NFS HA

eagredanom

I am running the Exercise A.3. Practicing Skills for NFS and I am re using the HA environment from course.

I have not had success on it and I am just thinking this is because of HA and I need a different approach than the one reviewed on trianing

Is my assumption correct?

Any hints on this?

TY

chrispokorni

Hi @eagredanom,

The NFS example assumes the NFS server running on CP1, while the WORKER is the only client mounting the NFS. After adding HA-proxy, CP2 and CP3, how is your PV defined and is CP1 still operational in the cluster?

Regards,
-Chris

fazlur.khan

I think if the pods gets scheduled on cp2 or cp3, they might fail to work, and If you install NFS clients on CP2 and CP3, It should work fine.

eagredanom

My PV has defined the server just for CP1 , I have installed the NFS server only at CP1, not on CP2 or CP3, and when issuing "showmount -e k8scp" at the worker node, I got the error "clnt_create: RPC: Unable to receive"

I was wondering that because of the HA environment ha-proxy and CP2 CP3, i should deploy NFS server under an HA scenario

chrispokorni

Hi @eagredanom,

If closely following the lab guide, the k8scp alias is re-assigned to the HA proxy server. After the alias re-assignment, the showmount -e k8scp command targets the HA proxy server. Instead, it should target the system running the NFS server, that is CP1. Update the commands referencing the k8scp alias with the actual CP1 hostname or its private IP.

Regards,
-Chris