[10:39] <cjwatson> Anyone fancy Sphinx documentation for Storm?  https://code.launchpad.net/~cjwatson/storm/doctest-cleanup/+merge/384530 https://code.launchpad.net/~cjwatson/storm/sphinx-doc/+merge/384532 https://code.launchpad.net/~cjwatson/storm/docstring-syntax/+merge/384534
[10:40] <cjwatson> I have some more stuff pending after that to add more docstrings for various things, but that'll do for now
[12:52] <tomwardill> pappacena: talking of QA, what were you trying to do that lead to this patch: https://git.launchpad.net/launchpad/commit?id=90eaad5839f6f684cedbab35f3f8030cedc294c1
[12:52] <tomwardill> something to do with editing OCI projects?
[13:11] <pappacena> It looks like. Something to do with the new oci_project_admin attribute on distribution?
[13:13] <pappacena> I don't quite rememeber... I think I couldn't edit OCIProject on dogfood, even though I was in the team that managed recipes for that distribution.
[13:13] <tomwardill> yeah, makes sense
[13:13]  * tomwardill tries
[13:13] <tomwardill> sigh, dogfood, not gofdood
[13:13]  * tomwardill can't even
[13:14]  * pappacena hahaha
[13:15] <tomwardill> okay,I can edit and rename an OCI PRoject
[13:15] <tomwardill> I think I'll call that good
[13:16] <tomwardill> and mark recipes as official
[13:16] <pappacena> 👍
[13:35] <pappacena> cjwatson, could you help me QAing dockerhub push by forcing a build for https://dogfood.paddev.net/~pappacena/ubuntu/+oci/pappacena-oci-project-1/+recipe/test-recipe ? I have just created a pushrule for that.
[13:35] <cjwatson> pappacena: Can do, can you hit "request builds"?
[13:35] <pappacena> (actually, created 2 push rules, but the first one was wrong and I do not have permission to delete it...)
[13:36] <pappacena> sure. I'll do it on the API. The interface shows a strange error saying that I should select a processor, but no processor is shown at the page.
[13:37] <pappacena> Requested the build: https://api.dogfood.paddev.net/devel/~pappacena/ubuntu/+oci/pappacena-oci-project-1/+recipe/test-recipe/+build-request/47675195
[13:43] <cjwatson> The recipe might be wrongly configured without any architectures.  Should be fixable on +edit
[13:43] <pappacena> Right! I forgot about that on edit page
[13:43] <cjwatson> 2020-05-26 13:43:09 INFO    Running <OCIRecipeRequestBuildsJob for <lp.oci.model.ocirecipe.OCIRecipe object at 0x7f47f0d87a50>> (ID 47675195) in status Waiting
[13:44] <cjwatson> Ah, and because it had no architectures, it didn't create any builds
[13:44] <cjwatson> So you'll need to fix that and try again
[13:44]  * pappacena doing it
[13:44]  * pappacena done!
[13:45] <pappacena> https://dogfood.paddev.net/~pappacena/ubuntu/+oci/pappacena-oci-project-1/+recipe/test-recipe/+build/20
[13:45] <cjwatson> Yep
[13:45] <cjwatson> Let me know when I need to run the push job
[13:45] <pappacena> Ok! I'll ping you as soon as the build finishes
[13:53] <pappacena> cjwatson, build finished
[14:00] <cjwatson> https://pastebin.canonical.com/p/yxkc4dBpMN/ hmm
[14:00] <cjwatson> MissingSchema: Invalid URL 'registry.hub.docker.com/v2/': No schema supplied. Perhaps you meant http://registry.hub.docker.com/v2/?
[14:01] <cjwatson> pappacena: https://pastebin.canonical.com/p/xB8jR8TgMm/ is the traceback
[14:01] <pappacena> Yes, it was the frist push rule that I've created. But I have created a second one fixing it
[14:01] <pappacena> It doesn't try the second push rule if the first one fails?
[14:02] <tomwardill> not atm
[14:02] <tomwardill> I think
[14:02] <pappacena> uhm... I don't have a way to fix nor delete the first rule :-(
[14:02] <pappacena> (using the API)
[14:03] <pappacena> Could you delete it for me on iharness, cjwatson?
[14:03] <cjwatson> Oh right
[14:03] <pappacena> https://api.dogfood.paddev.net/devel/~pappacena/ubuntu/+oci/pappacena-oci-project-1/+recipe/test-recipe/+push-rule/3
[14:03] <cjwatson> sec
[14:03] <pappacena> thanks!
[14:03] <cjwatson> We should also have some URL validation in there I suspect
[14:03] <cjwatson> There's a gadget somewhere for validating schemes
[14:04] <cjwatson> I SQLed it out, anyway
[14:04] <pappacena> I agree. I'll create a MP today to validate that.
[14:05] <cjwatson> Just fixing the job
[14:05] <pappacena> Should I ask a build again? Or just re-run the upload script will do?
[14:06] <cjwatson> I'm doing it
[14:06] <pappacena> Thanks!
[14:06] <cjwatson> I just set the job status back to Waiting
[14:11] <cjwatson> https://pastebin.canonical.com/p/zSPRpW35ZM/   hmm, hanging there
[14:12] <cjwatson> timed out eventually
[14:13] <cjwatson> This will probably need use_proxy=True to urlfetch, plus possibly some IS proxy configuration
[14:14] <pappacena> uhm...  it's possible. We are not using HTTPS for ROCKS?
[14:14] <cjwatson> Generally speaking these systems don't have direct access to external sites, only via squid
[14:14] <cjwatson> ROCKs is internal to our network
[14:14] <cjwatson> (We are using HTTPS; but that's not the relevant distinction, anyway)
[14:15] <pappacena> Right. Makes sense
[14:15] <pappacena> I'll open a MP to fix that...
[14:15] <cjwatson> The proxy configs are in lp:canonical-is-internal-proxy-configs
[14:15] <pappacena> Thanks
[14:15] <cjwatson> launchpad_script_servers, launchpad_stg_script_servers, launchpad_dogfood_servers will need access to dockerhub
[14:16] <cjwatson> And in fact also to whatever the ROCKs endpoints are
[14:17] <cjwatson> (While that works at the moment, it'd probably otherwise stop working once we switch to using the proxy)
[14:17] <cjwatson> So probably best to get the proxy config changes landed first
[14:17] <pappacena> :+1
[14:41] <pappacena> cjwatson, does that make sense? https://code.launchpad.net/~pappacena/canonical-is-internal-proxy-configs/lp-docker-images-pushing/+merge/384549
[15:05] <cjwatson> pappacena: Is that the API endpoint for ROCKs?  And I think we'll need upload.image-registry.staging.canonical.com too, for staging
[15:11] <tomwardill> both urls I have for ROCKS have 'staging' in them
[15:11] <tomwardill> I'm not actually sure what the endpoint for production would be
[15:13] <pappacena> I'll add a rule for upload.image-registry.staging.canonical.com. The current acl is for rocks.canonical.com. Maybe that's the production one?
[15:20] <cjwatson> Could be
[15:22] <pappacena> Ok, updated the MP. I'll open that for review. Should I ping someone from IS to review/deploy it? What is the process here?
[15:23] <cjwatson> Ask the help contact in the #is topic as a first step
[15:23] <pappacena> Great! Thansk!
[15:23] <cjwatson> They might redirect you to RT depending on how swamped they are
[15:40] <pappacena> On another topic: does anyone has any idea on how can I simulate a 500 error or similar, to see the OOPS page? In my machine it was quite easy to force a `raise Exception` somewhere, but it's not that easy on dogfood... :)
[15:49] <cjwatson> A git ref page should do I think due to the fairly broken way those are set up on dogfood
[15:49] <cjwatson> https://code.dogfood.paddev.net/~cjwatson/turnip/+git/turnip/+ref/master
[15:50] <pappacena> Cool! Thanks!
[16:19] <pappacena> cjwatson or tomwardill , I'll have lunch now, but if you find some minutes to review it: https://code.launchpad.net/~pappacena/launchpad/+git/launchpad/+merge/384557
[16:19] <pappacena> It's the MP to use proxy when uploading images to ROCKS and DockerHub.
[16:26] <cjwatson> r=me
[16:50] <cjwatson> pappacena: I misclicked and marked your OOPS links commit green on deployable, but I wasn't sure what its previous state was.  Feel free to change that if my change was wrong
[16:51] <cjwatson> pappacena: For your sync-signingkeys change, I'm just going to run it again on dogfood and mark that green if it doesn't crash, since the new test should deal with the specific case at hand.  Sound OK?
[16:54] <cjwatson> But I think I'm done for the day; will pick it up tomorrow
[17:00] <pappacena> Ok for both messages, cjwatson.