I've installed RH 6.1 on my test partition, and it shows a
very unstable PPP dial setup. I've configured the PPP items
just like the production RH 6.0 system, using netconf. I've
repeatedly checked all settings but the problem remains. The
RH 6.1 system fails to connect to my ISP, restarting the
pppd daemon incessantly. In the /var/log/messages file I see
complaints about pppd not being able to find a suitable
secret to authenticate the link. Occasionally it does
connect (odd) but most of the time netconf says after 15
seconds that the ifup-watch program has failed and asks me
to kill it. I believe this is caused by the new initscripts,
but I haven't demonstrated that.
I have the same problem.
A workaround to this is to add
to PPP options (there is an option for this in netconf)
After this workaround pppd usually works more or less OK.
The next release of initscripts will fix this problem by using "noauth"
automatically in this case. For now, you can add noauth to the extra
PPP options as Pim suggests.
The problem is that pppd's behaviour regarding CHAP authentication appears
to have changed.
In the process of releasing initscripts-4.67, which will fix this.