Netonix Forums

Need some help on this as I can't find anything on the UBNT forums... We are running 2 AF5X links. At our "hub" they plug into the power injectors that came with them, and then into a HP ProCurve switch. We also have two Rocket M5 links from this location. Those pass IPv6 traffic and leases just fine, while the AF5X's won't.

Running 3.2.1 on all the AF5X Units. At the remote ends, they are run through Netonix switches, one being a WS-12-250B and the other a WS-10-250 AC. The former runs firmware 1.4.5RC8 and the other is 1.4.2. We are doing nothing fancy with the switches; each is in a single VLAN. No STP, no LAG or anything like that. Simple switch configs all around.

The Rocket M5 links also end with Netonix switches, all of them being WS-10-250-AC's with 1.4.2. At the hub location, they go through a mid-span power injector to get power, but terminate in the same ProCurve switch. I can't find anyone who has had a similar problems with these AF's so reaching out here in case anyone has seen this behavior before. Our core router is a MikroTik CCR1009-8G-1S+PC running very recent firmware.

Thanks,

Doug N

The Netonix switches pass IPV6 fine so this is not your issue.

However you are running very old firmware (which is not the cause just an observation)

I didn't have any suspicion of the Netonix products, and I do plan to upgrade the firmware soon. Just trying to lay out the problem as best as I can describe.

I wish I could give you more help on this matter but sadly we are not yet routing IPV6 at my WISP.

I now have a support number with Ubiquiti on this issue. But really surprised that no other WISP has tried running IPv6 on their networks. Or maybe it's just they have never tried it with AirFiber units. Still, the old Rocket M5's handle it just fine.

Also a little worried that the last firmware release for the AF5X was close to a year ago. Maybe this will prompt a new release.

Doug

Oh, we did find a config error in our pool setting for the IPv6 DHCP server. That fixed one of our links. But the other one, with the bulk of our customers, still would not get v6 addresses at the CPE. So this morning I upgraded the WS-12-250B to 1.4.6. Since it was already running 1.4.5RC8, there was no problem with the upgrade (or so I thought). This afternoon, I adjusted the MTU rate on all the ports from 1518 to 1528, the latter being the default on the other Netonix switches that we have. About 10 minutes after that, I lost all links to that tower, and could no longer reach the switch.

Arriving at the tower location, I noted that all the POE lights were still lit. Upon a hard reboot, though, none of them would come back on. Only two ports would link up at all, but I could not reach the switch on its assigned address or the default. Pressing the reset button on the front has no effect. Thankfully, I had a spare WS-10-250-AC, but we are one port short. At least all of our main links are back up. Not sure if it's worth an RMA since the board revision is pretty old. Told my boss to get one with at least 12 ports since two always seem to be SFP cages.

Not sure what happened there. It is just one page and saved at about 8Mb. Very strange. Since then I have reinstalled WB12 and the links now seem to work fine. Sorry to have taken up your time Pablo.