Ethernet Connectivity

I am seeing a trend with bridge table on my SMs sporatically…

Usually once/month I’ll get calls from customers saying they have no connectivity, but I’ll login to their radio and determine their link is 100/100 and RSSI/Jitter is OK.

I noted the uptime on the radios to be 25 days. I know the bridge table timeout is default to 30 days.

The radios are setup with static public IPs with DHCP server enabled and DHCP client disabled.

The software is 7.3.6 on most of the radios I’ve seen issues with.

I’ve noted that customers that have this problem have routers or they are directly connected to the SM, either/or so I can’t tie it to a specific router.

Is there a direct reason that this screws up their connectivity? I really don’t want to upgrade to 8.2.2 at this time if at all possible…like I said, I’m just recently starting to see this.

A simple reboot of the radio resolves the issue. And if I telnet to the radio and try to ping 169.254.1.2 (their router or computer) I get no response and there is no ARP entry until after I reboot. Could I simply adjust the bridge table timeout from 30 days to longer or never without implications temporarily until I determine what the heck is going on?

Thanks,

Rich

I am seeing the same thing. I’d sure like to figure out what is going on here.

Guys, the Bridge Timeout defaul is 25 minutes NOT days.

Have you got ANY radios in your network running version 8 ? Do you see any CRC errors on the Ethernet Stats? My experience is that Version 8 software on your network sometimes causes some SMs running 7.3.6 to show CRC errors. The cable is always fine.

Yeah, our 900 system has some mixed 8.1.5 radios. Those are not an issue though.


I did increase the timeout on the radio to 500 minutes last night on a customer. I have one today that I might change.


Would extending the time out to the maximum alloted minutes make more of a change right now, but cause the issue to crop up at a later time, forcing them to reboot?

I’m sorry – I was having brainfarts as I was typing the original message.

This is related (I THINK) to the DHCP Server Lease Timeout (Which defaults to 30 Days).

It seems that it’s not renewing or something is choking.

0 CRC errors on the radios that croak and need rebooted.

One other thing is that it seems to be the same people monthly that have this problem; but as we install new customers, it’s a month’s time before we realize it could be one of the new customers too.

I’m probably going to reflash the radios in question that have the problems and wait another month to see what happens.

Edit: I wonder if it’s related to the bridge table entry? Just a thought I suppose.

we have the bridge entry timeout at 500 minutes or longer. So far so good.

What advantage do you see with having this timeout set so high?

It’s longer than the ARP timeout on the router.

advantage = no problems and less ARP traffic.