> I've had ppp working in -auto -alias mode for quite some time, but
> lately I've been getting error messages like the following when
> something I do on the Unix box causes a reconnect to the net.
> RTM_LOSING: Kernel Suspects Partitioning: len 124, pid: 0, seq 0, errno 0, flags:<UP,GATEWAY,HOST,DONE,WASCLONED>
> locks: inits:
> sockaddrs: <DST,GATEWAY>
> <machine-name-I'm-trying-to-reach> <ISP's-dynamic-name-for-my-computer>
> It doesn't happen all the time, but when it does, the access to the
> outside world is blocked up, and I must either:
> a) Telnet to the control port of my local machine, then do:
> delete! all
> though the dial will sometimes get done for me, and if it starts
> before the delete! all is finished, then I must follow plan b) below.
> b) Find the ppp process number and send it a kill, and then restart
> Is there some more elegant way of coping, preferably one that doesn't
> require manual intervention?
I should probably add that this is quite likely due to something that
the ISP have changed of late. For the longest time things worked
flawlessly, then it went bad. I reported it, and after three rounds
of begging it got better for a while. Now it's back.
Getting the ISP to do the right thing is not the easiest -- they're in
the middle of a take-over, and I understand that all of tech support
have just been told that they'll need to re-apply for their jobs with
the new owners (which move I'm sorry to say I quite understand, and
have to support). At any rate, they don't support anything but Mac
and windows -- the typical reaction that I get when I call up and tell
them I can connect with FreeBSD (or even NT) is utter amazement.
But so long as one understands the kind of thing that the
behind-the-sce4nes techies might have been trying to dow when they
made this change, it should be easy to come up with a workaround for
it. That's the sort of elpe I'm seeking.