Date: Wed, 23 Apr 2008 15:48:48 +0100 (BST) From: Robert Watson <rwatson@FreeBSD.org> To: Dmitriy Kirhlarov <dimma@higis.ru> Cc: freebsd-current@freebsd.org Subject: Re: Http Accept filters (accf_http) Message-ID: <20080423154626.F64388@fledge.watson.org> In-Reply-To: <480EE8B2.2020907@higis.ru> References: <8481.1208889581@critter.freebsd.dk> <480E3E66.3000303@samsco.org> <fuli49$pa3$1@ger.gmane.org> <480E589C.8010108@delphij.net> <480EE8B2.2020907@higis.ru>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 23 Apr 2008, Dmitriy Kirhlarov wrote: > Xin LI wrote: >> | Does anyone know why accf_accept is disabled by default in the ports' >> | stock Apache 2.2 (it's disabled in the default config files)? I thought >> >> Yes, I wonder that too. Personally I think it might be a good idea to just >> bind accf_http and accf_data into GENERIC, as I always add them > > We are using accf_http and accf_data on 6.1-6.3 with apache and proftpd. Now > we have periodical sockets leak. We can't reproduce it, but accf_* under > suspiction. If you kill apache and wait a few seconds, does the socket count go back down to normal? Shutting down to single user mode and comparing allocated memory with vmstat -z and vmstat -m against similar measurements made before going up to multi-user (ignoring file system stuff) might well shed some light. I'm aware of a few problems relating to accept filters, and "fixing" them has been on my todo list for several years. Unfortunately, other things keep getting ahead of that in the todo list. One known issue is that accept filters aren't entirely happy with the new multi-processor locking world order. Robert N M Watson Computer Laboratory University of Cambridge
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20080423154626.F64388>