cisco 2924 errors

[chicken1212]

Hello everyone, i need some comments or help on my 2924 which has some errors at the moment " 17545 input errors, 17545 CRC, 0 frame, 40159 overrun, 40159 ignored" any advice would be appreciate on how to reduce or eliminate those errors, as some of the servers often get sudden disconnection on FE ports

here is the complete status of my UPLINK port

FastEthernet0/1 is up, line protocol is up
Hardware is Fast Ethernet, address is 00b0.64b5.70c1 (bia 00b0.64b5.70c1)
Description: UPLINK
MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec,
reliability 255/255, txload 32/255, rxload 10/255
Encapsulation ARPA, loopback not set
Keepalive not set
Full-duplex, Auto Speed (100), 100BaseTX/FX
ARP type: ARPA, ARP Timeout 04:00:00
Last input never, output 00:00:00, output hang never
Last clearing of "show interface" counters never
Queueing strategy: fifo
Output queue 0/40, 0 drops; input queue 0/75, 0 drops
5 minute input rate 4256000 bits/sec, 4922 packets/sec
5 minute output rate 12642000 bits/sec, 4205 packets/sec
1873245754 packets input, 3020828584 bytes
Received 325419 broadcasts, 0 runts, 0 giants, 0 throttles
17545 input errors, 17545 CRC, 0 frame, 40159 overrun, 40159 ignored
0 watchdog, 0 multicast
0 input packets with dribble condition detected
621202313 packets output, 3886290131 bytes, 0 underruns
0 output errors, 0 collisions, 4 interface resets
0 babbles, 0 late collision, 0 deferred
0 lost carrier, 0 no carrier
0 output buffer failures, 0 output buffers swapped out

 

[KiscoKid]

To be fair, the amount of errors you're seeing there is an exceptionally number (less than 0.001%) of total packets passing through that port. I personally would only really be concerned if my input error ratio was 1% or greater.

However, CRC errors are typically caused by hardware faults - NIC and cable are common causes. It may also be the switch port itself possibly.

The overrun errors are just an indicator that the interface receiver was unable to hand received data to a hardware buffer because the input rate exceeded the receiver's ability to handle the data - i.e. congestion. Again the percentage of these errors compared to the total packets through the device is extremely low and probably not a concern.

I notice the duplex has been hard set but the speed hasn't. I always *ALWAYS* try to fix speed and duplex for my server connections to avoid any potential issues here. In my opinion, I would suggest fixing the server to 100mb full duplex and similarly for the switch port.

Incidentally any idea what the load on this device/port during busy periods. I see a 12.5mb load on it in your snapshot and was wondering, if the load tends to be much higher than this, that perhaps a Gig Ethernet upgrade/port aggregation approach could be useful.

HTH

 

[vipergg]

Seeing that you have the speed set as auto , set your duplex as auto also , you could have a mismatch the way it is .

 

[jneiberger]

I would also say that you have a duplex mismatch. Set *both* sides of the connection to autonegotiate both speed and duplex.

 

[chicken1212]

i must really thank everyone for useful tips and comments. This UPLINK port is connected to my ISP's router which is fixed at 100mbps FULL duplex. So i guess i also need to fix my UPLINK port as 100mbps FULL duplex. This will eliminate any mismatch between my UPLINK port and ISP's router.

However i would prefer the FE ports to operate at 10mbps FULL duplex, to control BW and prevent any svr from abusing BW. In this case, what is the best way to set on each FE ports? Speed operation will have to be fixed 10mbps. but Duplex is auto or fixed "FULL Duplex" ? currently all the servers do auto-negotiate to best settings upon start up, and follows according to FE port's. I am running Fedora Core and is there any way to Force FULL Duplex & 10mbps operation on my NICs ? ^__^

 

[chicken1212]

Oh yes, just to add on, here is my ISP's router config that is connected to my UPLINK port..

Sh int
<<
FastEthernet15 is up, line protocol is up
Hardware is FastEthernet, address is 00e0.520b.11b6 (bia 00e0.520b.11b6)
Configured speed 100Mbit, actual 100Mbit, configured duplex fdx, actual fdx
Member of L2 VLAN ID 1, port is untagged, port state is FORWARDING
STP configured to ON, priority is high, flow control enabled
mirror disabled, monitor disabled
Not member of any active trunks
Not member of any configured trunks
Port name is AXGN
Internet address is xxx.xxx.xx.xx/xx, MTU 1500 bytes, encapsulation ethernet
5 minute input rate: 6676064 bits/sec, 4077 packets/sec, 7.32% utilization
5 minute output rate: 3502936 bits/sec, 5190 packets/sec, 4.32% utilization
1842826658 packets input, 3752496022 bytes, 0 no buffer
Received 1686 broadcasts, 0 runts, 3868 giants
6023209 input errors, 2395014 CRC, 2371887 frame, 0 ignored
188776 multicast
2170159830 packets output, 3242110555 bytes, 0 underruns
0 output errors, 0 collisions
>>

And this UPLINK port have to handle thousands of packets and as much as 30Mbps at peak hours. So there may not be enough buffer or whatever. I also have 2950-T with Gigabit ports on my another 100Mbps line. Is this 2950 switch better than 2924 in terms of ability to handle large amount of packets & heavy BW? i am not so sure about the technical specifications.

 

[vipergg]

Simple rule , if you hardcode your switchports to 10/full then you must hardcode the server nics to 10/full also otherwise you will have a mismatch and the corresponding errors and slow response that goes along with it. Hardcode your link to your isp at 100/full , he has his hardcode and auto will not work.