debugging a volume ¶
the kubectl debug command is a nice way to mess around with an
existing pod by adding ephemeral containers to it. annoyingly, it does
not seem to provide a convenient way to access volumes, instead
requiring a custom profile json file
echo '{"volumeMounts":[{"mountPath":"/data","name":"datavol"}]}' > profile.json
kubectl debug -it --image busybox --custom profile.json my-cool-pod
longhorn in helm controller ¶
if you've used longhorn for a bit, you'll know that it occasionally
shits itself. but if you use the recommended setup for helm controller
(setting failurePolicy: abort), it'll still error out even after the
original issue is resolved.
Release status is 'failed' and failure policy is 'abort', not 'reinstall'; waiting for operator intervention
so how does one intervene?
by digging around in longhorn's secrets with kubectl get -n longhorn-system secrets
NAME TYPE DATA AGE longhorn-webhook-ca kubernetes.io/tls 2 42d longhorn-webhook-tls kubernetes.io/tls 2 42d sh.helm.release.v1.longhorn-install.v1 helm.sh/release.v1 1 42d sh.helm.release.v1.longhorn-install.v2 helm.sh/release.v1 1 42d sh.helm.release.v1.longhorn-install.v3 helm.sh/release.v1 1 8d sh.helm.release.v1.longhorn-install.v4 helm.sh/release.v1 1 8d sh.helm.release.v1.longhorn-install.v5 helm.sh/release.v1 1 77m
and then just deleting the most recent helm.sh one!?
kubectl delete -n longhorn-system secrets sh.helm.release.v1.longhorn-install.v5
and only then will it re-attempt running the helm chart. as far as i can tell, longhorn just does not document this anywhere, i only know about this because of one comment on the github issue about adding helm controller support...
