Rollout restart all meshed pods

[shaikatz]

Does someone already created some script to restart all meshed pods (for example after version upgrade)?

Thanks guys.

[grampelberg]

kubectl get deploy \
    -o custom-columns=NAME:.metadata.name,"ANNOTATIONS":".spec.template.metadata.annotations.linkerd\.io/inject"  \
    --no-headers | \
grep enabled | \
awk '{ print $1 }' | \
xargs kubectl rollout restart deploy

[shaikatz]

Thanks!

[shaikatz]

Actually, @grampelberg answer didn't cover pods that got injected by namespace annotation + has to be run for each namespace separately.

I've crafted some one-liner that deletes all pods in the cluster that has a specific container, in that example edge-20.10.5, delayed by 5 seconds between each to make it safer.
kubectl get pod --all-namespaces -o=json | jq '.items[]|select(any( .spec.containers[]; .image=="ghcr.io/linkerd/proxy:edge-20.10.5"))|"\(.metadata.namespace) \(.metadata.name)"' | awk '{ system("sleep 5"); print $1,$2 }' | xargs -L1 bash -c 'kubectl delete pod -n $0 $1'