Very Computer

  I recently upgraded my kernel to 2.0.0 and to modules 2.0.0 and now for
some reason my request-route script no longer gets called by kerneld.  For
the life of me I can't figure out why not.  Here is the output from kdstat:

16353:job (0804b878) scheduled, type request_route ('166.82.180.10')
16353:job (0804b878) stored, pid=512
16353:finish_jobs
16353:SIGCHLD: job (0804b878), pid=512, status=2
16353:finish_jobs
16353:job (0804b878) unlinked
16353:job (0804bca0) scheduled, type request_route ('166.82.1.3')
16353:job (0804bca0) stored, pid=513
16353:finish_jobs
16353:SIGCHLD: job (0804bca0), pid=513, status=2
16353:finish_jobs
16353:job (0804bca0) unlinked
16353:job (0804b878) scheduled, type request_route ('166.82.180.10')
16353:job (0804b878) stored, pid=514
16353:finish_jobs
16353:SIGCHLD: job (0804b878), pid=514, status=2
16353:finish_jobs
16353:job (0804b878) unlinked
16353:job (0804bca0) scheduled, type request_route ('166.82.1.3')
16353:job (0804bca0) stored, pid=515
16353:finish_jobs
16353:SIGCHLD: job (0804bca0), pid=515, status=2
16353:finish_jobs
16353:job (0804bca0) unlinked

  The two IP addresses that show up are my nameserver.  It looks to me from
this output that the request is just falling through.. It never actually
calls the script.  I've checked the filename and directory of my
request-route script in the code and its fine. If anyone could give me a
clue about where to start looking I'd appreciate it.

        -Scott

---------------------------------------------
Scott Lampert

"Sing the Hare Hare, Dance the Hoochie Koo!"
---------------------------------------------

Firstly, the script file should be /sbin/request-route. You said this was
correctly set, so I'll believe you. =)
Secondly, check for any stray lockfiles running about. It's probably
/tmp/request-route unless you modified the stock script...

--

+----------------------------------------------+Do not meddle in the affairs |
|Mobius strip for sale... see back for details.|of DEC VMS 6.0, for it is    |
|Klein bottle for sale... inquire within.      |subtle and quick to anger.   |
+----------------------------------------------+-----------------------------+

->Firstly, the script file should be /sbin/request-route. You said this was
->correctly set, so I'll believe you. =)
->Secondly, check for any stray lockfiles running about. It's probably
->/tmp/request-route unless you modified the stock script...
->

  This isn't it either.  I have no lock file in tmp.  I've been ripping my
hair out over this for a week now and I'm no closer to a solution.  If I try
to ping an IP address (by number) while I'm not connected to the net ping
comes back and tells me "no route to host" over and over which is what I
would expect if I didn't have kerneld set up to autodial in.  However it
just never calls the script.  I thought perhaps it was something in my
modules.conf but deleting it and making a new one with the defaults has no
effect either.  I'm certainly at a complete loss.  What bugs me the most is
that it was working fine until I upgraded the kernel and modules to 2.0.0.

          -Scott

--
Scott Lampert

http://www.ioa.com/home/fortunato
"Sing the Hare Hare, Dance the Hoochie Koo"

:   I recently upgraded my kernel to 2.0.0 and to modules 2.0.0 and now for
: some reason my request-route script no longer gets called by kerneld.  For
: the life of me I can't figure out why not.  Here is the output from kdstat:

(kdstat output cut)

:   The two IP addresses that show up are my nameserver.  It looks to me from
: this output that the request is just falling through.. It never actually
: calls the script.  I've checked the filename and directory of my
: request-route script in the code and its fine. If anyone could give me a
: clue about where to start looking I'd appreciate it.

:         -Scott

: ---------------------------------------------
: Scott Lampert

: "Sing the Hare Hare, Dance the Hoochie Koo!"
: ---------------------------------------------

I was having a similar problem, although I haven't yet upgraded from 1.3.69k
to 2.0.0. I finally discovered the problem:

I apparently had hit a function key or something when I first entered pico
to modify the script, as a result a ~4 was inserted at the beginning of
the #!/bin/sh line.

Try executing the script by hand, if something similar to my problem is
causing yours, the script probably won't execute. If it does work and the
shell doesn't return any error messages, then you at least know that the
problem probably doesn't lie within the script itself.

------

On a related note, I can't seem to get the idle-disconnect option to
disconnect. According to pppstat, the link is never idle for more than 35
seconds, even though I've just been sitting there watching the pppstat
output since the link was brought up. Any ideas why?

Mike Delaney