[Openswan Users] [Announce] xl2tpd-1.1.07 released

[Paul Wouters]

On Thu, 22 Feb 2007, Nels Lindquist wrote:

> Okay, I'm still being bit by this.  It had been bad enough with previous
> releases that I rolled back to l2tpd-0.69cvs20051030.
>
> On hearing the news that 1.1.07 was out, I upgraded yesterday hoping for
> the best.
>
> Unfortunately by this morning our L2TP users were unable to connect.
> There were two pppd sessions present, but neither of them were
> functioning.  Restarting xl2tpd (also had to "kill -9" one of the pppd
> processes) allowed the users to successfully connect again.
>
> I'm seeing lines like this in /var/log/pppd:
>
> > Feb 22 13:07:02 yxdgate pppd[14048]: LCP terminated by peer (^VuhX^@<M-Mt^@^@^@^@)
> > Feb 22 13:07:02 yxdgate xl2tpd[12929]: check_control: control, cid = 1, Ns = 4, Nr = 12
> > Feb 22 13:07:02 yxdgate xl2tpd[12929]: control_finish: Connection closed to 24.xx.xx.xx, serial 0 ()
> > Feb 22 13:07:02 yxdgate xl2tpd[12929]: Lingering pppd: sending SIGTERM to pid 14048.
> > Feb 22 13:07:02 yxdgate xl2tpd[12929]: Lingering pppd: sent SIGTERM, running waitpid.

Excellent! This confirms that pppd indeed does not die after getting the SIGTERM, because
your debug would have shown "Lingering pppd: waitpid returned, pppd died" after this
message.

> Any other logging I should enable/look for?

Nope, just edit call.c and change line 442 from:

kill (pid, SIGTERM);

to

kill (pid, SIGKILL);

and that should fix your problem.

> PPP version is ppp-2.4.2-6.4.RHEL4 running on CentOS 4.

I'll file a bug report with them. Thanks!

Thanks to Tuomo for spotting this one. I'll release a new version with the debug
code removed and this change incorporated.

Paul
-- 
Building and integrating Virtual Private Networks with Openswan:
http://www.amazon.com/gp/product/1904811256/104-3099591-2946327?n=283155