Issue with NFS Cluster and Proxmox

K

kitsune242

Member
Oct 27, 2022
5
4
8
Hi,

I’m facing a rather strange issue with my NFS cluster. I have HAProxy set up in front of it, and when I manually mount the NFS share on a Proxmox node using:

mount -t nfs -o vers3,nolock,proto=tcp server destination
everything works fine. However, when I run the command:


pvesm nfsscan server
I don't see any available resources.

If I set up the mount specifically for a single node, everything works correctly, but this defeats the purpose of having an NFS cluster, causing inconsistencies.

I suspect the issue might be related to rpcbind, as HAProxy only handles TCP connections.

Do you have any insights on how to resolve this?
 
Seems your HA Proxy don't play nice in the game. Normally you would define a "flying" nfs-cluster IP (or name) in datacenter as storage (which is even normally initiated as nfs v4.2). Assuming you have a minimum of two pve servers too, so temporary export a local dir on 2nd pve and see if "pvesm nfsscan 1st-pve" is still working fine.
 
The issue is likely due to the fact that "scan" relies on response from rpcinfo command, which uses UDP and TCP. Since you are not proxying UDP, the scan does not work.
Keep in mind that periodic checks also use the same method, so you'd want to resolve this if you were to use haproxy/nfs in production with PVE.

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
 
  • Like
Reactions: waltar
You must log in or register to reply here.
Top Bottom
Back