already exists 965f090b9482: Layer already exists 918efb8f161b: Layer already exists 27dd43ea46a8: Layer already exists 9f3bfcc4a1a8: Layer already exists 2dc9f76fb25b: Layer already exists dial tcp 3.222.52.89:443: i/o timeout ariel@Ivory:~/datasets/ra_prod$ sh push_command.sh The push refers to repository [docker.synapse.org/syn21499370/ra_model] Get https://docker.synapse.org/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)
@arielis - you may have gotten some notifications about your previous (invalid) submission running through the queue - please disregard. The queue is still not live. Thanks!
Hmm. Can you email me a copy (allawayr@synapse.org) of your terminal session? I am not having trouble pushing images so I don't think the issue is on Synapse's end.
dial tcp: lookup repo-prod.prod.sagebase.org: no such host
Can you copy and paste the error? I've not encountered that issue myself. Thanks, Robert
Hi @allawayr , Thanks for your answer. I've rebuilt the docker file, and I'm trying to push it, but I keep getting disconnected during the push command. It seems there are some connectivity problems to the synapse docker server... .
Hi @arielis - We updated some software for running the scoring harness that required this change. ("Would the testing infrastructure know to run the "/run.sh" even if it does not appear as an entrypoint?" no - unfortunately not). Would you mind rebuilding and resubmitting your Docker container to conform to the updated requirements when you have a chance? Apologies for the change! The image you submitted already will not use up one of your quota of attempts - our scoring code will only "count" submissions that pass validation against your limit (i.e. those that produce a valid prediction file in our hands). Cheers, Robert
@lars.ericson Thanks for your concern... I don't mind using one of my 5 attempts, as long as the solution gets scored.
@arielis, if I was hearing them right, by submitting a dockerized solution to test the framework, you just used up 1 of 5 chances. @allawayr please confirm.
@allawayr , Great. I have already pushed and submitted a dockerized solution a few days ago, according to the instructions I found on the wiki. It does not use a GPU, but, I did not put the entrypoint line at, that I see now in the github at the end. "ENTRYPOINT ["/bin/bash", "/run.sh"]" Would the testing infrastructure know to run the "/run.sh" even if it does not appear as an entrypoint? If it works anyway, I'd be happy to keep it on the queue, for when the scoring engine would work.
We are not yet accepting submissions but hopefully will be soon. We are still working out some compute provisioning issues and testing the infrastructure to make sure it is stable but none of these should change the submission process. I will provide an update as soon as possible. Thanks, Robert
Does this mean scoring engine is running?
For those needing GPU access, please note that the run.sh file contains some critical PATHs that must be exported in your container to map NVIDIA drivers correctly.

hello-world example Docker container page is loading…