12-15-2003 07:27 PM - edited 03-02-2019 12:22 PM
We recently installed a WS-C3750-48TS-E and loaded the newest Enterprise IOS image on it.
There are a few trunks configured on the switch and there are also about 10 servers connected to it. We periodically get flashing amber lights on only the server ports. Errors occur occasionally on these ports. I have included the controller output of one of the periodic blinking amber ports below.
I opened up a TAC case, but the tech was less than helpful. He quoted the following bug, but this is definitely not the problem.
-------------------------------
CSCec14238
Externally found moderate defect: Verified (V)
3750 connected using dot1q cause Runts increase on interface.
-------------------------------
The Catalyst 3750 documentation states that ports should only blink amber if "the port is blocked by STP and is transmitting or receiving packets". Well, we have "switchport host" issued on each of the server ports with BPDU guard and BPDU filter enabled. I know that the STP port state never changes, so I know this is not the problem. We never lose connectivity to the server when its port is blinking amber.
There are only 2 possibilities that I can think of:
1) The documentation is incorrect - the port blinks amber if there are errors
2) There is a cosmetic bug
-----------------------------------------
aus-core1#show controllers ethernet-controller fast 1/0/38
Transmit FastEthernet1/0/38 Receive
2470495362 Bytes 1518647103 Bytes
24502910 Unicast frames 32513196 Unicast frames
749706 Multicast frames 0 Multicast frames
2956574 Broadcast frames 2608 Broadcast frames
0 Too old frames 1518187216 Unicast bytes
0 Deferred frames 0 Multicast bytes
0 MTU exceeded frames 319302 Broadcast bytes
0 1 collision frames 456 Alignment errors
0 2 collision frames 0 FCS errors
0 3 collision frames 0 Oversize frames
0 4 collision frames 0 Undersize frames
0 5 collision frames 73 Collision fragments
0 6 collision frames
0 7 collision frames 2935484 Minimum size frames
0 8 collision frames 7232091 65 to 127 byte frames
0 9 collision frames 1820568 128 to 255 byte frames
0 10 collision frames 1355735 256 to 511 byte frames
0 11 collision frames 1379397 512 to 1023 byte frames
0 12 collision frames 17792985 1024 to 1518 byte frames
0 13 collision frames 0 Overrun frames
0 14 collision frames 0 Pause frames
0 15 collision frames
0 Excessive collisions 0 Symbol error frames
0 Late collisions 0 Invalid frames, too large
0 VLAN discard frames 0 Valid frames, too large
0 Excess defer frames 73 Invalid frames, too small
15990751 64 byte frames 0 Valid frames, too small
6453030 127 byte frames
1092956 255 byte frames 0 Too old frames
1234441 511 byte frames 0 Valid oversize frames
551685 1023 byte frames 0 System FCS error frames
2886327 1518 byte frames 0 RxPortFifoFull drop frame
0 Too large frames
0 Good (1 coll) frames
0 Good (>1 coll) frames
Receive
1518647103 Bytes
32513196 Unicast frames
0 Multicast frames
2608 Broadcast frames
1518187216 Unicast bytes
0 Multicast bytes
319302 Broadcast bytes
456 Alignment errors
0 FCS errors
0 Oversize frames
0 Undersize frames
73 Collision fragments
2935484 Minimum size frames
7232091 65 to 127 byte frames
1820568 128 to 255 byte frames
1355735 256 to 511 byte frames
1379397 512 to 1023 byte frames
17792985 1024 to 1518 byte frames
0 Overrun frames
0 Pause frames
0 Symbol error frames
0 Invalid frames, too large
0 Valid frames, too large
73 Invalid frames, too small
0 Valid frames, too small
0 Too old frames
0 Valid oversize frames
0 System FCS error frames
0 RxPortFifoFull drop frame
---------------------------------------
Can anyone offer any advise or suggestions?
Thanks,
Jordan
12-16-2003 07:23 PM
Well, I think I found the problem. It appears that the 3750 port lights will blink amber when errors occur on a port (which indicates and error with the switch documentation). I was using an older Intel Pro/100+ driver on numerous servers connected to the switch. The older drivers were generating alignment errors and collision fragments periodically. I upgraded to the new version of the Intel drivers and haven't received further errors. Since the upgrade, I haven't seen the lights flash amber
It would be nice if Cisco could confirm or deny my findings...
Jordan
12-17-2003 10:39 PM
There is a bug that addresses this issue, CSCec53648. Basically, the interface LED will blink amber when receiving valid dot1q packets which are less than 68 bytes (tag included). This will be expected behavior and no fix. Hope this helps.
12-18-2003 08:29 AM
Our problem is a different bug that has not yet been reported. This problem is not related to the above issue because it has nothing to do with trunks. We have portfast enabled on the ports with the global bpdufilter command enabled, so all BPDU's should be blocked anyway. We were seeing both collision fragments and algnment errors on the ports, which appeared to fire off the amber lights. We updated the Intel NIC drivers on the servers and that eliminated the port errors, which in-turn eliminated the blinking amber lights.
12-18-2003 10:49 AM
The bug states that Runts will trigger the LED to flash amber. What I meant was any packets less than 68 bytes with a dot1Q tag .. which would bring it below 64 bytes w/o the tag .. which is a Runt.
12-18-2003 10:59 AM
Then that must be it. I did notice some "collision fragments", which I would assume would be runts.
Thanks for your help!
Jordan
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide