Hi, On Wed, Oct 03, 2012 at 10:15:32PM +0200, Andre Oppermann wrote: > On 03.10.2012 22:03, Adrian Chadd wrote: > > > > somaxconn is the connection queue depth. If it's sitting at a couple > > hundred thousand then something else is going crazily wrong. > > > > I understand your frustration, but there's a lot of instances where > > the application just isn't doing things "right" and the OS tries to > > hide it as much as psosible. Blowing out somaxconn to chew up a whole > > lot of resources seems a bit silly. I'd rather investigate why the > > userland application is not servicing the connect queue often enough. > > > > I've written network services that supported tens of thousands of new > > TCP connections a second on a LAN and I never once had to bump > > somaxconn past 32767. I'm not saying that it won't apply to your > > scenario, I'm just trying to explain that there's likely more going > > on. > > I guess the problem is rather kern.ipc.maxsockets which is only 25600. > > The name somaxconn is confusing as it specifies the listen queue limit > instead of the maximum number of connections as the it suggests. If we want to change that name to something more sensible and less error-prone like "somaxbacklog", does the project has a policy to change sysctl names? I'm thinking of something like renaming the sysctl to "somaxbacklog" and make "somaxconn" compatibility shim during RELENG_10 which still works but prints a warning in the dmesg. -- Jeremie Le Hen Scientists say the world is made up of Protons, Neutrons and Electrons. They forgot to mention Morons.Received on Mon Oct 08 2012 - 08:49:46 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:31 UTC