dagu: run dagu-tools on the stock docker CLI image #293
No reviewers
Labels
No labels
audit-work
bug
docs
general-admin
major-upgrade
needs-vps-sync
new-service
on-hold
outside-work
post-podman
renovate
upstream
vps
No milestone
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
mysticalsoap/docker!293
Loading…
Reference in a new issue
No description provided.
Delete branch "dagu-tools-stock"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Problem
dagu-tools was a locally built image (alpine plus docker-cli, curl, unzip, bash) when the stock
docker:<version>-cliimage already ships the docker CLI and a busybox with wget and unzip. It was the one scheduled-job sidecar not on a stock upstream image, and after #290 and #291 the onlybuild:left besides config-renderer and the two apps tracked in #292.Fix
dagu-toolsbecomesdocker.io/library/docker:28.5.2-clipinned by digest, the same image the validate workflow runs check-compose.sh in, with its docker-wrapping entrypoint overridden by the idlesleep.dagu/tools/Dockerfileremoved.ip2location/update.sh,crowdsec-hub-update/update.sh,image-scan/list.sh) rewritten for busybox sh: wget instead of curl, no bashisms. DAGs invoke them withsh.scan.shwas already sh.Verification
On the branch,
just deploy dagurecreated dagu-tools from the stock image (entrypointsleep infinity, user 948:948). Then every DAG that uses it, run by hand:image-scan: list step in sh, 58/58 images ok, registry fallback still kicks in for invidious-companion.docker-image-prune: succeeded.crowdsec-hub-update: hub update and upgrade ran, crowdsec restarted and came back healthy.ip2location: DB downloaded with wget through the real token, 11.7 MB written in place, traefik restarted and came back healthy.Also the first real-merge run of #290's auto-detect happened just before this branch: after pulling #290 and #291,
just deploypicked infra, monitoring, invidious and web from the two Dockerfile pins alone, rebuilt config-renderer and knock, and every container came back healthy.