Hello, For a testing purpose I've replaced the ha_socket.c implementation by a corosync implementation (using the cpg library, see http://en.wikipedia.org/wiki/Corosync_%28project%29 for more information, and send/recv queues) So far, it seems to solve all the issues I got (tested on a 7K tunnels configuration, with resync) but of course it will need a lot more testing.
The big advantage of corosync is that it can multicast messages to several nodes: it is not limited to a single remote node. The problem is that I am not sure about the reorganization of the HA plugin to properly handle this. Maybe we would have to remove the ha_socket direct references and introduce a new communication layer? Regards, Emeric ----- Mail original ----- De: "Martin Willi" <[email protected]> À: "Emeric POUPON" <[email protected]> Cc: [email protected] Envoyé: Mardi 9 Septembre 2014 14:18:27 Objet: Re: [strongSwan-dev] HA resync issue Hi Emeric, > Do you have some news or new ideas about this issue? > > Would you like that I fill in a bug report about it? Feel free to do so. However, please don't expect an immediate fix from the strongSwan core developers for this non-trivial issue. We have limited resources for non-funded projects, and a long list of feature and bugfix requests from our community. Regards Martin _______________________________________________ Dev mailing list [email protected] https://lists.strongswan.org/mailman/listinfo/dev
