[Info-vax] PEA0: Errors

[JF Mezei]

Christoph Gartmann wrote:

>    FCS Rx          : 0                  Drops Tx        : 2,654,260
>    Alignment Rx    : 0                  Collisions Tx   : 117,101
>    Runts Rx        : 0                  Late Colln Tx   : 0
>    Giants Rx       : 0                  Excessive Colln : 0
>    Total Rx Errors : 0                  Deferred Tx     : 0


I asked my friend Mrs Google for help and she told me to look at:

http://www.hp.com/rnd/library/troubleshoot_lan.htm


Looks like the "Drops Tx" is really dropped packets (as opposed to loss
of carrier). And dropped packets are to be expected when you have more
traffic being sent than a slow link can handle.


A switch can handle a burst of traffic followed by silence since it can
then send the few packets it received during the burst over the slow
link. BUt with sustained throughput, its buffers will overflow and it
then has to drop packets.

It is possible your 3com switch may have larger buffers. It may also
have a different philosophy of which packets to drop when buffers are
full which might preserve enough SCS keepalive packets to keep that node
in the cluster.

It is also possible that your HP switch may have had some
misconfiguration wich would have caused your spam box to go nuts and
start generating bad IP packets by the dozen.

Does the spambox misbehave when on the 3COM switch ?

Note: whenever you have traffic going from fast to slow links, you have
to expect problems unless you architect the links properly.