08-21-2012 04:19 PM - edited 03-07-2019 08:28 AM
Folks
We have a brand new 2 x 3750-X stack running 15.0(1)SE3 with C3KX-NM-10Gs installed in each switch. An HP C-Class server enclosure with Virtual Connect Flex-10 modules is connected to each of the 10G ports using twinax cables and they seem to behave OK apart from some random packet drop which appears to be within the enclosure and not without. However, when we were investigating this, we found a little oddity ... even though we have jumbo frame support enabled on the 3750-Xs, we are getting a significant "Valid frames, too large" recieved count on the interface, and a "Too large" transmitted count. I've looked around and can't find a definitive reason for this count to go up - it doesn't seem to be a jumbo frames thing, more an unexpected-additional-field-in-a-header type thing, like when you have a VLAN tag in a DTP frame. Any thoughts as to why I might be getting these from a new server enclosure?
SW-STACK-CORE-3750X#sho int t2/1/1 controller
TenGigabitEthernet2/1/1 is up, line protocol is up (connected)
Hardware is Ten Gigabit Ethernet, address is a44c.116e.55b5 (bia a44c.116e.55b5)
Description: *** Server Enclosure B2P3 ***
MTU 9000 bytes, BW 10000000 Kbit/sec, DLY 10 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive not set
Full-duplex, 10Gb/s, link type is auto, media type is SFP-10GBase-CX1
input flow-control is off, output flow-control is unsupported
ARP type: ARPA, ARP Timeout 04:00:00
Last input 00:00:05, output 00:00:00, output hang never
Last clearing of "show interface" counters 2d15h
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 3453000 bits/sec, 1142 packets/sec
5 minute output rate 6076000 bits/sec, 1064 packets/sec
52092348 packets input, 20546616470 bytes, 0 no buffer
Received 15193 broadcasts (15193 multicasts)
0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
0 watchdog, 15193 multicast, 0 pause input
0 input packets with dribble condition detected
45009519 packets output, 23719388911 bytes, 0 underruns
0 output errors, 0 collisions, 0 interface resets
7596 unknown protocol drops
0 babbles, 0 late collision, 0 deferred
0 lost carrier, 0 no carrier, 0 pause output
0 output buffer failures, 0 output buffers swapped out
Transmit TenGigabitEthernet2/1/1 Receive
23722640051 Bytes 20546716139 Bytes
43629235 Unicast frames 52077194 Unicast frames
393269 Multicast frames 15935 Multicast frames
1009161 Broadcast frames 0 Broadcast frames
0 Too old frames 20336358788 Unicast bytes
0 Deferred frames 2053800 Multicast bytes
0 MTU exceeded frames 0 Broadcast bytes
0 1 collision frames 0 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 0 Collision fragments
0 6 collision frames
0 7 collision frames 1306 Minimum size frames
0 8 collision frames 25046107 65 to 127 byte frames
0 9 collision frames 10356470 128 to 255 byte frames
0 10 collision frames 5009947 256 to 511 byte frames
0 11 collision frames 2082849 512 to 1023 byte frames
0 12 collision frames 5590720 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 4005730 Valid frames, too large
0 Excess defer frames 0 Invalid frames, too small
17812 64 byte frames 0 Valid frames, too small
16287332 127 byte frames
14170891 255 byte frames 0 Too old frames
1172987 511 byte frames 0 Valid oversize frames
611101 1023 byte frames 0 System FCS error frames
2147552 1518 byte frames 0 RxPortFifoFull drop frame
10623990 Too large frames
0 Good (1 coll) frames
0 Good (>1 coll) frames
08-21-2012 11:21 PM
There's a scarcity in documentation which would explain, line-by-line, as to the meaning of the output above.
I, for one, am no expert but I've seen "valid frames, too large" even with ordinary PC and I just tend to ignore it. As far as I can see, the output to your "sh controller" is pretty much a two-thumbs-up for me. No issues there. No alignment, no FCS, no Symbols errors on the receive side. On the transmit side, I don't see any collisions frames. Again. Looks good.
I, however, am a bit curious as to the amount of "unknown protocol drops" in the "sh interface" output. I've never seen that before.
I am also curious about the SFP "media type" of "SFP-10GBase-CX1". I've never seen an output like this before.
08-22-2012 12:09 AM
Thanks for that, leolaohoo - I thought much the same but with the other issues around the Virtual Connect side of things, I need to be sure
The "unknown protocol drops" may be LLDP, which I know is turned on down at the HP end of things, but not on the 3750-X. Then again, knowing HP it could be any number of other odd things ...!
The media type is how the 3750-X represents the SFP-H10GB-CU5M twinax cables when connected to the C3KX-NM-10G - I've only used them before in Nexus 2Ks and 5Ks so not seen this before myself!
01-21-2016 09:07 AM
Hi there DarrenV.
I have the same 3750x x 2 Stack, same C7000 HP enclosure (with Flex 10/D, 10gbit card) and am seeing the exact same increment of "Valid Frames, too large". Same IOS version as well.
slight difference in the transceivers i'm using (we have the basic SR fibre 10gbit transceivers) .
I cannot see any drops on the interface statistics but the "show controller ethe ten 1/1/1" is showing the "valid frames, too large" increment all the time at quite a high rate
we've recently had some of our linux vm guests on the chassis start to get packetloss and although i've given the switch a clean bill of health, that statistic is rather annoying me. I'm sure the packetloss is at the HP/Vmware level but without doing a packet capture on the 10gbit interface not sure what else i can do to look at those (i.e. whether they are jumbo frames prancing through the network even though MTU is set at 1500 on the interfaces or they are other protocols like LLDP etc).
i know this thread is a little old now, but i thought i'd also join it and see if you happen to have had any problems from what you stated above.
thanks all.
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