You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
fix(networking): cleanup service/endpoint if needed
We meet a corner case that the service/endpoint would not be
cleanup. That will cause the service keep the ClusterIP `None`.
With this config, the endpoint of sharemanager would not correct.
So the CSI driver cannot perform the mountpoint well.
We would like to have a checking mechanism to know if the service/
endpoint did not cleanup. Then we will cleanup the service/endpoint
to ensure the correct endpoint.
Remove the cleanup function in the setting controller, we could do
the cleanup on the sm controller
Signed-off-by: Vicente Cheng <[email protected]>
return&types.ErrorInvalidState{Reason: fmt.Sprintf("failed to apply %v setting to Longhorn components when there are attached volumes. It will be eventually applied", types.SettingNameStorageNetworkForRWXVolumeEnabled)}
360
360
}
361
361
362
-
// Perform cleanup of the share manager Service
363
-
// This is to allow the creation of the correct Service
364
-
// and Endpoint when switching between cluster network
err=errors.Wrapf(err, "failed to cleanup share manager service and endpoints for %s setting update", types.SettingNameStorageNetworkForRWXVolumeEnabled)
0 commit comments