This message was deleted.
# longhorn-storage
a
This message was deleted.
c
yeah, upstream docs specifically tell you not to use nodeName with WaitForFirstConsumer.
you can change the binding mode, or use node selector instead of nodeName.
this is a limitation of the Kubernetes PV binding process, not anything in the provisioner.
b
Awesome, thanks for pointing this out, I missed it. I was trying to recreate an already existing Pod (with the Go client) just reusing the same configuration that I received from the Get. It took me a while to find out that the nodeName was causing the issue