PPP not exiting

Michael Collard <quadfour@iinet.net.au> wrote:

> Sure, that last connection was a test one only. A demo of a connection
> where we did a little traffic etc is here:

....

> Jun 9 14:14:50 natalie pppd[19803]: Script /etc/ppp/ip-up finished
> (pid 20406), status = 0x0
> Jun 9 17:40:16 natalie pppd[19803]: rcvd [LCP TermReq id=0x2 "User
> request"]
> Jun 9 17:40:16 natalie pppd[19803]: LCP terminated by peer (User request)

The peer requested PPP termination, and pppd agreed.

> Jun 9 17:40:16 natalie pppd[19803]: Script /etc/ppp/ip-down started
> (pid 23969)Jun 9 17:40:16 natalie pppd[19803]: sent [LCP TermAck id=0x2]
> Jun 9 17:40:16 natalie pppd[19803]: Script /etc/ppp/ip-down finished
> (pid 23969), status = 0x0
> Jun 9 17:40:19 natalie pppd[19803]: Connection terminated.
> Jun 9 17:40:19 natalie pppd[19803]: Connect time 205.6 minutes.
> Jun 9 17:40:19 natalie pppd[19803]: Sent 1825263 bytes, received
> 578076 bytes.

This is a clean disconnection by pppd.

> Jun 9 17:40:19 natalie pppd[19803]: using channel 82
> Jun 9 17:40:19 natalie pppd[19803]: Using interface ppp1
> Jun 9 17:40:19 natalie pppd[19803]: Connect: ppp1 <--> /dev/tts/0
> Jun 9 17:40:19 natalie pppd[19803]: sent [LCP ConfReq id=0x2
> <asyncmap 0x0> <auth pap> <magic 0xc52c931f> <pcomp> <accomp>]
> Jun 9 17:40:46 natalie last message repeated 9 times
> Jun 9 18:17:45 natalie pppd[19803]: Hangup (SIGHUP)
> Jun 9 18:17:45 natalie pppd[19803]: Modem hangup
> Jun 9 18:17:45 natalie pppd[19803]: Connection terminated.

Pppd attempted a reconnect immediately - I don't know why. The result
of the attempt looks like a LCP timeout, which is normally terminated
cleanly, i.e., a kill -9 or any other kill signal should not be necessary,
pppd should just clean up and terminate.

> Just noticed the end of that log, pretty strange stuff. Maybe its
> related to the passive option? I will remove it and observe the next
> connection's hanup. I've also just created a script that will kill
> -9 pppd's PID at the point where it runs the ip-down script. That
> should help me out :)

The ip-down script is run by pppd after the PPP interface is brought down.
Pppd should cleanly terminate itself after that, unless the persist option
is used.

>> What bothers me is the /dev/tts/0 that appears in the output above. That
>> should read /dev/pts/0, or /dev/pts/<some integer> . I suppose Mandrake
>> could have mangled things and created /dev/tts but it's very doubtful
>> that it would be used by pppd. But if "looks like" means you
>> hand-copied the messages then that might explain it. In fact, comparing
>> this with my own logs strongly implies that it *is* a hand-copy. I'd
>> expect a line with "Using interface ppp1" before the "Connect" line and
>> a 1-second delay between the "Connect" and the first LCP request.

> It wasn't hand copied, straight from the logs :)

Then perhaps the /dev/tts/0 is a devfs thing. I never used the devfs
and it is going away in the 2.6.x kernel series to be replaced by the
udev filesystem - whatever that means.

I still don't understand why there is no 1-second delay between the
"Connect" line and the first LCP request. See man pppd, the option
"connect-delay n." The delay is always present here.

>> That's surprising; I'd expect the PID file to be around after an abrupt
>> hangup. Or does "it hangs up" mean that it is pppd that actually hangs
>> up, after closing the PPP session?

> I think its a case where it hangs up but maybe crashes on terminating?

Perhaps so, except if it "crashes" then I'd not expect it to continue
running(?).

> I really don't know though. Perhaps its something to do with Mandrake's
> PPP. On a couple of times only when it hung, there was also an
> ip-down <defunct> process but I couldn't kill this directly, had to
> kill pppd.

It could be a Mandrake induced problem. I've never seen any other post
describing the problems you are encountering.

> Thanks for your help thus far. I will let you know how I go with
> the disconnect script I made.

I started this reply last night but other things took precedence, so it
wasn't finished until just now. I see from your most recent post that
you've got a work-around; sometimes it's "whatever works."

--
Clifford Kite Email: "echo xvgr_yvahk-ccc@ri1.arg|rot13"
PPP-Q&A links, downloads: http://ckite.no-ip.net/
/* 97.3% of all statistics are made up. */