fix: default to 50% max unavailable for longhorn manager daemonset

[NilsGriebner]

When using 100% max unavailable for the longhorn-manager daemonset a helm upgrade fails with the following error.

 Upgrade failed: an error occurred while rolling back the release. original
│ upgrade error: failed to create resource: Internal error occurred: failed
│ calling webhook "mutator.longhorn.io": failed to call webhook: Post
│ "https://longhorn-admission-webhook.longhorn-system.svc:9502/v1/webhook/mutation?timeout=10s":
│ no endpoints available for service "longhorn-admission-webhook": failed to
│ create resource: Internal error occurred: failed calling webhook
│ "mutator.longhorn.io": failed to call webhook: Post
│ "https://longhorn-admission-webhook.longhorn-system.svc:9502/v1/webhook/mutation?timeout=10s":
│ no endpoints available for service "longhorn-admission-webhook"

This can be fixed by setting max unavailable lower than 100%.

[derekbit]

@NilsGriebner It is intentional.Once all old longhorn-manager pods are terminated, a new longhorn-manager pod will start and handle the upgrade and provide webhook services. failed to call webhook: Post... is a temporary service unavailability.

[NilsGriebner]

@NilsGriebner It is intentional.Once all old longhorn-manager pods are terminated, a new longhorn-manager pod will start and handle the upgrade and provide webhook services. failed to call webhook: Post... is a temporary service unavailability.

Hey @derekbit,

thanks for your quick reply. It might be that this is a temperory service interruption, but this leads to a failed helm upgrade (we tried 5 times in a row with the same result). Maybe we're running in an edge case here?

[derekbit]

Hello @NilsGriebner Does the issue happen starting from v1.11.0?

[NilsGriebner]

Hello @NilsGriebner Does the issue happen starting from v1.11.0?

Yes, this started with v1.11.0.

[derekbit]

Hello @NilsGriebner Does the issue happen starting from v1.11.0?

Yes, this started with v1.11.0.

I see. Please check longhorn/longhorn#12578.
We found the root cause and will provide a hotfixed image soon.

[NilsGriebner]

Hello @NilsGriebner Does the issue happen starting from v1.11.0?

Yes, this started with v1.11.0.

I see. Please check longhorn/longhorn#12578. We found the root cause and will provide a hotfixed image soon.

Thanks. We'll try this out and report back.