[strongSwan] [Strongswan-5.3.0] - Ikev2 fragmentation Question

2016-07-28 Thread Sriram
Hi, We are using strongswan - 5.3.0. To make use of the ikev2 fragmentation feature that is available since 5.2.1, we enabled fragmentation=yes in ipsec.conf and fragment_size = 1200. The device mtu is 1500. Feature gets enabled as the security gateway also supports IKEV2_FRAGMENTATION_SUPPORTED p

Re: [strongSwan] [Strongswan-5.3.0] - Ikev2 fragmentation Question

2016-07-29 Thread Tobias Brunner
Hi Sriram, > But the concern is fragment size, though it is set as 1200, > fragment_size of 576 is seen in the wireshark. I'm assuming for packets sent by the gateway. The fragment size is not negotiated, so the gateway might just default to the minimum datagram size a host must be able to accep

Re: [strongSwan] [Strongswan-5.3.0] - Ikev2 fragmentation Question

2016-07-29 Thread Sriram
Hi Tobias, Thanks for the reply. I have set fragment_size = 1200 in strongswan.conf and fragmentation=yes in the ipsec.conf in the client side . Even though it is 1200, ike packets that are sent from the client are of the size 576. I have not changed the configuration file, as the generation of t

Re: [strongSwan] [Strongswan-5.3.0] - Ikev2 fragmentation Question

2016-07-29 Thread Tobias Brunner
Hi Sriram, > So I think, since the strongswan file is not proper, charon would have > defaulted to 576. Please clarify. Yes, if the file is invalid it gets rejected completely and no options in it will get applied. You should have seen an error message like "invalid config file '...'" in the log

Re: [strongSwan] [Strongswan-5.3.0] - Ikev2 fragmentation Question

2016-07-29 Thread Sriram
Hi Tobias. Yes you are right, strongswan complains about this. Since logs got rolled over and there is no console access, I could not catch this issue soon. Thanks for the help. Regards, Sriram. On Fri, Jul 29, 2016 at 6:32 PM, Tobias Brunner wrote: > Hi Sriram, > > > So I think, since the st