fix(ci): expose the dind docker host to job containers
ci / typecheck (map[dir:apps/booking name:booking]) (push) Successful in 45s
ci / typecheck (map[dir:apps/operator name:operator]) (push) Successful in 50s
ci / build (map[dir:apps/operator name:operator]) (push) Failing after 5s
ci / deploy (push) Has been skipped
ci / typecheck (map[dir:apps/portal name:portal]) (push) Successful in 27s
ci / typecheck (map[dir:apps/website name:website]) (push) Successful in 23s
ci / typecheck (map[dir:services/platform-api name:platform-api]) (push) Successful in 24s
ci / test (push) Successful in 35s
ci / build (map[dir:apps/booking name:booking]) (push) Failing after 7s
ci / build (map[dir:apps/portal name:portal]) (push) Failing after 5s
ci / build (map[dir:services/platform-api name:platform-api]) (push) Failing after 6s

gitea/runner 1.x no longer auto-mounts the docker daemon into job
containers (act_runner 0.2.x did), so 'docker build' in the build jobs
failed with 'cannot connect to /var/run/docker.sock'. container.docker_host
"" restores find-and-mount.
This commit is contained in:
Ronni Baslund
2026-06-10 08:34:54 +02:00
parent 3590c356a4
commit 1114be6c93
@@ -49,6 +49,12 @@ data:
# We don't use the Actions cache (setup-node cache was removed); disabling
# avoids the cache server the DinD job containers can't reach anyway.
enabled: false
container:
# "" = find an available docker host automatically AND expose it to job
# containers (the dind daemon's socket gets bind-mounted into jobs, which
# is what `docker build` in the build jobs needs). gitea/runner 1.x no
# longer does this by default — act_runner 0.2.x did.
docker_host: ""
---
apiVersion: apps/v1
kind: Deployment