VMware Cloud Community
FredPeterson
Expert
Expert

VMkernel traffic routing for NFS - subnet mask not 24 bit

Hello,

So we've got a consultant in here setting up our NFS on our NetApp and we've of course going with an unrouted network and two VLANs for the traffic.

However, these VLANs were created as 10.1.26.0/23

In other words, when the two different NFS mounts are created, they will both be technically on the same first 3 octet subnet, however since the mask is different, something that is .5 will be on VLAN 25 and .70 will be VLAN 26.

My vSwitch 1 is configured so that

Portgroup NFS-1 is a VMkernel with an IP address of 10.1.26.7 subnet mask 255.255.255.192 and set to use vmnic4 as primary, with vmnic8 as standby

Portgroup NFS-2 is a VMkernel with an IP address of 10.1.26.72 subnet mask 255.255.255.192 and set to use vmnic8 as primary, vmnic4 as standby

My question is - is the VMkernel smart enough that if I mount NFS with an address of 10.1.26.2 it will use vmnic4 and if I mount NFS with IP 10.1.26.71 it will use vmnic8?

In other words, will the VMware vmkernel figure out the subnet mask difference or will it just go "oh, vmkernel #3 is the first one with that first 3 IP's so I'm going to use that port"?

Obviously we havn't gotten to the point of testing yet which will prove one way or the other, but I'd like folks to weigh in also.

0 Kudos
1 Reply
J1mbo
Virtuoso
Virtuoso

Classful addressing hasn't been used for years..., and in any case, if it were then it would be /8 as 10.x.x.x is class A (so would consider only the first octet in this case).

BTW the mask length for 255.255.255.192 is 26.

HTH

0 Kudos