[vortex] Possible FAQ: 3c905C driver

[Sam Wilson]

Hi,

Bogdan Costescu wrote:
> > Firstly, apologies as this maybe a FAQ ...
> 
> In fact, you're the first to ask this question in clear! Beware, you'll be
> the one remembered for starting the war 8-)

lol!
<SNIP> 
> What are the symptoms of the "dropping dead" case ? How are you noticing
> it ? Do you need to do anything to get it running again (like an "ifconfig
> eth0 down; ifconfig eth0 up") or it just recovers by itself ? If it just
> recovers, what is the time needed for this ?

Well, after it "drops dead" (as I said, all it normally takes is to
restart named), sometimes all one needs to do is ifdown, ifup (this is
rarely the case though). However, on bad cases it seems to kill syslogd
(syslogd is dumped to the terminal), can't login and the only solution
is to press the reset button! The worst host has (running RedHat 6.1 +
RH updates) a GigaByte GA bx2000 motherboard and CPU is a PII 233.

Once it's been restarted, problem doesn't appear again for about 2-4
weeks. But again, every time it is triggered by restarting named!

> What is the load and how do you measure it ? What does "namely restarting
> named" means ?

;-) Ok. The one host that seems to "drop dead" the most has
http/ftp/smtp/pop3 and primary dns (with about 30 domains). Servicing
these 30 domains plus 1500 ISP customers. Do you want some more
quantitative load measures? 

Um, the other thing is that we had two epic100 NICs in this host,
similar things happened, but I was able to find a posting to
linux-kernel that pointed to the exact problem. Perhaps it's got nothing
to do with the 3com or epic100 driver?

> Are you running in 100Mbit FullDuplex mode ?

Yep.

<SNIP> 
> > o Andrew Morton's 16Aug00 (as in 2.2.17-pre17)
> > o Andrew Morton's 24May00 (as in RedHat kernel 2.2.16-3)
> 
> >From stability point of view, these should be almost equivalent. AFAIK,
> the latest (from 2.2.17-pre18) has some small modifications to handle some
> Tx reset cases, but when using a switch, these conditions should not
> appear. As this driver is currently distributed with the kernel, it gets
> quite a lot of testing...
> I personally only use these two (the RedHat one on our desktops and the
> latest one plus my polling mode modifications for 905C on our clusters).

I've put the 16Aug00 driver on one host that has little/no load. 
 
<SNIP>

Anyway thanks heaps guys :-)

Regards,
Sam