Interface collisions

[Jack L. Stone]

At 12:20 PM 3.31.2003 -0600, David Kelly wrote:
>On Mon, Mar 31, 2003 at 10:34:05AM -0600, Jack L. Stone wrote:
>[...]
>> 
>> Also, I agree that the collisions are very small and were cached by the
>> switch, not lost necessarily. However, the sudden appearance over the past
>> 2-3 days indicates a change that is not for the better and more concerned
>> about the trend.
>
>Not "cached by the switch" else your rl driver would not have known
>about it. The rl driver logged the "collision" because it started
>sending a packet and was not able to copy it 100% in real time so it
>concluded somebody else was transmitting at the same time.
>
>If the card is configured in full duplex mode it should not be verifying
>copy of its own data when sending, by definition. Unless there is some
>sort of out-of-band communications between ethernet ports operating via
>full duplex.
>
>-- 
>David Kelly N4HHE, dkelly at hiwaay.net

1) I changed cables - NO fix
2) I switched switches - NO fix
3) I set the NIC to full-duplex "on the fly" - YES - Fixed

Setting the interface to full-duplex "on the fly" has now stopped all
collisions on this NIC.

And, you are right of course about where the caching is occurring with this
symptom.

Many thanks again....

Best regards,
Jack L. Stone,
Administrator

SageOne Net
http://www.sage-one.net
jackstone at sage-one.net