Waiting for 3c509 to discard packet, status 2011.

Waiting for 3c509 to discard packet, status 2011.

Post by Jose R.Cordone » Wed, 05 Mar 1997 04:00:00


I'm running Slackware (the one with kernel v2.0.0) on one of my older
machines, an IBM 350-100DX4, and I'm using Etherlink III TP Cards.

Since it was the first machine I got my hands on when I started with my
employer, I had to quickly install whichever Linux I got my hands on first.
After that, I have downloaded Debian 1.2 and RedHat 4.1 to its disk and
so I've been using it as my PPP box and Fileserver.

Due to several factors, I have had to run the machine with *2* network
interfaces.  One for the company's LAN (very busy) and one to my private
switch.  Slackware sees both interfaces OK with the "Linux
ether=0,0,eth1" etc. LILO boot parameter (I haven't wanted to make it

However, as an NFS Server, it is *AWFULLY* SLOW.
I get tons of messages in /var/adm/message of the type:

Waiting for 3c509 to discard packet, status 2011.

Where the status number varies.

The error message is not verbose enough for me to figure out whether it
comes from eth0 or eth1, but I get the error message even if I function
with a single card.  I sometimes reduced the number of complaints by
installing an Etherlink III and instead of another of the same, an
Etherlink II.. but gesh, do I have to resort to this? :(

In either case, NFS performance is horrible (how can I push for my good
ol' Linux to take over tasks performed by machines worth 10x as much, but
which *do* the job? :-)

I get a *load of:
NFS server aaa.bbb.ccc.ddd not responding, still trying.

messages from its clients.

I know I could tweak things on the client side (with NFS options such as
rsize, wsize, etc.), and define a bigger window at the route table of the
server, but when I'm installing, say, RedHat over NFS.. it can take
*hours* before it gets to the Package Selection option and I don't know
how to control the nfs options that its install script uses to mount my
nfs-exported tree.  On a good day, said installation over nfs would take
1.5 hours.  On a bad day, it can take a whole night! :-(

I'm trying to go after one problem at a time, so I'd suposse my NFS Problem
has to do with the Ethercard problem (too long to service interrupts?)
CPU Load is usually below 1.00 on the NFS Server.
The Ethercards are configured with 3Com's 3c5x9cfg.exe utility,
relevant choices:

For each card,
Machine Type: "Optimize as Server"
Machine's Serial Port 38400bps (I suposse the utility uses what I say
here to predict what kind of interrupt latency the card will have)
Plug-n-Play: Disabled

Then one card is on 0x200 and the other on 0x300.  One at IRQ 10 and the
other at IRQ 11.  No Conflicts AFAIK.
The only other relevant thing as to why it may be * and taking too
long to service the network cards, is its usage of IDE HDs (but they are
off the on-board IDE Controller.)

Can anyone out there HELP?!? :-|

Thanks in advance,
Jose R. Cordones


1. 3c509 : Waiting for 3c509 to discard packet, status 2001. ???

I get a lot of "Waiting for 3c509 to discard packet, status 2001."
with my 3c509b-combo network card... (I think this problem appears
somewhere between 2.0.0 and 2.0.24).
For example, I get a lot of them (sometimes more than 1 per minute)
with X (over the network, of course).

Where's the problem ?
Any idea ?


        Cedric Maion

2. ISPELL: dictionnaire en francais.

3. Waiting for 3C509 to discard packet - status 2001

4. failed connection attempts

5. Waiting for 3c509 to discard packet, status 2001.

6. initramfs patch

7. Waiting for 3c509 to discard packet, status 2001

8. Booting via Fibrechannel

9. Q: Waiting for 3c509 to discard packet???

10. waiting for 3c509 to discard packet

11. 3c509b - infinite loop in interrupt status 2011

12. "eth0 Infinite loop in interrupt, status 2011"

13. eth0: Status 2011