Comcast provides 1200mbps downstream, probably more with over provisioning
I have an mb8600 which supports bonding ports 1 and 2 for a LAG channel so i could get an aggregate 1.2gbps throughput. And this does work with multithreaded downloads.
I have a surplus ex-4300 juniper switch i got from work.
Cable modem ports 1 and 2 are plugged into the Juniper switch's rj45 ports. The Switch is configured to recognize a LAG and handles the port bonding. That LAG Channel is assigned to a private "cable modem" vlan that only my pfsense VM is able to access. Pfsense has no problem communicating with the cable modem and picking up an IP. If it matters i'm running esx with an enterprise license i also got from work, and pfsense is the only vm configured to access the cable modem vlan. ESX is connected to the switch via a 40gbps link, it's a little archaic, but it works great.
Generally everything works fine for about a 5-9 days, then eventually everything goes sideways. The cable modems activity lights are blinking, i see no errors on the switch, but i can't communicate with the modem at all, i can't even pull up the diagnostic page 192.168.100.1 on the modem. It all just goes non responsive until i reboot the modem.
If i turn LAG off, unplug port 2 from the switch, and turn off port bonding on the switch, everything works fine.
Has anyone experienced this and have any ideas what the problem might be? I get the feeling that the modem probably has some kind of resource leak that just builds up over time.
↧