I'm setting up a virtual file server and have iSCSI connected using MPIO and the bandwidth doesn't seem to max out both NICs but allows about 20-40MB/s higher transfer so it is working but I'm not sure why it isn't higher than about 1.2Gbps. It does about 700Mbps on one NIC and 400Mbps on the other. The NAS has an NVMe cache so it should be able to handle 10Gbps easy IMO.
Next I teamed two NICs going to the same switch. That switch is connected via 1Gx4 LAG to our 1x per switch in a 4x core switch stack. If I transfer a file simultaneously from two computers on the core switch, the total throughput is around 800Mbps and bounces all over the place. Maybe if both clients are on the same core switch in the stack it's bottlenecking going over the same 1G LAG line? This is my first time working with teaming or a LAG so I only have basic knowledge. I know teaming doesn't increase P2P bandwidth but allows load balancing and failover. To me load balancing would be 2 clients each getting directed to one of the two connections allowing 1G each more or less.
Is my understanding wrong?
I just want to make sure our virtual windows file server has the same or better performance compared to our current NAS hosted SMB share with 2-NIC teaming.
No comments:
Post a Comment