[Openswan Users] Facing a problem with updown script

[Anirudh Kamatgi]

On Tue, Nov 18, 2008 at 10:09 PM, Paul Wouters <paul at xelerance.com> wrote:

> On Tue, 18 Nov 2008, Anirudh Kamatgi wrote:
>
> > I have a custom updown script written in perl.
> > Whenever this script gets called with 'down-client', I want to parse the
> > output of 'ipsec auto --status' inside this script.
> > I am assigning the output to a perl array( my @status = `ipsec auto
> > --status`; ).
> > However, the script hangs at this point.
>
> Does the script rely on DNS at the other end of a tunnle that is down?

No, I am using static IP private addresses, it is a gateway-gateway tunnel
and the gateways are connected by a crossover cable.

>
>
> > I can see 2 shell processes executing the <path to ipsec>/ipsec/auto
> command
> > and one whack process.
> > After I kill the whack process, things get back to normal.
>
> add plutodebug=controlmore and see if the logs say more.

I did not get any extra information after adding the above parameter.
I did notice one interesting  thing.

The problem happens when I bring down a tunnel administratively with the
'ipsec auto --down tunnel-name' command.
This results in the updown script getting called which executes the command
'ipsec auto --status' and hangs.

If I see the processes which are running at this point, I see 2 whack
processes, with the arguments '--name tunnel-name --terminate' and
'--status' respectively.
If I terminate( i.e send SIGINT ) the '--down' command which has hung, the
first whack process is terminated, but the second whack process( the one
with argument '--status' ) is still not terminated.
After this, all ipsec commands hang until I kill the original whack process
which had not terminated.
Is this a reentrancy issue?
Paul,
Do you need any more information like the barf output?

>
>
> Paul
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.openswan.org/pipermail/users/attachments/20081119/74805555/attachment-0001.html