Re: [strongSwan] IPTables NAT Query

2015-04-07 Thread Tormod Macleod
Yeah, that was my mistake. I only need the second one. I thought I needed the first one because IPTables Postrouting didn't take place until after the kernel had determined whether the traffic was to be tunneled but I was mistaken. I see now that Postrouting takes place before the kernel determin

Re: [strongSwan] IPTables NAT Query

2015-04-07 Thread Noel Kuntze
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Hello Tormod, Okay. But still only one CHILD_SA is up and the other one not. Mit freundlichen Grüßen/Kind Regards, Noel Kuntze GPG Key ID: 0x63EC6658 Fingerprint: 23CA BB60 2146 05E7 7278 6592 3839 298F 63EC 6658 Am 07.04.2015 um 13:35 schrieb T

Re: [strongSwan] IPTables NAT Query

2015-04-07 Thread Tormod Macleod
Hi Noel, I need the SNAT as the network on the right want to see the traffic originate from the 1.1.1.0/24 range for internal routing purposes. I thought (Bryan Duff set me straight) I needed two Child SAs. Because the right device is a Cisco device I had to configure two separate Child SAs rat

Re: [strongSwan] IPTables NAT Query

2015-04-07 Thread Noel Kuntze
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Hello Mirko, Yes, I meant that one. It seems I forgot to put the reference in the email. Mit freundlichen Grüßen/Kind Regards, Noel Kuntze GPG Key ID: 0x63EC6658 Fingerprint: 23CA BB60 2146 05E7 7278 6592 3839 298F 63EC 6658 Am 07.04.2015 um 13:

Re: [strongSwan] IPTables NAT Query

2015-04-07 Thread Mirko Parthey
On Mon, Apr 06, 2015 at 07:01:42PM +0200, Noel Kuntze wrote: > There is a graph[1] that describes the path of the traffic in the kernel. Did you mean this one? [1] http://inai.de/images/nf-packet-flow.svg Regards Mirko ___ Users mailing list Users@lists

Re: [strongSwan] IPTables NAT Query

2015-04-06 Thread Noel Kuntze
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Hello Tormod, There is a graph[1] that describes the path of the traffic in the kernel. Why do you believe, that you have to apply SNAT/MASQUERADE? By the way, your tunnel setup is wrong. You define two IPsec tunnels, but there is only one being u

Re: [strongSwan] IPTables NAT Query

2015-04-06 Thread Bryan Duff
If I recall correctly your step 5 is where things matter - make sure at that point (basically after nat POSTROUTING) that the traffic source/dest matches your left/rightsubnets. Don't worry about the routing decision (your step 3). You should only need one DNAT and one SNAT for your traffic. -Br

[strongSwan] IPTables NAT Query

2015-04-06 Thread Tormod Macleod
Hello, I'm currently testing a site to site VPN. I need to change both the source and destination address on the left device before forwarding the packets over the VPN to the right device. I believe it all happens in the order below but I may be wrong. 1 IPTables Prerouting 2 Route selected