Comware

hi,

on a 5406 copper-module i've configured speed-duplex auto-10-100.

the status (show interface brief) for every port is 100FDx.

but all ports have a lot of tx drops (no other errors). the log is full of warnings like "high collision or drop rate". the port utilization has a maximum of 1%.

what's the problem in detail? i don't think that's a performance-problem. should i enable flow control?

hi Sven

1-check utp cable
you have cable cat5 or better (cat5e or cat6)

you make reinstall punch down patch panel and connector.

2-check nic or nic driver
please you check nic include 802.1 and 802.1u standart ifneed change your pc nic

standart 802.1u Fast ethernet connection porotocol number nic must be give this standart.

cenk

one little question...

tx drops are only packets are going out of the switchport directly to the edge-device, right? that means, the switch can't submit all packets to the device - hence the switch drops packets.

what i mean...

a tx-drop should be only a client- or cabling-problem or not??

please you can control pc nic support 802.1u protocol
if don't include 802.1u protocol pc nic
please change
(for 100 base FDx connection nic include 802.1u protocol)

and check utp cable

essentially... for good working infrastructure
we make many test method on utp cable
next,elfext,psnext,pselfext,RL,in Loss etc.

cenk

what type device connect on 5406 switch this port router ,switch,access point or pc

cenk

whatever vendor edge switch

Usually TX drops on 10/100 ports are caused by brief transient loops on Gigabit ports while spanning-tree converges.

Another reason could be due to a Gigabit device which is sending out traffic at line-rate to a broadcast/multicast address. The buffers for the 100Mbit ports simply cannot keep up so packets will get dropped.

The main question is - are the TX drops continuously increasing or are they now stable? If they are stable, I wouldn't worry about it but would keep an eye on it.

Hallo Herr Bergmann

Wie nutzen diese Einstellung im Zusammenhang mit 5400zl auf der Switch Seite und meistens Windows XP oder Vista Systemen auf der Hostseite.
Drop Fehler treten nicht im beschriebenen Maà auf.

Schlieà t man mögliche Ursachen wie Kabel und Dosen aus, liegt es Nahe auf der Hostseite zu suchen. Zumal, wenn der Groà teil der Systeme gleiche Treiber, gleiche Einstellungen, gleiches Betriebssystem usw. verwendet.

Sind wirklich alle Ports betroffen?
Wenn nein, wo liegen die Unterschiede auf der Hostseite?
Wenn die Fehler nur beim Starten der Hosts auftreten, welche Anwendungen sind da im Spiel?

Viel Erfolg!
Carsten

hi Carsten
please speak english
I'm not understand germany

many people not understand germany in this forum

update:

the drops only occur when the pc's on the switchports are boot up or shut down!

Sven Hello!
That is precisely the time when an image of a workstation is drawn
and where the image on the server is back.
Can you see that only those communications is concerned?

If it's only happening at those times, I'd ignore it. If you do a 'show int <PORT>' what's the average broadcast/multicast packets per second?</PORT>