From owner-freebsd-fs Sun Oct 15 22:51:28 2000 Delivered-To: freebsd-fs@freebsd.org Received: from fledge.watson.org (fledge.watson.org [204.156.12.50]) by hub.freebsd.org (Postfix) with ESMTP id 6D61B37B503 for ; Sun, 15 Oct 2000 22:51:26 -0700 (PDT) Received: from fledge.watson.org (robert@fledge.pr.watson.org [192.0.2.3]) by fledge.watson.org (8.9.3/8.9.3) with SMTP id BAA21505; Mon, 16 Oct 2000 01:51:15 -0400 (EDT) (envelope-from robert@fledge.watson.org) Date: Mon, 16 Oct 2000 01:51:15 -0400 (EDT) From: Robert Watson X-Sender: robert@fledge.watson.org To: Kherry Zamore Cc: freebsd-fs@freebsd.org Subject: Re: Overview of UFS In-Reply-To: <00ad01c03364$6f07c420$0202a8c0@majorzoot> 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 Wed, 11 Oct 2000, Kherry Zamore wrote: > Hi, I was just wondering if there are any places I could go to get an > overview of UFS and an understanding how it works "deep down". The usual authoritative source of information about BSD stuff is the "BSD Book", or "The Design and Implementation of the 4.4bsd Operating System" by McKusick, et al. While some parts of the book are relatively dated with respects to FreeBSD, the description of UFS should be fairly accurate (leaving aside some recent additions such as extended attributes, and soft updates in FFS). The normal computer book sources should have it in stock--try Barnes and Noble, et al. Robert N M Watson FreeBSD Core Team, TrustedBSD Project robert@fledge.watson.org NAI Labs, Safeport Network Services To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Mon Oct 16 2:27:36 2000 Delivered-To: freebsd-fs@freebsd.org Received: from relay2.mail.uk.psi.net (relay2.mail.uk.psi.net [154.32.107.6]) by hub.freebsd.org (Postfix) with ESMTP id B187637B502; Mon, 16 Oct 2000 02:27:30 -0700 (PDT) Received: from mail.plasmon.co.uk ([193.115.5.217]) by relay2.mail.uk.psi.net with smtp (Exim 2.12 #2) id 13l6Xr-0002RL-00; Mon, 16 Oct 2000 10:26:59 +0100 Received: by mail.plasmon.co.uk(Lotus SMTP MTA v4.6.4 (830.2 3-23-1999)) id 8025697A.0034103B ; Mon, 16 Oct 2000 10:28:40 +0100 X-Lotus-FromDomain: PLASNOTES From: dbhague@allstor-sw.co.uk To: Andre Albsmeier Cc: freebsd-scsi@FreeBSD.org, freebsd-fs@FreeBSD.org, smcintyre@allstor-sw.co.uk Message-ID: <8025697A.00340E6C.00@mail.plasmon.co.uk> Date: Mon, 16 Oct 2000 10:28:34 +0100 Subject: Re: Stressed SCSI subsystem locks up the system Mime-Version: 1.0 Content-type: text/plain; charset=us-ascii Content-Disposition: inline Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Andre, What were your SCSI errors ? We have one system that has now run for five days without failure. Today we will start to deconstruct this unit, any advice would be welcome. We also ran five system over the weekend and all but the one, the IDE system, failed. These were: A repeat of the passing system above, failed with Bad blocks 135666304, inode 5142534 6 seconds later, Bad blocks 135666304, inode 5634466 then, panic ffs_blkfree: freeing free frag, this is on the /RAID partition. Test run against an IDE disk, still running but slowly Test run against a SCSI disk Test run using a Symbios dual SCSI card, Test running FreeBSD 3.0 Two of the above tests have got struck in iowait, for example. root 451 0.0 0.1 368 172 p0 D Fri06PM 0:17.77 rm -rf /RAID/5 root 454 0.0 0.2 368 196 p0 D Fri06PM 0:17.85 rm -rf /RAID/7 root 455 0.0 0.2 368 196 p0 D Fri06PM 0:17.42 rm -rf /RAID/1 root 457 0.0 0.2 368 196 p0 D Fri06PM 0:17.44 rm -rf /RAID/2 root 459 0.0 0.2 368 196 p0 D Fri06PM 0:17.71 rm -rf /RAID/6 root 461 0.0 0.2 368 196 p0 D Fri06PM 0:17.10 rm -rf /RAID/4 root 463 0.0 0.2 368 196 p0 D Fri06PM 0:17.56 rm -rf /RAID/3 Just a few minutes ago cron started to die with a signal 10, we don't think this is relevant but... Oct 16 09:55:02 birch /kernel: pid 3551 (cron), uid 0: exited on signal 10 (core dumped) Oct 16 10:00:00 birch /kernel: pid 3555 (cron), uid 0: exited on signal 10 (core dumped) Oct 16 10:00:00 birch /kernel: pid 3556 (cron), uid 0: exited on signal 10 (core dumped) Oct 16 10:05:01 birch /kernel: pid 3558 (cron), uid 0: exited on signal 10 (core dumped) Oct 16 10:10:00 birch /kernel: pid 3560 (cron), uid 0: exited on signal 10 (core dumped) Oct 16 10:15:00 birch /kernel: pid 3562 (cron), uid 0: exited on signal 10 (core dumped) Oct 16 10:20:00 birch /kernel: pid 3564 (cron), uid 0: exited on signal 10 (core dumped) Regards Dave To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Mon Oct 16 2:47:34 2000 Delivered-To: freebsd-fs@freebsd.org Received: from goliath.siemens.de (goliath.siemens.de [194.138.37.131]) by hub.freebsd.org (Postfix) with ESMTP id BF68737B66C; Mon, 16 Oct 2000 02:47:27 -0700 (PDT) X-Envelope-Sender-Is: andre.albsmeier@mchp.siemens.de (at relayer goliath.siemens.de) Received: from mail2.siemens.de (mail2.siemens.de [139.25.208.11]) by goliath.siemens.de (8.11.0/8.11.0) with ESMTP id e9G9lOq13594; Mon, 16 Oct 2000 11:47:24 +0200 (MET DST) Received: from curry.mchp.siemens.de (curry.mchp.siemens.de [139.25.42.7]) by mail2.siemens.de (8.11.0/8.11.0) with ESMTP id e9G9lND04089; Mon, 16 Oct 2000 11:47:23 +0200 (MET DST) Received: (from localhost) by curry.mchp.siemens.de (8.11.1/8.11.1) id e9G9lNq17303; Date: Mon, 16 Oct 2000 11:47:23 +0200 From: Andre Albsmeier To: dbhague@allstor-sw.co.uk Cc: Andre Albsmeier , freebsd-scsi@FreeBSD.org, freebsd-fs@FreeBSD.org, smcintyre@allstor-sw.co.uk Subject: Re: Stressed SCSI subsystem locks up the system Message-ID: <20001016114723.A22193@curry.mchp.siemens.de> References: <8025697A.00340E6C.00@mail.plasmon.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <8025697A.00340E6C.00@mail.plasmon.co.uk>; from dbhague@allstor-sw.co.uk on Mon, Oct 16, 2000 at 10:28:34AM +0100 X-Echelon: BND CIA NSA Mossad KGB MI6 IRA detonator nuclear assault strike Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Mon, 16-Oct-2000 at 10:28:34 +0100, dbhague@allstor-sw.co.uk wrote: > Andre, > What were your SCSI errors ? Oct 13 10:05:28 server /ktry: (da2:ahc0:0:2:0): data overrun detected in Data-out phase. +Tag == 0xe. Oct 13 10:09:40 server /ktry: (da2:ahc0:0:2:0): Have seen Data Phase. Length = 65536. +NumSGs = 16. These appeared with the 3940AU. When replacing it with two 2940 everything worked great for several days now. I am keeping it this way now. When Justin does some driver changes, I will try my 3940AU again... -Andre > > We have one system that has now run for five days without failure. Today we > will start to deconstruct this unit, any advice would be welcome. > > We also ran five system over the weekend and all but the one, the IDE system, > failed. > These were: > A repeat of the passing system above, failed with > Bad blocks 135666304, inode 5142534 > 6 seconds later, Bad blocks 135666304, inode 5634466 > then, panic ffs_blkfree: freeing free frag, this is on the /RAID partition. > Test run against an IDE disk, still running but slowly > Test run against a SCSI disk > Test run using a Symbios dual SCSI card, > Test running FreeBSD 3.0 > > Two of the above tests have got struck in iowait, for example. > root 451 0.0 0.1 368 172 p0 D Fri06PM 0:17.77 rm -rf /RAID/5 > root 454 0.0 0.2 368 196 p0 D Fri06PM 0:17.85 rm -rf /RAID/7 > root 455 0.0 0.2 368 196 p0 D Fri06PM 0:17.42 rm -rf /RAID/1 > root 457 0.0 0.2 368 196 p0 D Fri06PM 0:17.44 rm -rf /RAID/2 > root 459 0.0 0.2 368 196 p0 D Fri06PM 0:17.71 rm -rf /RAID/6 > root 461 0.0 0.2 368 196 p0 D Fri06PM 0:17.10 rm -rf /RAID/4 > root 463 0.0 0.2 368 196 p0 D Fri06PM 0:17.56 rm -rf /RAID/3 > > Just a few minutes ago cron started to die with a signal 10, we don't think this > is relevant but... > Oct 16 09:55:02 birch /kernel: pid 3551 (cron), uid 0: exited on signal 10 (core > dumped) > Oct 16 10:00:00 birch /kernel: pid 3555 (cron), uid 0: exited on signal 10 (core > dumped) > Oct 16 10:00:00 birch /kernel: pid 3556 (cron), uid 0: exited on signal 10 (core > dumped) > Oct 16 10:05:01 birch /kernel: pid 3558 (cron), uid 0: exited on signal 10 (core > dumped) > Oct 16 10:10:00 birch /kernel: pid 3560 (cron), uid 0: exited on signal 10 (core > dumped) > Oct 16 10:15:00 birch /kernel: pid 3562 (cron), uid 0: exited on signal 10 (core > dumped) > Oct 16 10:20:00 birch /kernel: pid 3564 (cron), uid 0: exited on signal 10 (core > dumped) > > Regards Dave > -- Micro$oft: Which virus will you get today? To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Mon Oct 16 10:12:16 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 DF34937B670 for ; Mon, 16 Oct 2000 10:12:10 -0700 (PDT) Received: (qmail 17378 invoked by uid 1000); 16 Oct 2000 17:12:05 -0000 Date: Mon, 16 Oct 2000 19:12:04 +0200 From: "Karsten W. Rohrbach" To: Greg Lehey Cc: Bruce Evans , Andre Albsmeier , Marc Tardif , freebsd-fs@FreeBSD.org, freebsd-scsi@FreeBSD.org Subject: Re: ccd with other filesystems Message-ID: <20001016191204.A17349@rohrbach.de> Reply-To: karsten@rohrbach.de References: <20001001114540.G43885@wantadilla.lemis.com> <20001001190350.B304@rohrbach.de> <20001014193120.D1489@sydney.worldwide.lemis.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0i In-Reply-To: <20001014193120.D1489@sydney.worldwide.lemis.com>; from grog@lemis.com on Sat, Oct 14, 2000 at 07:31:20PM -0700 X-Arbitrary-Number-Of-The-Day: 42 X-Sender: karsten@rohrbach.de Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org let's discuss this in the lounge at the con in two days ;-) whoever wins pays the dinner *grin* /k Greg Lehey(grog@lemis.com)@Sat, Oct 14, 2000 at 07:31:20PM -0700: > On Sunday, 1 October 2000 at 19:03:50 +0200, Karsten W. Rohrbach wrote: > > Bruce Evans(bde@zeta.org.au)@Sun, Oct 01, 2000 at 11:32:29PM +1100: > >> On Sun, 1 Oct 2000, Greg Lehey wrote: > >> > >>> On Sunday, 1 October 2000 at 4:09:37 +0200, Karsten W. Rohrbach wrote: > >>>> i dont quite know why it is still possible doing a newfs on a 'c' > >>>> partition, since the partition type is 'unused' and not > >>>> '4.2BSD'. newfs should check this and throw an error while providing > >>>> an expert-only-feature command line option to explicitly override > >>>> it. > >>> > >>> I think this is a bug in newfs. > >> > >> This is a feature of newfs. It is almost device-independent. E.g., to > >> create a filesystem in a regular file with no label in sight: > >> > >> dd if=/dev/zero of=foo oseek=2779 count=1 > >> # -v and ./foo work around device dependence. > >> # -T floppy is so that I don't have to type a lot of args for this example. > >> newfs -v -T floppy ./foo > > > > and thats exactly the point... if you use -v it should work, if you dont > > the partition type must be 4.2BSD and it must not be mounted. > > no more no less? not reasonable? > > No, that's not really the issue. -v just says "don't make assumptions > about the partition from the device name". What we really need here > is a "force" flag ("yes, I know it's not a 4.2BSD partition, but do it > anyway"). Unfortunately, -f is already taken. > > Greg > -- > Finger grog@lemis.com for PGP public key > See complete headers for address and phone numbers -- > "Under capitalism, man exploits man. Under Communism, it's just the > opposite." -- John Kenneth Galbraith KR433/KR11-RIPE -- http://www.webmonster.de -- ftp://ftp.webmonster.de To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Mon Oct 16 12:30: 5 2000 Delivered-To: freebsd-fs@freebsd.org Received: from echunga.lemis.com (freebie.lemis.com [192.109.197.137]) by hub.freebsd.org (Postfix) with ESMTP id 891BC37B66C; Mon, 16 Oct 2000 12:29:56 -0700 (PDT) Received: (from grog@localhost) by sydney.worldwide.lemis.com (8.11.0/8.9.3) id e9F2VK603298; Sat, 14 Oct 2000 22:31:20 -0400 (EDT) (envelope-from grog) Date: Sat, 14 Oct 2000 19:31:20 -0700 From: Greg Lehey To: "Karsten W. Rohrbach" Cc: Bruce Evans , Andre Albsmeier , Marc Tardif , freebsd-fs@FreeBSD.org, freebsd-scsi@FreeBSD.org Subject: Re: ccd with other filesystems Message-ID: <20001014193120.D1489@sydney.worldwide.lemis.com> References: <20001001114540.G43885@wantadilla.lemis.com> <20001001190350.B304@rohrbach.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0i In-Reply-To: <20001001190350.B304@rohrbach.de>; from karsten@rohrbach.de on Sun, Oct 01, 2000 at 07:03:50PM +0200 Organization: LEMIS, PO Box 460, Echunga SA 5153, Australia Phone: +61-8-8388-8286 Fax: +61-8-8388-8725 Mobile: +61-418-838-708 WWW-Home-Page: http://www.lemis.com/~grog X-PGP-Fingerprint: 6B 7B C3 8C 61 CD 54 AF 13 24 52 F8 6D A4 95 EF Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Sunday, 1 October 2000 at 19:03:50 +0200, Karsten W. Rohrbach wrote: > Bruce Evans(bde@zeta.org.au)@Sun, Oct 01, 2000 at 11:32:29PM +1100: >> On Sun, 1 Oct 2000, Greg Lehey wrote: >> >>> On Sunday, 1 October 2000 at 4:09:37 +0200, Karsten W. Rohrbach wrote: >>>> i dont quite know why it is still possible doing a newfs on a 'c' >>>> partition, since the partition type is 'unused' and not >>>> '4.2BSD'. newfs should check this and throw an error while providing >>>> an expert-only-feature command line option to explicitly override >>>> it. >>> >>> I think this is a bug in newfs. >> >> This is a feature of newfs. It is almost device-independent. E.g., to >> create a filesystem in a regular file with no label in sight: >> >> dd if=/dev/zero of=foo oseek=2779 count=1 >> # -v and ./foo work around device dependence. >> # -T floppy is so that I don't have to type a lot of args for this example. >> newfs -v -T floppy ./foo > > and thats exactly the point... if you use -v it should work, if you dont > the partition type must be 4.2BSD and it must not be mounted. > no more no less? not reasonable? No, that's not really the issue. -v just says "don't make assumptions about the partition from the device name". What we really need here is a "force" flag ("yes, I know it's not a 4.2BSD partition, but do it anyway"). Unfortunately, -f is already taken. Greg -- Finger grog@lemis.com for PGP public key See complete headers for address and phone numbers To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Thu Oct 19 3:52: 3 2000 Delivered-To: freebsd-fs@freebsd.org Received: from relay1.mail.uk.psi.net (relay1.mail.uk.psi.net [154.32.105.6]) by hub.freebsd.org (Postfix) with ESMTP id E6F4337B4C5; Thu, 19 Oct 2000 03:51:57 -0700 (PDT) Received: from mail.plasmon.co.uk ([193.115.5.217]) by relay1.mail.uk.psi.net with smtp (Exim 2.12 #2) id 13mDIb-00019i-00; Thu, 19 Oct 2000 11:51:49 +0100 Received: by mail.plasmon.co.uk(Lotus SMTP MTA v4.6.4 (830.2 3-23-1999)) id 8025697D.003BD474 ; Thu, 19 Oct 2000 11:53:30 +0100 X-Lotus-FromDomain: PLASNOTES From: dbhague@allstor-sw.co.uk To: dbhague@allstor-sw.co.uk Cc: freebsd-scsi@FreeBSD.org, freebsd-fs@FreeBSD.org, gibbs@scsiguy.com, mjacob@feral.com, Andre Albsmeier , smcintyre@allstor-sw.co.uk, paul@originative.co.uk Message-ID: <8025697D.003BD2D1.00@mail.plasmon.co.uk> Date: Thu, 19 Oct 2000 11:53:25 +0100 Subject: Re: Stressed SCSI subsystem locks up the system Mime-Version: 1.0 Content-type: text/plain; charset=us-ascii Content-Disposition: inline Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Further to my earlier mails we have now trapped a failure and I enclose the stack trace. I have typed this in but do have a picture of the screen if required. We have also had several more of the >panic ffs_blkfree freeing free frags type errors. I believe by moving to the current 4.1 stable we now have a file system problem, Any comments would be welcome. Regards Dave start = 0, len =2, fs = /RAID panic: ffs_alloccg: map corrupted Debugger("panic") Stopped at Debugger+0x34: movb $0,in_Debugger.396 db> tr Debugger(c02d0063) at Debugger+0x34 panic(c02e5f7f,c02e5f60,0,2,c0b388d4) at panic+0x70 ffs_mapsearch(c0b38800,c47cc000,2bc0008,3,578) at ffs_mapsearch+0x128 ffs_alloccg((c107e400,578,2bc0008,400) at ffs_alloccg+0x21e ffs_hashalloc(c107e400,578,2bc0008,400,c0246748) at ffs_hashalloc+0x23 ffs_alloc(c107e400,0,2bc0008,400,c0b2c000) at ffs_alloc_0x129 ffs_balloc(c871ae2c,c8cc6780,c11d0a40,c871aedc,c88b0300) at fss_balloc+0x476 ffa_write(c871ae6c,c7cf01e0,c7cf01e0,400,1) at ffs_write+0x33d vn_write(c11d0a40,c871aedc,c0b2c000,0,c7cf01e0) at vn_write+0x160 dofilewrite(c7f01e0,c11d0a40,3,80567a0,400) at dofilewrite+0xb1 write(c7cf01e0,c871af80,0,180,bfbffbcc) at write+0x33 syscall2(bfbf002f,bfbf002f,bfbf002f,bfbffbcc,180) at syscall+0x205 Xint0x80_syscall() at Xint0x80_syscall+0x25 db>_ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Thu Oct 19 18:57:17 2000 Delivered-To: freebsd-fs@freebsd.org Received: from relay.butya.kz (butya-gw.butya.kz [212.154.129.94]) by hub.freebsd.org (Postfix) with ESMTP id 1B1E237B4D7; Thu, 19 Oct 2000 18:57:13 -0700 (PDT) Received: by relay.butya.kz (Postfix, from userid 1000) id 594B22878E; Fri, 20 Oct 2000 08:57:00 +0700 (ALMST) Received: from localhost (localhost [127.0.0.1]) by relay.butya.kz (Postfix) with ESMTP id 1AFFF2878D; Fri, 20 Oct 2000 08:57:00 +0700 (ALMST) Date: Fri, 20 Oct 2000 08:56:59 +0700 (ALMST) From: Boris Popov To: freebsd-hackers@freebsd.org, freebsd-fs@freebsd.org Subject: smbfs-1.3.0 released 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 Hello, At first, I'm want to say 'thank you' for everybody who provided me with feedback on my work. So, here is a records from the HISTORY file for last two releases: 20.10.2000 1.3.0 - Network IO engine significantly reworked. Now it uses kernel threads to implement 'smbiod' process which handles network traffic for each VC. Previous model were incapable to serve large number of mount points and didn't work well with intensive IO operations performed on a different files on the same mount point. Special care was taken on better usage of MP systems. Unfortunately, kernel threads aren't supported by FreeBSD 3.X and for now it is excluded from the list of supported systems. - Reduce overhead caused by using single hash table for each mount point. 26.09.2000 1.2.8 (never released) - More SMP related bugs are fixed. - Make smbfs compatible with the Linux emulator. - smbfs now known to work with IBM LanManager (special thanks to Eugen Averin) - Fix problem with files bigger than 2GB (reported by Lee McKenna) - Please note that smbfs may not work properly with FreeBSD 3.X. New version can be downloaded from: ftp://ftp.butya.kz/pub/smbfs/smbfs.tar.gz -- Boris Popov http://www.butya.kz/~bp/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Fri Oct 20 7: 4:59 2000 Delivered-To: freebsd-fs@freebsd.org Received: from ns.draper.com (ns.draper.com [140.102.2.1]) by hub.freebsd.org (Postfix) with ESMTP id 3D80E37B4CF for ; Fri, 20 Oct 2000 07:04:57 -0700 (PDT) Received: from mb2.draper.com ([140.102.16.18]) by ns.draper.com (PMDF V5.2-32 #37580) with ESMTP id <0G2Q00MMOF3YZR@ns.draper.com> for freebsd-fs@FreeBSD.ORG; Fri, 20 Oct 2000 10:04:46 -0400 (EDT) Received: from CONVERSION-DAEMON by mb2.draper.com (PMDF V5.2-32 #37579) id <0G2Q00I01F3WZD@mb2.draper.com>; Fri, 20 Oct 2000 10:04:44 -0400 (EDT) Received: from jbg2408 (jbg2408.draper.com [140.102.22.33]) by mb2.draper.com (PMDF V5.2-32 #37579) with ESMTP id <0G2Q00IOFF3V9L@mb2.draper.com>; Fri, 20 Oct 2000 10:04:44 -0400 (EDT) Date: Fri, 20 Oct 2000 10:05:40 -0400 From: Jason Goodman Subject: journaled file system X-Sender: jbg2408@imap.draper.com To: freebsd-fs@FreeBSD.ORG, don@calis.blacksun.org Message-id: <4.2.2.20001020100054.00b33e00@imap.draper.com> MIME-version: 1.0 X-Mailer: QUALCOMM Windows Eudora Pro Version 4.2.2 Content-type: text/plain; charset="us-ascii" Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hi. I found your email addresses by accident while surfing the web. I am looking for a product for windows NT or windows 2000 that has the journaled file system feature similar to what Unix offers. Basically what I need is a product that has an API that would allow me to either send an Internet Datagram to a Listener, when a file is added to a folder or changed, or has an API that would allow me to interact with JDBC in some way? Any help in this search would be appreciated. Thanks, Jason To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message From owner-freebsd-fs Fri Oct 20 8:57:30 2000 Delivered-To: freebsd-fs@freebsd.org Received: from smtp03.primenet.com (smtp03.primenet.com [206.165.6.133]) by hub.freebsd.org (Postfix) with ESMTP id 1562037B4C5 for ; Fri, 20 Oct 2000 08:56:06 -0700 (PDT) Received: (from daemon@localhost) by smtp03.primenet.com (8.9.3/8.9.3) id IAA01192; Fri, 20 Oct 2000 08:54:12 -0700 (MST) Received: from usr06.primenet.com(206.165.6.206) via SMTP by smtp03.primenet.com, id smtpdAAAFYaqpc; Fri Oct 20 08:54:06 2000 Received: (from tlambert@localhost) by usr06.primenet.com (8.8.5/8.8.5) id IAA05637; Fri, 20 Oct 2000 08:54:24 -0700 (MST) From: Terry Lambert Message-Id: <200010201554.IAA05637@usr06.primenet.com> Subject: Re: journaled file system To: jgoodman@draper.com (Jason Goodman) Date: Fri, 20 Oct 2000 15:54:24 +0000 (GMT) Cc: freebsd-fs@FreeBSD.ORG, don@calis.blacksun.org In-Reply-To: <4.2.2.20001020100054.00b33e00@imap.draper.com> from "Jason Goodman" at Oct 20, 2000 10:05:40 AM X-Mailer: ELM [version 2.5 PL2] MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > Hi. I found your email addresses by accident while surfing the > web. I am looking for a product for windows NT or windows 2000 > that has the journaled file system feature similar to what Unix > offers. NTFS is log structured; for what you are requesting, the type of FS is irrelevant. Please see: Inside the Windows NT File System Helen Custer Microsoft Press ISBN: 1-55615-660-X > Basically what I need is a product that has an API that would > allow me to either send an Internet Datagram to a Listener, when > a file is added to a folder or changed, You will need to write your own daemon for this. The NTFS supports the concept of "Monitors", which is half of what you are asking for in a product. Monitors are a generic FS term meaning "tell me when something changes". Many systems support Monitors. Windows and MacOS have pretty much supported them since the beginning, since if you add a file to a directory, a file browser with a view open onto the directory will need to update itself. The orginal Windows 3.x series did this by periodically polling the directory to determine if contents have changed. The Macintosh approach was to implement volume change notification, so that if the data anywhere on a volume changed, then the views onto directories on that volume would be updated. Appletalk implements this in their network file system by doing a poll for the volume modification timestap, once every 11 seconds. Most Appletalk implementations for UNIX will always state that the modification time is the current time, causing a network poll. In FreeBSD, there is a true volume modification timestamp that can be used, but of course, the Appletalk servers are user space processes, and unless they limit volume spanning over mount points on exports, this quickly degrades back to polling everything (I introduced this in 1994, when I was working at Novell on NWFS, in support of much lower network overhead for Appletalk services served by Novell servers). In any case, you can get what you want, but you will need to hook the IFSMgr interface to get your events, or use the Windows API, if you can find the documentation (it's pretty well hidden, and I do not have my Windows develeopement system out of mothballs right now, so I can't "just tell you"). The reason you will need to write your own daemon is because you want to do datagram notification; since this is inherently lossy (there is no retransmission and no receipt failure notification), most people would never implement this or find it useful, except in very specific applications where the lossy nature of the media wouldn't effect the integrity of the output of the system as a whole. You may also want to check with Veritas, since last I heard, they has a VXFS for Windows, and they probably document their stuff as well on NT as they did on UNIX. > or has an API that would allow me to interact with JDBC in some > way? That's really way out in left field. File system modification events really have nothing to do with the Java Database connector cruft. In general, an FS _can_ be considered to be a very special purpose database, but you would have a hell of a time wedging FS semantics into the very small model space that JDBC represents; JDBC is pretty useless for everything but SQL, so wedging FS change notifications into JDBC SQL "triggers" is probably a non-starter as an approach (and yes, I know, some idiots have built JDBC connectors for LDAP: I maintain they aren't useful, except for a tiny set of problems, or where there is enough horse power to brute-force something like an LDAP/SQL synchronization, and they are too lazy to implement to the JNDI, or have to deal with a legacy system integration and the short-sighted people requesting it). NB: If you were to wedge FS change notifications in somehow, you sure as heck would not be using datagrams for event delivery! The JDBC SQL "trigger" semantics are such that dropping events on the floor, as datagrams might, wil compromise your whole transaction model. PS: Are you sure you aren't really looking for B2Bi software, such as IBM's MQ-Series or Active Software's offerings (or, to date myself, USL's Tuxedo)? If you are going to do a job, you should probably use the right tools, instead of kludging it. A CORBA intterface might be the right approach, if you are insistent. PPS: Of course, you have only said "how do I get this soloution which I have already arrived at?", and haven't really told us the problem you are trying to solve, so there's really no way we can tell you what is or isn't appropriate technology. PPPS: Probably you want to reexamine your problem, instead of assuming the soloution, since there may be other soloutions that are more natural for the problem. PPPPS: Or you could just tell us your problem, instead, and ask for recommendations on a soloution... Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message