I have a ppp interface set up to come up on boot, and this works pretty much as expected (the link does actually come up on boot). Unfortunately the initscripts seem to have some sort of timeout somewhere which does not give my modem enough time to negotiate and establish a connection. This means that the startup script will timeout and consider the interface to have failed, and it then skips on past starting up other services and tasks as listed. The interface comes up and is fully ready about 5 seconds after the initscripts have decided that there was a problem, and all works fairly normally after that. Could the timeout before the initscripts decide that an interface has failed to come up be either lengthened or user specified, perhaps in the ifcfg-* files?
Exists in RH6.2. Dials in background and connects.
Yes you're correct, but this behaviour is not ideal when other daemons in the startup (for example ntpd) are started _before_ the PPP link is fully up and routing. This is because any other service which relies on DNS resolution when it starts does not get a satisfactory resolve and thus sometimes fails (unless of course there is access to DNS via another link). In this case pppd being backgrounded after 30s is of no use whatsoever. The ability to control this behaviour would be nice.
This will be fixed as of initscripts-5.60-1; there is a BOOTTIMEOUT option.
*** Bug 16920 has been marked as a duplicate of this bug. ***