On Thu, Mar 23, 2017 at 09:35:28PM +0100, Cyril Bonté wrote: > Hi Olivier and Willy, > > Le 23/03/2017 à 14:50, Willy Tarreau a écrit : > > Hi Olivier, > > > > On Thu, Mar 23, 2017 at 02:17:38PM +0100, Olivier Doucet wrote: > > > I was also thinking about dumping how many connections were force-killed > > > in > > > log, if it's easy. > > > > I thought about it as well while responding but forgot to mention it. That's > > also why I'd prefer to leave one extra second for the connections to cleanly > > close and log. Yes it's important. > > Of course I can add some logs. Currently, when haproxy enters "soft-stop", > it sends one log line for each proxy, indicating it's stopping : > [WARNING] 081/213056 (19942) : Stopping proxy test in 0 ms. > [WARNING] 081/213056 (19942) : Proxy test stopped (FE: 500 conns, BE: 0 > conns). > > I think we can add a new log line for each proxy having opened connections, > where the number of remaining connections is read from "feconn": > [WARNING] 081/213101 (19942) : Proxy test hard-stopped (500 remaining conns > will be closed). > > or do you prefer a global counter ?
I didn't think about the per-proxy approach but I think that emitting a log only for those having got killed connections would be nice given that most of the time there will be very few such proxies. But I think it can make your patch more complicated, especially if we target simplicity for the backport. So maybe in fact having only the global counter first is a good first step and if we can later improve on this we'll see if it's worth backporting. Thanks! Willy
