From owner-freebsd-hackers Sat Oct 26 11:26:42 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id LAA19684 for hackers-outgoing; Sat, 26 Oct 1996 11:26:42 -0700 (PDT) Received: from critter.tfs.com (disn1.cybercity.dk [194.16.57.1]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id LAA19658; Sat, 26 Oct 1996 11:26:32 -0700 (PDT) Received: from critter.tfs.com (localhost.tfs.com [127.0.0.1]) by critter.tfs.com (8.7.5/8.7.3) with ESMTP id UAA17439; Sat, 26 Oct 1996 20:26:57 +0200 (MET DST) To: Joe Greco cc: jgreco@ns.sol.net, hackers@FreeBSD.org Subject: Re: HELP! :-( Hitting datasize limit In-reply-to: Your message of "Sat, 26 Oct 1996 12:31:19 CDT." <199610261731.MAA08279@solaria.sol.net> Date: Sat, 26 Oct 1996 20:26:57 +0200 Message-ID: <17437.846354417@critter.tfs.com> From: Poul-Henning Kamp Sender: owner-hackers@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk In message <199610261731.MAA08279@solaria.sol.net>, Joe Greco writes: >> You may want to >> >> ln -sf H /etc/malloc.conf >> >> I've not gotten a complete picture, but people indicate that it helps >> a little bit. > >Hi Poul, > >I did not mean to imply that I think it has ANYTHING to do with >phkmalloc... the growth in size is EXPECTED... I know, I just wanted to help you :-) >innd allocates per-channel receive buffers. When the maximum allowed >size for a Usenet article is 1MB and you have fifty inbound NNTP channels, >it is reasonable to expect that the receive buffer for each channel will >eventually need to be grown close to 1MB to allow for the large >alt.binaries articles that everyone shovels around. Hmm, does it realloc them back down again ? If it does then the 'H' trick should work even better, what happens is that it calls madvise(MADV_FREE) on free(3)'ed pages. -- Poul-Henning Kamp | phk@FreeBSD.ORG FreeBSD Core-team. http://www.freebsd.org/~phk | phk@login.dknet.dk Private mailbox. whois: [PHK] | phk@ref.tfs.com TRW Financial Systems, Inc. Future will arrive by its own means, progress not so.