This message was deleted.
# longhorn-storagea
adamant-kite-43734
02/14/2024, 9:16 AMThis message was deleted.
c
creamy-pencil-82913
02/14/2024, 5:07 PMno, NFS is what makes it RWX (ReadWriteMany). The underlying volume is still xfs or ext4, but it’s shared to multiple nodes via NFS. if you disabled NFS all you’d have was RWO (ReadWriteOnce).
creamy-pencil-82913
02/14/2024, 5:07 PMIf you don’t want NFS, make your volumes RWO instead.
creamy-pencil-82913
02/14/2024, 5:09 PMIf your LH volumes are already all RWO then you’re not using LH NFS and it would not be related to the issue you’re experiencing.
c
crooked-cat-21365
02/15/2024, 9:07 AMProbably a misunderstanding: I have RWX based on NFS using the nfs-provisioner and an external NFS server. Problem is, nobody is using it, since Longhorn's RWX is the default. Can't be bad, if it was set as a default, is it?
Sorry to say, but NFS within the cluster is bad. Maybe I missed something, but apparently all I can do is removing the default bit for Longhorn entirely, because I cannot turn of the NFS part bundled with it.
c
creamy-pencil-82913
02/15/2024, 6:01 PMthere is no default for rwx vs default for rwo. There is just a default storage class. You should chose which makes the most sense to have as the default, or perhaps you don’t want to have a default at all so that people have to intentionally pick the correct one.