From owner-freebsd-fs Sun Mar 5 11:12:45 2000 Delivered-To: freebsd-fs@freebsd.org Received: from cs.columbia.edu (cs.columbia.edu [128.59.16.20]) by hub.freebsd.org (Postfix) with ESMTP id A437537BACF; Sun, 5 Mar 2000 11:12:36 -0800 (PST) (envelope-from ezk@shekel.mcl.cs.columbia.edu) Received: from shekel.mcl.cs.columbia.edu (shekel.mcl.cs.columbia.edu [128.59.18.15]) by cs.columbia.edu (8.9.1/8.9.1) with ESMTP id OAA04279; Sun, 5 Mar 2000 14:12:31 -0500 (EST) Received: (from ezk@localhost) by shekel.mcl.cs.columbia.edu (8.9.1/8.9.1) id OAA01726; Sun, 5 Mar 2000 14:12:30 -0500 (EST) Date: Sun, 5 Mar 2000 14:12:30 -0500 (EST) Message-Id: <200003051912.OAA01726@shekel.mcl.cs.columbia.edu> From: Erez Zadok To: Kirk McKusick Cc: Alfred Perlstein , Terry Lambert , fs@FreeBSD.ORG, jkh@FreeBSD.ORG, Bruce Evans Subject: Re: changing mount options still can cause damage? In-reply-to: Your message of "Sat, 04 Mar 2000 11:34:00 PST." <200003041934.LAA16343@flamingo.McKusick.COM> Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In message <200003041934.LAA16343@flamingo.McKusick.COM>, Kirk McKusick writes: [...] > just as it would for a file open for writing). This seems > a slightly odd semantic for a file open for reading, so I > have not done it. Does anyone have any views on whether the > filesystem should be changed in this way on forcible > downgrades? > > Kirk Traditionally Unix file systems had all kind of odd semantics, but were considered reliable and stable. So *if* I had to choose b/t adding some odd semantics and keeping the f/s clean, I'd go for cleanliness. If, however, there is no clear winning option, then the next best thing would be to do one of the following: (1) add a new mount flag that can select among the choices. This option could be added to the mount(2) when it's doing a remount. We can select some default behavior for the flag, and let users turn it off if they want. (2) print some console message from there kernel upon remount, warning of any known open+unlinked files. BTW, can one remount read-only a f/s that's already remounted read-only? One (true hack) might be to keep the current remount semantics the same, but change them upon a second read-only remount. Erez. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Sun Mar 5 21:47:42 2000 Delivered-To: freebsd-fs@freebsd.org Received: from mail.webmonster.de (datasink.webmonster.de [194.162.162.209]) by hub.freebsd.org (Postfix) with SMTP id EB4B037BC9B for ; Sun, 5 Mar 2000 21:47:38 -0800 (PST) (envelope-from karsten@rohrbach.de) Received: (qmail 61656 invoked by uid 1000); 6 Mar 2000 05:47:34 -0000 Date: Mon, 6 Mar 2000 06:47:34 +0100 From: "Karsten W. Rohrbach" To: Mattias Pantzare Cc: Edwin Mons , =?iso-8859-1?Q?Administra=E7=E3o_MailBR?= , freebsd-fs@FreeBSD.ORG Subject: Re: NFS: Linux Client and BSD Server Message-ID: <20000306064734.A61425@rohrbach.de> Reply-To: karsten@rohrbach.de References: <200001201602.RAA13433@queeg.ludd.luth.se> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0i In-Reply-To: <200001201602.RAA13433@queeg.ludd.luth.se>; from pantzer@ludd.luth.se on Thu, Jan 20, 2000 at 05:02:25PM +0100 X-Arbitrary-Number-Of-The-Day: 42 X-Sender: karsten@rohrbach.de Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Mattias Pantzare(pantzer@ludd.luth.se)@Thu, Jan 20, 2000 at 05:02:25PM +0100: > > When I did this while using an NFS (BSD <-> BSD) the speed went up to 1 > > MB/sec (on 10baseT). Theoretical maximum for 10 Mbit ethernet :-) > > And break the NFS protocol. Pray that the NFS server won't crash if you have > that on. what does vfs.nfs.async do, then, that it breaks the protocol? /k -- > Did you know that there are 71.9 acres of nipple tissue in the U.S.? http://www.webmonster.de http://www.apache.de http://www.splatterworld.de (NIC-HDL KR433/KR11-RIPE) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Sun Mar 5 22:39: 8 2000 Delivered-To: freebsd-fs@freebsd.org Received: from gatekeeper.tsc.tdk.com (gatekeeper.tsc.tdk.com [207.113.159.21]) by hub.freebsd.org (Postfix) with ESMTP id DCB2337BC15; Sun, 5 Mar 2000 22:39:05 -0800 (PST) (envelope-from gdonl@tsc.tdk.com) Received: from imap.gv.tsc.tdk.com (imap.gv.tsc.tdk.com [192.168.241.198]) by gatekeeper.tsc.tdk.com (8.8.8/8.8.8) with ESMTP id WAA02165; Sun, 5 Mar 2000 22:38:50 -0800 (PST) (envelope-from gdonl@tsc.tdk.com) Received: from salsa.gv.tsc.tdk.com (salsa.gv.tsc.tdk.com [192.168.241.194]) by imap.gv.tsc.tdk.com (8.9.3/8.9.3) with ESMTP id WAA41497; Sun, 5 Mar 2000 22:38:49 -0800 (PST) (envelope-from Don.Lewis@tsc.tdk.com) Received: (from gdonl@localhost) by salsa.gv.tsc.tdk.com (8.8.5/8.8.5) id WAA16092; Sun, 5 Mar 2000 22:38:40 -0800 (PST) From: Don Lewis Message-Id: <200003060638.WAA16092@salsa.gv.tsc.tdk.com> Date: Sun, 5 Mar 2000 22:38:37 -0800 In-Reply-To: <200003041934.LAA16343@flamingo.McKusick.COM> References: <200003041934.LAA16343@flamingo.McKusick.COM> X-Mailer: Mail User's Shell (7.2.6 beta(5) 10/07/98) To: Kirk McKusick , Alfred Perlstein Subject: Re: changing mount options still can cause damage? Cc: Terry Lambert , fs@FreeBSD.ORG, jkh@FreeBSD.ORG, Bruce Evans Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Mar 4, 11:34am, Kirk McKusick wrote: } Subject: Re: changing mount options still can cause damage? } 2) In reviewing my bug logs for FFS I have found the `corruption' } case to which I believe the bug entry in the manual page was } alluding. It is possible to get lost inodes in a filesystem } that has been downgraded to read-only even if it never ran } in async mode. The senario causing trouble goes as follows: } } A) a process opens a file for reading. } B) the file is unlinked } C) the filesystem is downgraded to read-only } D) the process referencing the now unlinked } file exits or closes the file. } } In this case, the the inode cannot be freed as the filesystem } is now in read-only mode. Corruption of this sort is not } particularly threatening as the lost inodes will be cleaned } up the next time `fsck -p' is run, but the resulting loss } of space may be annoying if the filesystem is nearly full. } The alternative is to vgone files with link counts of zero } when doing a (forcible) downgrade just as we do with files } that are open for writing. This would result in the inode } being released and the process seeing a dead file (again } just as it would for a file open for writing). This seems } a slightly odd semantic for a file open for reading, so I } have not done it. Does anyone have any views on whether the } filesystem should be changed in this way on forcible } downgrades? This is probably OK when doing a forceable downgrade. A non-forceable downgrade should probably fail with EBUSY. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Mon Mar 6 1:25:47 2000 Delivered-To: freebsd-fs@freebsd.org Received: from anchor-post-31.mail.demon.net (anchor-post-31.mail.demon.net [194.217.242.89]) by hub.freebsd.org (Postfix) with ESMTP id 53F1B37BCCA for ; Mon, 6 Mar 2000 01:25:37 -0800 (PST) (envelope-from dfr@nlsystems.com) Received: from nlsys.demon.co.uk ([158.152.125.33] helo=herring.nlsystems.com) by anchor-post-31.mail.demon.net with esmtp (Exim 2.12 #1) id 12Rtlf-0003oX-0V; Mon, 6 Mar 2000 09:25:35 +0000 Received: from salmon.nlsystems.com (salmon.nlsystems.com [10.0.0.3]) by herring.nlsystems.com (8.9.3/8.8.8) with ESMTP id JAA30053; Mon, 6 Mar 2000 09:31:09 GMT (envelope-from dfr@nlsystems.com) Date: Mon, 6 Mar 2000 09:28:11 +0000 (GMT) From: Doug Rabson To: "Karsten W. Rohrbach" Cc: Mattias Pantzare , Edwin Mons , =?iso-8859-1?Q?Administra=E7=E3o_MailBR?= , freebsd-fs@freebsd.org Subject: Re: NFS: Linux Client and BSD Server In-Reply-To: <20000306064734.A61425@rohrbach.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Mon, 6 Mar 2000, Karsten W. Rohrbach wrote: > Mattias Pantzare(pantzer@ludd.luth.se)@Thu, Jan 20, 2000 at 05:02:25PM +0100: > > > When I did this while using an NFS (BSD <-> BSD) the speed went up to 1 > > > MB/sec (on 10baseT). Theoretical maximum for 10 Mbit ethernet :-) > > > > And break the NFS protocol. Pray that the NFS server won't crash if you have > > that on. > > what does vfs.nfs.async do, then, that it breaks the protocol? It replies to the RPC before the bits have actually hit stable storage. If the server crashes or loses power before it syncs, the client has no way of knowing that its data is lost. With NFSv3, there is a change to the protocol which allows most of the performance of async but still allows the client to re-send its changes after the server reboots. -- Doug Rabson Mail: dfr@nlsystems.com Nonlinear Systems Ltd. Phone: +44 181 442 9037 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Mon Mar 6 6:18:37 2000 Delivered-To: freebsd-fs@freebsd.org Received: from acutiator.nacamar.de (mail.nacamar.de [194.162.162.200]) by hub.freebsd.org (Postfix) with ESMTP id 7294737BD67 for ; Mon, 6 Mar 2000 06:18:26 -0800 (PST) (envelope-from karsten@rohrbach.de) Received: from hardcore (dchp13.nacamar.de [195.63.63.246]) by acutiator.nacamar.de (Postfix) with SMTP id 126CD5D6F; Mon, 6 Mar 2000 15:17:43 +0100 (CET) Message-ID: <009701bf8776$b93963e0$f63f3fc3@hardcore> From: "Karsten W. Rohrbach" To: "Doug Rabson" Cc: "Mattias Pantzare" , "Edwin Mons" , =?iso-8859-1?Q?Administra=E7=E3o_MailBR?= , References: Subject: Re: NFS: Linux Client and BSD Server Date: Mon, 6 Mar 2000 15:02:52 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2919.6600 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600 Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org so, if you mount -o nfsv3 and have vfs.nfs.async=1 it wont fail or does this mean that one has not to set vfs.nfs.async=1 ? /k -- The path of excess leads to the tower of wisdom. -W. Blake Karsten W. Rohrbach - Senior Research Engineer - Nacamar Data Communications one world. one net. nacamar - http://www.nacamar.net This was written on the road... Answers to your reply could be delayed. ----- Original Message ----- From: "Doug Rabson" To: "Karsten W. Rohrbach" Cc: "Mattias Pantzare" ; "Edwin Mons" ; "Administração MailBR" ; Sent: Monday, March 06, 2000 10:28 AM Subject: Re: NFS: Linux Client and BSD Server > On Mon, 6 Mar 2000, Karsten W. Rohrbach wrote: > > > Mattias Pantzare(pantzer@ludd.luth.se)@Thu, Jan 20, 2000 at 05:02:25PM +0100: > > > > When I did this while using an NFS (BSD <-> BSD) the speed went up to 1 > > > > MB/sec (on 10baseT). Theoretical maximum for 10 Mbit ethernet :-) > > > > > > And break the NFS protocol. Pray that the NFS server won't crash if you have > > > that on. > > > > what does vfs.nfs.async do, then, that it breaks the protocol? > > It replies to the RPC before the bits have actually hit stable storage. If > the server crashes or loses power before it syncs, the client has no way > of knowing that its data is lost. > > With NFSv3, there is a change to the protocol which allows most of the > performance of async but still allows the client to re-send its changes > after the server reboots. > > -- > Doug Rabson Mail: dfr@nlsystems.com > Nonlinear Systems Ltd. Phone: +44 181 442 9037 > > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Mon Mar 6 15:20:26 2000 Delivered-To: freebsd-fs@freebsd.org Received: from knecht.Sendmail.ORG (knecht.sendmail.org [209.31.233.176]) by hub.freebsd.org (Postfix) with ESMTP id 9077737C08B; Mon, 6 Mar 2000 15:20:23 -0800 (PST) (envelope-from mckusick@flamingo.McKusick.COM) Received: from flamingo.McKusick.COM (root@flamingo.mckusick.com [209.31.233.178]) by knecht.Sendmail.ORG (8.10.0.Beta10/8.10.0.Beta10) with ESMTP id e26NKB207062; Mon, 6 Mar 2000 15:20:12 -0800 (PST) Received: from flamingo.McKusick.COM (mckusick@localhost [127.0.0.1]) by flamingo.McKusick.COM (8.9.3/8.9.0) with ESMTP id PAA16884; Mon, 6 Mar 2000 15:14:52 -0800 (PST) Message-Id: <200003062314.PAA16884@flamingo.McKusick.COM> To: Erez Zadok Subject: Re: changing mount options still can cause damage? Cc: Alfred Perlstein , Terry Lambert , fs@freebsd.org, jkh@freebsd.org, Bruce Evans In-reply-to: Your message of "Sun, 05 Mar 2000 14:12:30 EST." <200003051912.OAA01726@shekel.mcl.cs.columbia.edu> Date: Mon, 06 Mar 2000 15:14:52 -0800 From: Kirk McKusick Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Date: Sun, 5 Mar 2000 14:12:30 -0500 (EST) From: Erez Zadok To: Kirk McKusick Cc: Alfred Perlstein , Terry Lambert , fs@FreeBSD.ORG, jkh@FreeBSD.ORG, Bruce Evans Subject: Re: changing mount options still can cause damage? In-reply-to: Your message of "Sat, 04 Mar 2000 11:34:00 PST." <200003041934.LAA16343@flamingo.McKusick.COM> In message <200003041934.LAA16343@flamingo.McKusick.COM>, Kirk McKusick writes: [...] > just as it would for a file open for writing). This seems > a slightly odd semantic for a file open for reading, so I > have not done it. Does anyone have any views on whether the > filesystem should be changed in this way on forcible > downgrades? > > Kirk Traditionally Unix file systems had all kind of odd semantics, but were considered reliable and stable. So *if* I had to choose b/t adding some odd semantics and keeping the f/s clean, I'd go for cleanliness. If, however, there is no clear winning option, then the next best thing would be to do one of the following: (1) add a new mount flag that can select among the choices. This option could be added to the mount(2) when it's doing a remount. We can select some default behavior for the flag, and let users turn it off if they want. (2) print some console message from there kernel upon remount, warning of any known open+unlinked files. BTW, can one remount read-only a f/s that's already remounted read-only? One (true hack) might be to keep the current remount semantics the same, but change them upon a second read-only remount. Erez. In contemplating these choices, I am inclined away from special mount flags or double remounting, as I think it would be hard to describe to a non-techinical person what these flags meant beyond down-grade cleanly and down-grade broken. The more I think about it the more I am inclined to say that non-forcible down-grade should fail if there are files open for writing or files open for reading with zero nlink counts. If a forcible downgrade is done, then all of these are vgone'd and the filesystem will be clean as a result. I had contemplated adding a -F (very forcible) unmount which unmounts without doing any additional I/O operations. Obviously such a filesystem would be dirty, but it would be useful if the disk (or NFS server) had died, and you just wanted the mount gone. But short of such an option, I think that a downgrade should leave the filesystem clean. Kirk To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Tue Mar 7 0:29:54 2000 Delivered-To: freebsd-fs@freebsd.org Received: from tele-post-20.mail.demon.net (tele-post-20.mail.demon.net [194.217.242.20]) by hub.freebsd.org (Postfix) with ESMTP id 4226437BF2C for ; Tue, 7 Mar 2000 00:29:50 -0800 (PST) (envelope-from dfr@nlsystems.com) Received: from nlsys.demon.co.uk ([158.152.125.33] helo=herring.nlsystems.com) by tele-post-20.mail.demon.net with esmtp (Exim 2.12 #2) id 12SFNE-00005M-0K; Tue, 7 Mar 2000 08:29:49 +0000 Received: from salmon.nlsystems.com (salmon.nlsystems.com [10.0.0.3]) by herring.nlsystems.com (8.9.3/8.8.8) with ESMTP id IAA96690; Tue, 7 Mar 2000 08:35:34 GMT (envelope-from dfr@nlsystems.com) Date: Tue, 7 Mar 2000 08:32:20 +0000 (GMT) From: Doug Rabson To: "Karsten W. Rohrbach" Cc: Mattias Pantzare , Edwin Mons , =?iso-8859-1?Q?Administra=E7=E3o_MailBR?= , freebsd-fs@freebsd.org Subject: Re: NFS: Linux Client and BSD Server In-Reply-To: <009701bf8776$b93963e0$f63f3fc3@hardcore> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Mon, 6 Mar 2000, Karsten W. Rohrbach wrote: > so, if you mount -o nfsv3 and have vfs.nfs.async=1 it wont fail or does this > mean that one has not to set vfs.nfs.async=1 ? > /k Note that you would set vfs.nfs.async=1 on the server, not the client. I think the server will still use the setting of this variable to change the way it interprets the protocol for NFSv3 but there is really no need to do this since NFSv3 has a better, safer way to get this kind of performance. -- Doug Rabson Mail: dfr@nlsystems.com Nonlinear Systems Ltd. Phone: +44 181 442 9037 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Tue Mar 7 15:15:21 2000 Delivered-To: freebsd-fs@freebsd.org Received: from cs.columbia.edu (cs.columbia.edu [128.59.16.20]) by hub.freebsd.org (Postfix) with ESMTP id 5675937B598; Tue, 7 Mar 2000 15:15:18 -0800 (PST) (envelope-from ezk@shekel.mcl.cs.columbia.edu) Received: from shekel.mcl.cs.columbia.edu (shekel.mcl.cs.columbia.edu [128.59.18.15]) by cs.columbia.edu (8.9.1/8.9.1) with ESMTP id SAA18884; Tue, 7 Mar 2000 18:15:09 -0500 (EST) Received: (from ezk@localhost) by shekel.mcl.cs.columbia.edu (8.9.1/8.9.1) id SAA04578; Tue, 7 Mar 2000 18:15:03 -0500 (EST) Date: Tue, 7 Mar 2000 18:15:03 -0500 (EST) Message-Id: <200003072315.SAA04578@shekel.mcl.cs.columbia.edu> From: Erez Zadok To: Kirk McKusick Cc: Erez Zadok , Alfred Perlstein , Terry Lambert , fs@freebsd.org, jkh@freebsd.org, Bruce Evans Subject: Re: changing mount options still can cause damage? In-reply-to: Your message of "Mon, 06 Mar 2000 15:14:52 PST." <200003062314.PAA16884@flamingo.McKusick.COM> Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org In message <200003062314.PAA16884@flamingo.McKusick.COM>, Kirk McKusick writes: [...] > I had contemplated adding a -F (very forcible) > unmount which unmounts without doing any additional I/O operations. > Obviously such a filesystem would be dirty, but it would be > useful if the disk (or NFS server) had died, and you just wanted > the mount gone. Puting my amd-maintainer hat on, such an option would be *extremely* useful to recover from a hung/dead automounter that left "hard" mount points. Since these are virtual mount points created by amd w/o any underlying persistent storage, there really won't be any corruption on disk. The current alternative to recover from a hung/dead amd is a reboot, where all I want is the offending struct vfs nuked with extreme prejudice... > > Kirk Erez. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Tue Mar 7 21:56:34 2000 Delivered-To: freebsd-fs@freebsd.org Received: from dt051n0b.san.rr.com (dt051n0b.san.rr.com [204.210.32.11]) by hub.freebsd.org (Postfix) with ESMTP id 0B35837B5CD for ; Tue, 7 Mar 2000 21:56:33 -0800 (PST) (envelope-from DougB@simplenet.com) Received: from simplenet.com (doug@master [10.0.0.2]) by dt051n0b.san.rr.com (8.9.3/8.9.3) with ESMTP id VAA28581; Tue, 7 Mar 2000 21:56:28 -0800 (PST) (envelope-from DougB@simplenet.com) Message-ID: <38C5EB8B.89F336B0@simplenet.com> Date: Tue, 07 Mar 2000 21:56:27 -0800 From: Doug Barton Organization: Triborough Bridge & Tunnel Authority X-Mailer: Mozilla 4.72 [en] (X11; U; FreeBSD 4.0-CURRENT-0302 i386) X-Accept-Language: en MIME-Version: 1.0 To: Erez Zadok Cc: Kirk McKusick , fs@FreeBSD.ORG Subject: Re: changing mount options still can cause damage? References: <200003072315.SAA04578@shekel.mcl.cs.columbia.edu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Erez Zadok wrote: > Puting my amd-maintainer hat on, such an option would be *extremely* useful > to recover from a hung/dead automounter that left "hard" mount points. > Since these are virtual mount points created by amd w/o any underlying > persistent storage, there really won't be any corruption on disk. > > The current alternative to recover from a hung/dead amd is a reboot, where > all I want is the offending struct vfs nuked with extreme prejudice... I would be heartily in favor of this as well, for "regular" NFS as well as amd. Having to reboot a machine to fix a hung mount just chaps my hide. Doug -- "Welcome to the desert of the real." - Laurence Fishburne as Morpheus, "The Matrix" To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message