Hello,

I?m having some problem understanding the cause that makes the error counters increment on some ports on my Cat6513. Here is the show counters output:

Switch1> (enable) sh counters 5/5

64 bit counters

0 rxHCTotalPkts =2149565

1 txHCTotalPkts =67102015

2 rxHCUnicastPkts =889

3 txHCUnicastPkts =41016608

4 rxHCMulticastPkts =0

5 txHCMulticastPkts =7202712

6 rxHCBroadcastPkts =2148676

7 txHCBroadcastPkts =18882695

8 rxHCOctets =157301256

9 txHCOctets =9979756778

10 rxTxHCPkts64Octets =29219721

11 rxTxHCPkts65to127Octets =23588508

12 rxTxHCPkts128to255Octets =1367389

13 rxTxHCPkts256to511Octets =4630885

14 rxTxHCpkts512to1023Octets =884417

15 rxTxHCpkts1024to1518Octets =9560660

16 txHCTrunkFrames =0

17 rxHCTrunkFrames =0

18 rxHCDropEvents =2040268

32 bit counters

0 rxCRCAlignErrors =0

1 rxUndersizedPkts =0

2 rxOversizedPkts =0

3 rxFragmentPkts =0

4 rxJabbers =0

5 txCollisions =0

6 ifInErrors =2040068

7 ifOutErrors =0

8 ifInDiscards =2040068

9 ifInUnknownProtos =2040068

10 ifOutDiscards =0

11 txDelayExceededDiscards =0

12 txCRC =0

13 linkChange =0

14 wrongEncapFrames =2040068

You can see the errors on: rxHCDropEvents, ifInErrors, ifInDiscards, ifInUnknownProtos, wrongEncapFrames.

All ports where there is an IBM HS40 server with fiber optic gigabit ethernet interface intel PRO/1000 Server have the same kind of show counters output. There is no visible effect on server performance, they seem to work correctly, but the NMS keeps alerting for high input errors and discards on the switch ports and this makes hard to detect when other types of errors occur.

The switch is running hybrid CatOS 8.3(4) and IOS 12.2(17a)SX4. The module on which the servers are connected is WS-X6724-SFP.

Does anybody know how to solve this issue?

Thank you all