fix: get the correct corrosion version in ucind image (#285)

On arm64 the x86 version of corrosion was downloaded. This downloads
the correct version. Also added comments to HACKING.md on how to clean
up a broken cluster as I have done it twice now and forgot how to do it
both times.

Fixes: #284

~~~ shell
% ./ucind cluster create miektest
Cluster "miektest" initialised with machine "machine-1"
Waiting for cluster to be ready... done.
Cluster 'miektest' added to Uncloud config as the current context.
Cluster 'miektest' created.
~~~

`uc machine ls` also works.

Signed-off-by: Miek Gieben <miek@miek.nl>
This commit is contained in:
Miek Gieben
2026-03-25 13:24:40 +10:00
committed by GitHub
parent a0dbe4a983
commit ba73259887
2 changed files with 13 additions and 2 deletions
+3 -2
View File
@@ -39,10 +39,11 @@ CMD ["corrosion", "agent"]
FROM alpine:${ALPINE_VERSION} AS corrosion-image-tarball
ARG CORROSION_IMAGE="ghcr.io/psviderski/corrosion:latest"
ARG TARGETOS
ARG TARGETARCH
RUN apk --no-cache add crane
RUN crane pull "${CORROSION_IMAGE}" /corrosion.tar
RUN crane pull --platform ${TARGETOS}/${TARGETARCH} "${CORROSION_IMAGE}" /corrosion.tar
# Uncloud-in-Docker (ucind) image for running Uncloud test clusters using Docker.
FROM docker:27.3.1-dind AS ucind
+10
View File
@@ -47,6 +47,16 @@ Windows:
GOOS=linux GOARCH=amd64 go build -o uncloudd ./cmd/uncloudd
```
## Docker in Docker (ucind)
When running ucind and running in trouble, you must manually remove the cluster, by running:
```shell
docker stop CONTAINERID
docker rm CONTAINERID
docker network rm CLUSTERNAME
```
## Testing
Run all tests (unit and e2e):