I regularly look through various customers statistics to keep ahead of problems, and one thing I can't make sense of is when I see some numbers next to RcvFifoNoBuf in the ethernet statistics field.
Does anyone know what this means? Usually, it's accompanied by a similar count by inerrors, but I can't make much sense of it, because everything works great. The internet, and this board don't seem to have any information other than this thread... which is just conjecture.
Well technically speaking RcvFifoNoBuf or inerrors or CRC basically are the same. All three counters indicates the packet loss or packet loss because of OSI layer 2 FCS fail. In order to make sure that there is no issues on LAN side you should check :-
1.)Duplex settings on both end should be same.If same try to reduce one level down i,e if 100Mbps set,low this down to 10Mbps on both ends and then see if you still face counter increament.
2.)Faulty cabling and broken hardware could cause this issue.
3.) On the Wireless side I suspect there may be the interference issue. So please run the Spectrum Analyser on both AP and SM side and share it here.
Hoping to hear back from you soon.
I'm sorry I forgot about this inquiry, so I hadn't checked back for a while.
There are a lot of sites with this information, and I think I was getting to an issue with cabling or hardware on the customer side as the primary cause. The problem with this part of testing that idea is I don't have access to the customer's router without a lot of wasted time.
The numbers I see are typically on the scale of 1 every other week, at most, so they aren't really worth tracking.
I'll keep an eye out, and follow up if there is more action on this front.
Conjecture and unanswered posts ;)
Buffer errors are not the same as CRC errors at all. The former means that the device could not allocate memory to store the frame before forwarding, the latter means that the frame checksum was determined to be corrupt.
Are you seeing this on the SM or AP? We have issues in the network sometimes where micro-bursts of traffic will overwhelm the receive buffer of a device. This happens to us typically on switches where the site router is connected at 1 Gbps but an AP is connected at 100 Mbps - a siginficant amount of data will hit the 100 meg port at 10x the speed that the port can transmit and eventually overrun the transmit port's buffer. TCP traffic will not cause this problem, unless there are many TCP flows aggregated. UDP can absolutely create these problems.
How are your RF Overload stats looking?