I've just set up a "working" Multi-Tunnel VPN connection, but the balancing is weired.
Setup:
Site 1:
virtual UTM 9.1
Uplink1.1: 100Mbit down / 6Mbit up
Uplink1.2: 100Mbit down / 6Mbit up
Site 2:
virtual UTM 9.1
Uplink2.1: 25Mbit down / 5Mbit up
Uplink2.2: 25Mbit down / 5Mbit up
If I either connect Uplink1.1 Uplink2.1 or Uplink1.2 Uplink2.2 via IPSec-VPN, i get the maximum possible bandwith with NetIO
TCP connection established.
Packet size 1k bytes: 612.07 KByte/s Tx, 375.33 KByte/s Rx.
Packet size 2k bytes: 626.75 KByte/s Tx, 382.09 KByte/s Rx.
Packet size 4k bytes: 631.49 KByte/s Tx, 394.23 KByte/s Rx.
Packet size 8k bytes: 621.46 KByte/s Tx, 395.23 KByte/s Rx.
Packet size 16k bytes: 637.13 KByte/s Tx, 394.85 KByte/s Rx.
Packet size 32k bytes: 605.03 KByte/s Tx, 393.22 KByte/s Rx.
Done.
Now i've connected
Uplink1.1 Uplink2.1
Uplink1.2 Uplink2.2
via "Bind Tunnel to Local Interface" on both UTMs. Tunnels are up and running. Then I've set up MultiPath Rules on both UTMs so split "by Connection" over Uplinkinterfaces. This is working, since i can see ESP Pakets on all 4 Interfaces with tcptump. But NetIO shows me something very disappointing (tested at least 20 times, all similar):
TCP connection established.
Packet size 1k bytes: 347.49 KByte/s Tx, 347.51 KByte/s Rx.
Packet size 2k bytes: 339.28 KByte/s Tx, 411.48 KByte/s Rx.
Packet size 4k bytes: 324.03 KByte/s Tx, 423.60 KByte/s Rx.
Packet size 8k bytes: 407.81 KByte/s Tx, 444.98 KByte/s Rx.
Packet size 16k bytes: 324.88 KByte/s Tx, 398.78 KByte/s Rx.
Packet size 32k bytes: 478.62 KByte/s Tx, 414.66 KByte/s Rx.
Done.
Although it should be somethink like 1078.62 KByte/s Tx, 914.66 KByte/s Rx.
Even the UTM-Dashboard shows that the Data is split up through both Interfaces...
Qos = off
No Traffic-Selectors activated
No traffic on Uplinks
I've also testes Samba File Transfers with the same effect.
Am I missing something or have i done something wrong?
This thread was automatically locked due to age.
