- What should I be looking at to troubleshoot Ethernet In Discards.
2. What settings should I put in for my customers to have a 1M down 512 Up connection with no burst?
Thank You.
1 - On SM? I’d look first to attached device (make sure both device and SM auto-negotiate speed/duplex correctly, swap out different router/PC and clear stats, see if problem is tied to customer equipment) then check ethernet cable and lightning/surge arrestors, then SM’s ethernet port (swap radio).
2 - Try something in QoS on SM (NOT AP) like:
Sustained Uplink Data Rate : 500 (kbps)
Sustained Downlink Data Rate : 1000 (kbps)
Uplink Burst Allocation : 0 (kbits)
Downlink Burst Allocation : 0 (kbits)
Should limit them to 500k up, 1m down, no burst above that. If you provision all your customers the same way, it’s pretty easy to just hit those four values when you configure SMs. If you’re intending to limit only certain customers that way, look into central provisioning, the up-front setup is well worth the long-term savings in time/effort. (Prism, Radius, JungleAuth)
j
The Ethernet In Discards is on Backhaul. Direct shielded connection to a CMM Micro.
I was going to put the rate limit in like you mentioned but it seemed to simple. I just wanted to check myself.
eww. How long is the cable, what’s the nature of the tower (cell tower, radio/tv tower, water tank, building - CMM with gear, CMM at ground level, etc) and how much traffic is it actually handling? I’m assuming it’s a Canopy 10meg or 20meg backhaul unit, deriving sync from the CMM at the master end?
Yeah, QoS is pretty simple on Canopy, and therefore only mildly flexible. PMP320 gear on the other hand is ridiculously flexible/configurable, but as a result is quite complicated to manage well. (though for something like a single fixed class of service it’s dead-easy to manage)
j
Ditto on the cable length, proximity to high power transmitters, etc. To further what newkirk said, if your link speed or traffic requirements are low enough you could drop the ethernet to 10 Mbps/Full Duplex and probably not worry about it. This is just a workaround, though, for a cable-related fault. Other things to check are surge suppressors, cable ends for corrosion, water, etc, damage to the actual cable (I had one cable get climbed on on a colo tower… don’t ask), and of course the ethernet ports at either end.
The cable is pushing 260ft in length. It’s Belden 7919A shielded outdoor Cat5e. It sounds like I need to look at the cable itself. Again, this is happening at two AP’s on two different towers. I have the eliminated the surge suppresors and run the cables straight from teh BH to the CMM Micro.