XG1541 10G throughput
-
We have a customer that is running those and we are wondering if throughput problems are "normal":
Customer did an iperf measurement from 2 VLANs behind the XG. Both VLANs are on the 10G links that go to their switches (that do 10G in between all the way). So they measured from/to servers in different VLANs that terminate on pfSense and while they get near 10G performance in the same subnet (pfSense not involved), it drops down to a measly 2-2.5Gbps between two servers on different vlans (e.g. vlan100/200.lagg0 while lagg0 is setup over both ix0/ix1 links and configured with LACP to their switches).
So as they never even reached a single 10G I'm wondering: where's the big blockade in the setup? Simple two VLANs on an single 10G link should show more then ~2.5Gbps alone. If it would be around 5, I'd say it's some weird setup or they have to share the 10G between the two VLANs but with 2-2.5Gbps something smells fishy.
Any insight?
-
Additional insights:
- Version is 21.05.1
- BIOS is 1.2c
- no NAT done, pfSense is internal firewall/router in front of their core switches
- ix0/ix1 are combined to lagg0 and connected to a core switch each (no crossover cabling)
- lagg0 is running normally (both channels active)
- rules on test VLAN interfaces are simple and to test were pass alls
so nothing out of the ordinary that would hinder traffic flow or performance
