From owner-freebsd-hackers Thu Mar 7 8:12:29 2002 Delivered-To: freebsd-hackers@freebsd.org Received: from dire.bris.ac.uk (dire.bris.ac.uk [137.222.10.60]) by hub.freebsd.org (Postfix) with ESMTP id 3D72237B402 for ; Thu, 7 Mar 2002 08:12:24 -0800 (PST) Received: from mail.ilrt.bris.ac.uk by dire.bris.ac.uk with SMTP-PRIV with ESMTP; Thu, 7 Mar 2002 16:12:14 +0000 Received: from cmjg (helo=localhost) by mail.ilrt.bris.ac.uk with local-esmtp (Exim 3.16 #1) id 16j0Sr-0000iq-00; Thu, 07 Mar 2002 16:09:57 +0000 Date: Thu, 7 Mar 2002 16:09:57 +0000 (GMT) From: Jan Grant X-X-Sender: cmjg@mail.ilrt.bris.ac.uk To: freebsd-hackers@freebsd.org Subject: Bug? still looking, yet to knock up small test case. Suggestions solicited. Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Something odd seems to be happening; I'd appreciate "look here" suggestions. I suspect mmapped pages aren't being flushed but gawd alone knows why. Situation: vmware2, with a "fake disk", files in the /external FS (/external/vmware1/nt1.*). FBSD-stable. Behaviour's been like this for quite some time (since around 4.3-release, I think) Symptoms: run vmware, launch NT in virtual machine, blah, blah. Suspend the virtual machine. Former behaviour went like this: reports "synchronising disk status" (or imilar) and flushes memory contents to disk. (Irritatingly, one time in 20, this still happens.) Behaviour these days: vmware suspends immediately, and exits normally. However, the contents of the nt1.* files _aren't_ flushed, and stay unflushed for an open-ended period of time (days, weeks). sync(2)s don't make any change on this; umounting the FS will take several minutes as the contents are flushed to disk. fsyncing the appropriate files causes them to be flushed. Shouldn't the file contents be getting written back out within some finite time period? The system's not under tremendous load (it's a desktop workstn most of the time). Note: this doesn't result in ultimately broken behaviour, because those pages are eventually flushed on orderly shutdown. I'm just curious as to why sync(2) isn't forcing this. Cheers, jan -- jan grant, ILRT, University of Bristol. http://www.ilrt.bris.ac.uk/ Tel +44(0)117 9287088 Fax +44 (0)117 9287112 RFC822 jan.grant@bris.ac.uk Ceci n'est pas une pipe | To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message