Bandwidth down crossing ZS

[Nigrofasciatum]

Hi,

I have a bond with 2 VPNs between 2 ZS machines. On “server” side I have simetric 1Gb/s access, and on the “client” side I have 2 access:

Access 1 >> VPN1 >> 100/10Mb/s
Access 2 >> VPN2 >> 20/20Mb/s

I access to Internet trough the server side with servers public IP address.

Bonding is working fine, whit a maximum troughput 80/15. I think this is usual because of tcp retries & out of order packets due to different latencies on both ISP access.

But the question is, if on the “client side” I delete Bond interface like GW and I use the Access 1 like GW, I obtain 100/10Mb/s, but If I use only Access1 on the bond (only VPN1 like member), I obtain only 18Mb/s. ¿Whats the problem?…There is only 1 access working with the bond, so there shouldn´t be problems with retries or out of order packets due to different latencies. VPN or bond limit?

I,ve been monitoring both machines, and there is not overload. Where could be the problem? Why the default view of the bond appears only 10Mb/s on each VPN?

Bonding Mode: dynamic balancing
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 2000
Down Delay (ms): 0

Slave Interface: VPN00
MII Status: up
Speed: 10 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: ca:fe:ca:fe:ca:fe
Slave queue ID: 0

Slave Interface: VPN01
MII Status: up
Speed: 10 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 01:01:01:01:01:01
Slave queue ID: 0

[Author]

Posts