Jan Knutar wrote:

Tony Sivori wrote:

Since BOINK was due to replace Seti@Home I with Seti@Home II Real Soon,
I was just going to put up with this occasional yet annoying problem.
But Real Soon has stretched into many months and I tire of this glitch.

That no other internet traffic works either, would indiccate a problem
with the modem, modem drivers, or ISP, and not the seti@home client. The
other symptoms described, looking for modem which is never found, makes
me believe more in the modem + drivers acting up, although it's possible
that kppp leaves pppd running, keeping the modem open and unavailable,
but this is more unlikely, imo. My bet is the modem drivers.

I agree something is happening to the modem driver or kppp. But I am
convinced that seti is the cause (although the actual defect may lie
elsewhere). This problem comes and goes with running seti.

The binary only drivers have probably never been tested with a mandrake
'enterprise' kernel...

When it happens the next time, try ping as you already mentioned
yourself. Try see if there are any error messages in dmesg (type dmesg
in a shell), and maybe try look at the logs in /var/log/messages...

Checking them just now, I don't see anything incriminating.

Not being able to open state.sah properly sounds just like it would be
empty, the S@H client probably clears it at one point or another when a
work unit is completed...

You're probably right about that. The last time the connection crapped
out, it did happen to be right at the end of a work unit. But other times
when it would not open, it was in the middle of a work unit and it
couldn't have been empty. However, none of those times did I bother to
force it to open with Kwrite.

-- 
Tony Sivori