Disable the Weave agent to prevent it from overwriting the adjustments you make in the next steps by running:
kubectl -n weave scale deployment/weave-agent --replicas=0
-
Retrieve the Weave cloud manifests the agent would normally apply by requesting the following URL, replace
<k8s-version>
and<service-token>
with the respective values:curl -o weave-cloud.json -L "https://cloud.weave.works/k8s/v<k8s-version>/weave-cloud.json?service-token=<service-token>"
-
Change the affected images
quay.io/weaveworks/watch
andquay.io/weaveworks/flux
, so they are pulled fromdocker.io
. This can be done by running the following command:sed -i -e 's#quay.io/weaveworks/watch#docker.io/weaveworks/watch#g' -e 's#quay.io/weaveworks/flux#docker.io/weaveworks/flux#g' weave-cloud.json
-
Apply the modified manifests to your cluster:
kubectl apply -f weave-cloud.json
Once the issues with the images have been resolved you can simply
directly apply the manifests from cloud.weave.works
to your cluster
and scale up the agent again:
kubectl apply -f "https://cloud.weave.works/k8s/v<k8s-version>/weave-cloud.json?service-token=<service-token>"
kubectl -n weave scale deployment/weave-agent --replicas=1