Thank you! I may one day follow-up on persistent connections, but this does this trick for now.
On Wed, Mar 15, 2017 at 3:44 AM, Willy Tarreau <w...@1wt.eu> wrote: > On Mon, Mar 13, 2017 at 06:10:23PM +0100, Willy Tarreau wrote: > > > Just wanted to follow up. I've been running this patch for a couple > days on > > > an idle system and haven't noticed any problems. > > > Could this be merged? Is there anything else I can test? > > > > I'm personally fine with it but I'd rather have Emeric approve it, as > > he knows better than me the possible impacts of shutting down cleanly > > or not on SSL. > > > > Emeric, I've re-attached the patch. Using conn_data_shutw() instead of > > conn_data_shutw_hard() causes the "clean" flag to be set when calling > > ssl_sock_shutw() and SSL_set_quiet_shutdown() not to be called so that > > we perform a clean shutdown. The purpose is to give a chance to the > > server to store the context and avoid renegociating. > > Now applied with Emeric's blessing. > > Thanks, > Willy >