on a somewhat related issue...
I just had a horrible evening. Dropped into my local ISP to check up on
things and started playing with the "what is going on here" problem.
The details:
2.1.124, 64M RAM, P75, nothing special about this machine, save for the
fact that it has four ISA internal modems. It's located about 60km from
where I am, serving as one end of a 4x33k6 EQL link to a much faster
connection on eth0.
The other end is similar, 2.1.124, 64MB RAM, P166 I think. four external
33k6 modems connected via a BOCA 16 port thingy. Call the former Computer
A and this one computer B.
Please excuse the technical terms, I'm drinking a rather strong
screwdriver. :-)
This machine has been up for about 46 or so days and in the last two weeks
we've been experiencing strange probelms with the modems. More
specifically: dialling them directly works great, but try to get ppp to
work across them will yield any two of the four working. debug logs show
that B sends to A a config request and A sends to B a config
acknowledgement, but B never sees the acknowledgement. Indeed, the RX
light on computer B's modems never flickers once pppd has been started.
I've reset computer B at least 6 times to no avail. always two of the
four work. When both of these boxes were "fresh" I had no problem. Then
one wouldn't work... then two...
A cat of /proc/interrupts shows over 560 MILLION interrupts on one of the
modems, and just over 100 million on the other three for computer A.
Naturally I didn't want to reset computer A as it always gives me the
willies to do such a thing as it's a 40 minute drive if something goes
wrong.
Well I reset it. "/sbin/shutdown -r now" and waited.
Here comes problem 2... When I finally got someone local to there to
reset it for me, he had said that it was stuck on "can't exec runlevel 6"
or some such thing. in essence, it had never rebooted. he hit the switch
and all was roses. logs showed NOTHING unusual. just the "received
signal 15... received signal 9" and then nothing. wtmp shows only one
restart (the one that worked).
Funny thing is that after computer A reset, all four modems work happilly
again. Again -- these modems work FINE if you just dial in as a terminal
session, but PPPD on computer A doesn't seem to actually SEND anything,
even though it says it is.
If anyone has any ideas as to either of these problems, I'd be SO
grateful. I already owe one person a case of beer, so what's a few more
drinks? :-)
to recap:
problem 1 - modems with exceedingly high interrupt counts somehow confuse
pppd, but work fine otherwise?
problem 2 - not being able to exec rc.6 when it is indeed +x
Thank you for your time,
Andrew
oh yes -- box A is slack 3.4, box B is RH5.1. both are running at least
the miminum latest program versions as outlined in
/Documentation/Changes.txt
Andrew
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/