From owner-freebsd-hackers Fri Jun 6 10:21:02 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.5/8.8.5) id KAA00359 for hackers-outgoing; Fri, 6 Jun 1997 10:21:02 -0700 (PDT) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.50]) by hub.freebsd.org (8.8.5/8.8.5) with SMTP id KAA00353 for ; Fri, 6 Jun 1997 10:20:58 -0700 (PDT) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id KAA22861; Fri, 6 Jun 1997 10:15:18 -0700 From: Terry Lambert Message-Id: <199706061715.KAA22861@phaeton.artisoft.com> Subject: Re: Extremely poor interactive response under heave SCSI load To: jkh@time.cdrom.com (Jordan K. Hubbard) Date: Fri, 6 Jun 1997 10:15:18 -0700 (MST) Cc: henrich@crh.cl.msu.edu, freebsd-hackers@FreeBSD.ORG In-Reply-To: <18707.865573190@time.cdrom.com> from "Jordan K. Hubbard" at Jun 5, 97 09:59:50 pm X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-hackers@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk > > There has to be some solution, forcing the scsi command queue to search for > > alternate commands every so often or something.. It just sucks :) > > We await your diffs, Mr Henrich. :) I think the problem is the monopolization of the buffer pool by a single file. At one point in time, I had submitted diffs to (1) track the number of pages on a per vnode basis, and (2) recycle pages from the same vnode (LRU) instead of the system pool when the count hit some "quota". You should look at the -current list archives from March-May of 1996. I implemented the same thing for UnixWare at one time... they have a similar problem triggered by their ld, which mmap's object files and bounces all over in them (they didn't integrate the fix, either). Regards, Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.