I've been seeing this error quite often lately. Re-running the tests sometimes fixes the issue but other times it takes multiple re-runs to get it to pass. Any ideas????
[ERROR] DOCKER> Unable to build image [gdcc/configbaker:unstable] : "The command '/bin/bash -eu -c true && mkdir -p ${SCRIPT_DIR} ${SECRETS_DIR} ${SOLR_TEMPLATE} && apt-get update -q && apt-get install -qqy --no-install-recommends ${PKGS} && apt -qqy --no-install-recommends install python3 python3-venv pipx && pipx install pipx && apt purge -qqy --autoremove pipx && ~/.local/bin/pipx install --global pipx && rm -rf \"~/.local\" && rm -rf \"/var/lib/apt/lists/\" \"/var/cache/apt/archives/\"' returned a non-zero code: 100" ["The command '/bin/bash -eu -c true && mkdir -p ${SCRIPT_DIR} ${SECRETS_DIR} ${SOLR_TEMPLATE} && apt-get update -q && apt-get install -qqy --no-install-recommends ${PKGS} && apt -qqy --no-install-recommends install python3 python3-venv pipx && pipx install pipx && apt purge -qqy --autoremove pipx && ~/.local/bin/pipx install --global pipx && rm -rf \"~/.local\" && rm -rf \"/var/lib/apt/lists/\" \"/var/cache/apt/archives/\"' returned a non-zero code: 100" ]
100 generally means that the issue is happening within the apt commands.
But for better monitoring, it might be a good idea to decouple that single command into multiple steps and enable logging/monitoring
Phil mentioned it here: #containers > Unable to build image [gdcc/configbaker:unstable]
Thanks, I was going to mention that other topic. Actually. I'll just merge them into one. It's the same topic, really.
5 messages were moved here from #containers > Unable to build image [gdcc/configbaker:unstable] by Philip Durbin ๐.
Ok. Combined.
@Steven Winship like @Ash Manda is saying, last time it was an ephemeral problem.
@Oliver Bertuch any thoughts on this?
I just added it to our next container meeting agenda, at least.
Today I merged #12613 from @Ash Manda that should help with this problem.
It should, at least temporarily, but long term we definitely want to think about how we can improve build health. Considering how many times the tests run, some builds are bound to fail nonetheless, but this looks like something preventable.
Thank you for merging it!
You have @Oliver Bertuch to thank. And his "merge it! merge it! merge it!"
This might be one of these reasons why it may be a good idea to investigate into splitting the images out of the main repo again
pros and cons, obviously :smile:
One of the things I really really really would like to see before we do this: have a proper release flow for our Maven modules. If we make certain things we need to coordinate a part of the parent, we can ship that to Central and include it from other repos.
Back to the original topic for moment, I did add it to the agenda for our next meeting, in case people want to discuss the issue or the PR.
And it would be so cool to have the parent at the root of our project... Wondering if I should just test this. If we just move files via Git, one should still be able to See the history.
Yeah, bit offtopic. Sry. It's just all so connected
So many wrongs to right. We should also be using org.dataverse instead of edu.harvard :sweat_smile:
Maybe. Or just go for io.gdcc. ![]()
At the time we didn't own either!
True true. Maybe I'll try out shuffling things around in a separate repo and see what happens.
If it brings you joy! :smile:
Last updated: Nov 02 2026 at 07:01 UTC