[Openswan Users] 'No such process' And 'Resource temporarily unavailable'

[Paul Wouters]

On Fri, 1 May 2009, Faizan Syed wrote:

> I am experiencing some issue here,
>  
> On Centos 4.3 (Openswan 2.4.18) when vpn link is down and try to ping
> immediately this error occurs
> connect: Resource temporarily unavailable

I believe this is the bogus -EAGAIN error that Dave Miller decided
to return, breaking POSIX. This related to policies being loaded
into the kernel for which no IPsec SA s established, so a signal is
sent to userland to try and establish a new tunnel.

Patches were merged into 2.4.x at some point that should have addressed
this for NETKEY. On KLIPS it was never a problem, as it handles packet
caching when tunnels are temporarilly unavailable correctly.

> But on Centos 5 (Openswan 2.4 and 2.6 tried both) whne vpn is down it
> give this error but after waiting 30secs when try to ping dead peer..
> connect: No such process

Can you confirm that this is a kernel issue and not a openswan version
issue by trying the same 2.4.18 openswan on the centos 5 kernel?

Paul