From owner-freebsd-isp Sat Oct 25 09:23:22 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id JAA09325 for isp-outgoing; Sat, 25 Oct 1997 09:23:22 -0700 (PDT) (envelope-from owner-freebsd-isp) Received: from burka.carrier.kiev.ua (root@burka.carrier.kiev.ua [193.193.193.107]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id JAA09302 for ; Sat, 25 Oct 1997 09:22:52 -0700 (PDT) (envelope-from archer@grape.carrier.kiev.ua) Received: from sivka.carrier.kiev.ua (root@sivka.carrier.kiev.ua [193.193.193.101]) by burka.carrier.kiev.ua (8.8.6/8.8.6) with ESMTP id TAA19072 for ; Sat, 25 Oct 1997 19:22:30 +0300 (EEST) Received: (from uucp@localhost) by sivka.carrier.kiev.ua (8.8.7/8.8.7) with UUCP id TAA27998 for isp@freebsd.org; Sat, 25 Oct 1997 19:15:58 +0300 (EEST) Received: (from archer@localhost) by grape.carrier.kiev.ua (8.8.7/8.8.7) id TAA11494; Sat, 25 Oct 1997 19:06:12 +0300 (EEST) Date: Sat, 25 Oct 1997 19:06:12 +0300 (EEST) From: Alexander Litvin Message-Id: <199710251606.TAA11494@grape.carrier.kiev.ua> To: isp@freebsd.org Subject: Re: Squid 1.1.16 breaking under load In-Reply-To: <199710230745.RAA26944@spooky.eis.net.au> <344F7413.6CCA@terranova.net> X-Newsreader: TIN [UNIX 1.3 unoff BETA 970424; i386 FreeBSD 2.2.5-971010-BETA] Sender: owner-freebsd-isp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk In article <344F7413.6CCA@terranova.net> you wrote: > Ernie Elu wrote: > > > > Anyone had squid 1.1.16 break under load? > > I am keen to know if anyone else has had the same problem, or is it > > just unique to my server in which case I will set up a fresh drive > > and swap it out. > Yes, it just started happening after I upgraded to 1.NOVM.16 and hasn't > happened yet in the last 36 hours since I upgraded to 1.NOVM.17. > I think it was a problem they found and fixed: > Changes to squid-1.1.17 (October 19, 1997): > - Fixed ftpget null pointer coredump for mime encodings. > - Fixed setuid() bug when creating swap directories. Now > change to 'effective_user' before making dirs. > - Fixed cachemgr.cgi to give 'unknown host' message if > gethostbyname() fails. > - Fixed extra disk read() bug for swapins. Now a disk > read calls back immediately if any data was read instead > of trying to fill the 8k buffer. > - Fixed asciiHandleConn(). The 'while (1)' loop broke things > badly under very high loads. Limit to 10 times through > the loop, or until RESERVED_FD limit is reached. > I'd suggest upgrading to .17 and see if the problem ceases. No chance, it seems. I mean, I've upgraded to 1.NOVM.17, and it still dies sometimes with 6, cache.log reports "Sombody is putting NULL pointer". IMHO it has something to do with squid internal memory management. Haven't had time to track it down yet. > -T > -- > TerraNovaNet Internet Services - Key Largo, FL > Voice: (305)453-4011 > Fax: (305)451-5991 > http://www.TerraNova.net > ---------------------------------------------- > "Acceptance without proof is the fundamental characteristic of Western > religion, Rejection without proof is the fundamental characteristic of > Western science." > -- Gary Zukav, "The Dancing Wu Li Masters" -- Litvin Alexander