Re: [Cooker] ReiserFS killed my /var
Thierry Vignaud wrote: SI Reasoning [EMAIL PROTECTED] writes: Should I disable writeback mode? safer but slower on writes :-) if you want, put a hdparm -W 0 /dev/hd[XYZ] in your /etc/rc.d/rc.local man hdparm says that write-caching is disabled by default. Grégoire
Re: [Cooker] ReiserFS killed my /var
SI Reasoning [EMAIL PROTECTED] writes: about reiserfs ate my /var partition, note that ext3 writeback mode where metadata only are journalized, you can still have old date or mix of old and new data in your files on remount after crash. ext3 ordered mode should be safer. but : note also that for eide disks, most of them works in writeback mode by default, ie the eide controller lye to the OS and say ok, it has just been written on the disk whereas it's only in the disk cache. so eide disks can lead to corruption dispite the medata journal. scsi disks should be safer here. there're also other hw problems than can lead to metadata corruption like the via bugs, the dma engine still writing do disk when the power is shoot down because it's powered longer than ram (ram needs to be refresh thousands times per second to keep its data), ... does the eide disk problem also affect reiserfs? eide writeback mode affects all fs : by saying ok it's writing whereas it's only cached, the eide disk/controller couple waits for mode I/O to do them more intelligently if they can order them, they can achieve better write performance but if a power problem or a kernel oops occurs, then you can lost your data since they're not yet written. this can break journaled fs since they rely on physically written metadata. but the odds are high you see data corruption before metadata corruption. Should I disable writeback mode? safer but slower on writes :-) if you want, put a hdparm -W 0 /dev/hd[XYZ] in your /etc/rc.d/rc.local
Re: [Cooker] ReiserFS killed my /var
as far as eide hard drives is there not a bios setting to control that? --- Thierry Vignaud [EMAIL PROTECTED] wrote: Nima S. Panahi [EMAIL PROTECTED] writes: I have had mixed results. I tried to keep my mouth shut during the whole thread, but I can't any longer. I experimented with Reiser since 7.2. All it did was each my paritions and I think the VIA bug has something to do with it. But this was not a rare occation, it ate my /home and /var multiple times and on different machines (all three had some VIA chipset as I only use AMD). Then, the newer Reiser version cameout and changing the paritions to those helped but I still saw problems where it reported a directory to have 2 gigs of data in it when I only had 400 megs! I must have been smoking something, because for my server, I went all ReiserFS with the latest 8.0. I was doing a switch from the old server to a new one and in the excitement I forgot about my ReiserFS problems. I almost switch back as soon as I realized what I had done. However, I just bited down and did regular backups waiting for when it went south. IT NEVER DID! I hope I am not talking too soon, but for a few months it has been running w/o any problems. I would still waite for it to mature more or use XFS, which I have used in many of my desktops. IMHO anyways... about reiserfs ate my /var partition, note that ext3 writeback mode where metadata only are journalized, you can still have old date or mix of old and new data in your files on remount after crash. ext3 ordered mode should be safer. but : note also that for eide disks, most of them works in writeback mode by default, ie the eide controller lye to the OS and say ok, it has just been written on the disk whereas it's only in the disk cache. so eide disks can lead to corruption dispite the medata journal. scsi disks should be safer here. there're also other hw problems than can lead to metadata corruption like the via bugs, the dma engine still writing do disk when the power is shoot down because it's powered longer than ram (ram needs to be refresh thousands times per second to keep its data), ... reiserfs = SI Reasoning [EMAIL PROTECTED] gnupg/pgp key id 035213BC __ Do You Yahoo!? Make international calls for as low as $.04/minute with Yahoo! Messenger http://phonecard.yahoo.com/
Re: [Cooker] ReiserFS killed my /var
does the eide disk problem also affect reiserfs? Should I disable writeback mode? --- Thierry Vignaud [EMAIL PROTECTED] wrote: about reiserfs ate my /var partition, note that ext3 writeback mode where metadata only are journalized, you can still have old date or mix of old and new data in your files on remount after crash. ext3 ordered mode should be safer. but : note also that for eide disks, most of them works in writeback mode by default, ie the eide controller lye to the OS and say ok, it has just been written on the disk whereas it's only in the disk cache. so eide disks can lead to corruption dispite the medata journal. scsi disks should be safer here. there're also other hw problems than can lead to metadata corruption like the via bugs, the dma engine still writing do disk when the power is shoot down because it's powered longer than ram (ram needs to be refresh thousands times per second to keep its data), ... reiserfs = SI Reasoning [EMAIL PROTECTED] gnupg/pgp key id 035213BC __ Do You Yahoo!? Make international calls for as low as $.04/minute with Yahoo! Messenger http://phonecard.yahoo.com/
Re: [Cooker] ReiserFS killed my /var
I have had mixed results. I tried to keep my mouth shut during the whole thread, but I can't any longer. I experimented with Reiser since 7.2. All it did was each my paritions and I think the VIA bug has something to do with it. But this was not a rare occation, it ate my /home and /var multiple times and on different machines (all three had some VIA chipset as I only use AMD). Then, the newer Reiser version cameout and changing the paritions to those helped but I still saw problems where it reported a directory to have 2 gigs of data in it when I only had 400 megs! I must have been smoking something, because for my server, I went all ReiserFS with the latest 8.0. I was doing a switch from the old server to a new one and in the excitement I forgot about my ReiserFS problems. I almost switch back as soon as I realized what I had done. However, I just bited down and did regular backups waiting for when it went south. IT NEVER DID! I hope I am not talking too soon, but for a few months it has been running w/o any problems. I would still waite for it to mature more or use XFS, which I have used in many of my desktops. IMHO anyways... On Wed, 8 Aug 2001, Harry wrote: On 8/6/01 2:04 AM, Stefan Siegel [EMAIL PROTECTED] wrote: Since then I don't trust ReiserFS = I wouldn't install it currently on a critical system, only on playgounds for data loosing ... Depending on the nature of your electrical failure, no data is really safe if it is preceded by a spike while the drive writes to the disk. In general, I have found Reiserfs to be admirably reliable. Harry
Re: [Cooker] ReiserFS killed my /var
On 8/6/01 2:04 AM, Stefan Siegel [EMAIL PROTECTED] wrote: Since then I don't trust ReiserFS = I wouldn't install it currently on a critical system, only on playgounds for data loosing ... Depending on the nature of your electrical failure, no data is really safe if it is preceded by a spike while the drive writes to the disk. In general, I have found Reiserfs to be admirably reliable. Harry
Re: [Cooker] ReiserFS killed my /var
Oh it can't be possible : yesterday someone had problems with ReiserFS, and this morning my ReiserFS / said no more space avaliable on device! The df command says unable the read the partition information ! seems like someone has putten a not funny virus in ReiserFS... /\ | José Jorge [EMAIL PROTECTED]| | TEKLYNX International http://www.teklynx.com | \/ Brook Humphrey [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent by: cc: cooker-owner@linux-maSubject: Re: [Cooker] ReiserFS killed my /var ndrake.com 06/08/2001 21:03 Please respond to cooker On Monday 06 August 2001 10:09 am, you wrote: On Mon Aug 06, 2001 at 04:07:18PM +0200, Stefan Siegel wrote: This is really odd... I've been using reiserfs almost exclusively since 7.2 on all my machines. The only partition that I don't use it for is /var because I run qmail on all my machines and there was an issue with qmail and reiserfs in the past (no idea if it's fixed, but /var being ext2 is a matter of habit now). Anyways, I've had a number of power outtages despite having UPS's with the systems all of a sudden going down. I may have lost the odd file occassionally, but certainly never an entire partition. Considering this is on about 5 machines all showing the same thing, I'm a little hesitant to think it's reiserfs *alone*. Maybe there is something with the files on /var that causes this problem (again, maybe if I had reiserfs /var partitions I would have experienced the same thing, I don't know). I do know that I have never lost more than a handful of files and I've got many reiserfs partitions across various systems. Same here. I did loose two partitions on completetly different hard drives but tracked it down to hard drive controler failure. The only thing I can say about the failure is that the ext2 partitions survived it but reiser didn't However I run reiser on almost everything and this is the first time I've had a problem.
Re: [Cooker] ReiserFS killed my /var
Es schrieb Borsenkow Andrej: On 8/3/01 2:21 PM, Grégoire Colbert [EMAIL PROTECTED] wrote: 1- Is ReiserFS good enough nowadays for a production machine? Yes, in fact I use it on production machines with great success. I use it as well, but my systems are not mission critical :-) 2- Will it ever become the default filesystem for Mandrake? One can hope. I'm sure it will be made default after RedHat does so :-) Eh? It is default file system suggested by 8.0 install. What else do you mean under default filesystem? It killed my var partiton. When I did reboot my desktop machine after an electricity failure, my /var was gone. So I am currently left alone with a 8.0 WITHOUT any RPM support, as all relevant files reside there :-\ (In fact the whole system is relatively unusable without an intact /var, but most can be reconstructed, except the RPM-DB) As this was my first encounter with a power failure (and I was hoping the jounal would fix arising problems) I lost all data on the ReiserFS partition. Lucky me it was not /home with my (not yet finished) diploma thesis on it ... Since then I don't trust ReiserFS = I wouldn't install it currently on a critical system, only on playgounds for data loosing ... -- _ Tschüss und bis demnächst/à bientôt, _|_|_ () * Stefan /v\ / »( )« Penguin Powered! +(m-m)--+
RE: [Cooker] ReiserFS killed my /var
It killed my var partiton. When I did reboot my desktop machine after an electricity failure, my /var was gone. So I am currently left alone with a 8.0 WITHOUT any RPM support, as all relevant files reside there :-\ (In fact the whole system is relatively unusable without an intact /var, but most can be reconstructed, except the RPM-DB) What exactly do you mean? - it was not possible to mount /var? - it was possible to mount but /var was empty? - it was not entirely empty but some files were gone? Have you tried fsck it? -andrej
Re: [Cooker] ReiserFS killed my /var
Es schrieb Borsenkow Andrej: It killed my var partiton. When I did reboot my desktop machine after an electricity failure, my /var was gone. So I am currently left alone with a 8.0 WITHOUT any RPM support, as all relevant files reside there :-\ (In fact the whole system is relatively unusable without an intact /var, but most can be reconstructed, except the RPM-DB) What exactly do you mean? - it was not possible to mount /var? Yes, boot pricess gave ma an error message (don't remember any more) and left me with a prompt. - it was possible to mount but /var was empty? It was unwilling to mount it (still don't remember error message). - it was not entirely empty but some files were gone? As it was unwilling to mount, I have no idea. Have you tried fsck it? After it was unwillint to mount it, yes. It repoted a lot of note cecking/fixing line (by the way it warned me several times that ReiserFsck is totaly unstable and might destroy everything before it was starting to check the partition ...). After that I was able to mount the partition again, but it was totally empty !!! :-( As a result I ran mke2fs on that partition, remounted it and created some directories to get able to start my system again. It was a nightmare (I thaught sometines I would be sitting on a Window$ system, as all my data went away with no possibility to get it back). As a result my laptop only runs a stable file system (aka ext2) and I never had such problems ... -- _ Tschüss und bis demnächst/à bientôt, _|_|_ () * Stefan /v\ / »( )« Penguin Powered! +(m-m)--+ | Stefan Siegel | http://www.student.uni-kl.de/~siegel/ | | Kurt-Schumacher-Str. 34 / App.144 | mailto:[EMAIL PROTECTED]| | D-67663 Kaiserslautern| PGP Public Key: | | Tel.: +49-631-18269 | finger -l [EMAIL PROTECTED] | +---+
RE: [Cooker] ReiserFS killed my /var
What exactly do you mean? - it was not possible to mount /var? Yes, boot pricess gave ma an error message (don't remember any more) and left me with a prompt. ... Have you tried fsck it? After it was unwillint to mount it, yes. It repoted a lot of note cecking/fixing line (by the way it warned me several times that ReiserFsck is totaly unstable and might destroy everything before it was starting to check the partition ...). After that I was able to mount the partition again, but it was totally empty !!! :-( There is an option rebuild tree (sorry, do not have access to Mandrake just now, cannot check exactly). I guess you have not tried it? I ask because I'm really interested in how to recover a broken ReiserFS. So far, I could not find any real information, looks like reiser developers assume it never goes wrong. Any pointers are appreciated. And of course, I can well understand your feeling about it ... -andrej
Re: [Cooker] ReiserFS killed my /var
Es schrieb Borsenkow Andrej: [...] After that I was able to mount the partition again, but it was totally empty !!! :-( There is an option rebuild tree (sorry, do not have access to Mandrake just now, cannot check exactly). I guess you have not tried it? I did, as this was IIRC te only part doing anything, the other options all reported some errors but did nothing (still unable to mount afterwards). I ask because I'm really interested in how to recover a broken ReiserFS. So far, I could not find any real information, looks like reiser developers assume it never goes wrong. It seems so ... And of course, I can well understand your feeling about it ... Thx. I am sorry I can't provide you more informations, I've tried to wipe it out of my brian, as it is an awfull experiance, loosing relevant parts of an OS, specially when it is labeld rocksolid like GNU/Linux is these days ... -- _ Tschüss und bis demnächst/à bientôt, _|_|_ () * Stefan /v\ / »( )« Penguin Powered! +(m-m)--+
Re: [Cooker] ReiserFS killed my /var
Am 2001-08-06, um 15:41:26 (+0200) schrieb Guillaume Cottenceau: Stefan Siegel [EMAIL PROTECTED] writes: [...] thesis on it ... Since then I don't trust ReiserFS = I wouldn't install it currently on a critical system, only on playgounds for data loosing ... The question is whether it ameliorates mean crash recovery on critical systems. Nornally journaling would go in the right direction, but there are still potential problems. Any filesystem you're using, you'll still want to do periodic saves on tape devices, or even replication. Sure in theory you are right. But ReiserFS don't seem to be stable enough (atleast for me) as on the _FIRST_ crash (Power failure on the whole PC) I had, it shrederd my data. ext2 sometime eats data too, I know - but only files - at least on my side it never killed whole partitions. I am hoping for ext3 as it works on top of ext2 afaik. -- _ Tschüss und bis demnächst/à bientôt, _|_|_ () * Stefan /v\ / »( )« Penguin Powered! +(m-m)--+
Re: [Cooker] ReiserFS killed my /var
On Mon Aug 06, 2001 at 04:07:18PM +0200, Stefan Siegel wrote: thesis on it ... Since then I don't trust ReiserFS = I wouldn't install it currently on a critical system, only on playgounds for data loosing ... The question is whether it ameliorates mean crash recovery on critical systems. Nornally journaling would go in the right direction, but there are still potential problems. Any filesystem you're using, you'll still want to do periodic saves on tape devices, or even replication. Sure in theory you are right. But ReiserFS don't seem to be stable enough (atleast for me) as on the _FIRST_ crash (Power failure on the whole PC) I had, it shrederd my data. ext2 sometime eats data too, I know - but only files - at least on my side it never killed whole partitions. I am hoping for ext3 as it works on top of ext2 afaik. This is really odd... I've been using reiserfs almost exclusively since 7.2 on all my machines. The only partition that I don't use it for is /var because I run qmail on all my machines and there was an issue with qmail and reiserfs in the past (no idea if it's fixed, but /var being ext2 is a matter of habit now). Anyways, I've had a number of power outtages despite having UPS's with the systems all of a sudden going down. I may have lost the odd file occassionally, but certainly never an entire partition. Considering this is on about 5 machines all showing the same thing, I'm a little hesitant to think it's reiserfs *alone*. Maybe there is something with the files on /var that causes this problem (again, maybe if I had reiserfs /var partitions I would have experienced the same thing, I don't know). I do know that I have never lost more than a handful of files and I've got many reiserfs partitions across various systems. -- [EMAIL PROTECTED], OpenPGP key available on www.keyserver.net 1024D/FE6F2AFD 88D8 0D23 8D4B 3407 5BD7 66F9 2043 D0E5 FE6F 2AFD - Danen Consulting Serviceswww.danen.net, www.freezer-burn.org - MandrakeSoft, Inc. Security www.linux-mandrake.com Current Linux kernel 2.4.3-20mdk-win4lin uptime: 10 days 22 hours 51 minutes. PGP signature
Re: [Cooker] ReiserFS killed my /var
[Snip] I am using reiserfs since 7.2, including /var. And I never had any problems with it. And we had more than one powerfailure ;). I am so confident with it that windows-user-method I sometimes just click on reset if the system seems to be hanging /windows-user-method Everytime it came up perfectly clear. From my expirience I would say that you loss of /var filesystem was not due to reiserfs but due to something else. But thats just IMHO. Take care, Andre P.S.: Even if you use reiserfs, never underestimate the power of Murphy! (Do backups)!
RE: [Cooker] ReiserFS killed my /var
I have 7 raiserfs system including 4 SMP 4 proc boxes and I have never had problem with any of them except with one old box which I'm still using as my firewall. The system would started corrupting volumes if I used 2.4 kernel with reiserfs. When installed with 2.2 kernel and reiserfs or 2.4 kernel with ext2 the box could tick for months. The system in question was old 120Mhz Pentium (no mmx) with 96 Mb memory and Adaptec 2940 Scsi Controller ( no -ide drives) Andrew Bielecki -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Kelley Terry Sent: Monday, August 06, 2001 3:42 PM To: [EMAIL PROTECTED] Subject: Re: [Cooker] ReiserFS killed my /var On Mon, 6 Aug 2001 11:09:45 -0600 Vincent Danen [EMAIL PROTECTED] wrote: On Mon Aug 06, 2001 at 04:07:18PM +0200, Stefan Siegel wrote: This is really odd... I've been using reiserfs almost exclusively since 7.2 on all my machines. The only partition that I don't use it for is /var because I run qmail on all my machines and there was an issue with qmail and reiserfs in the past (no idea if it's fixed, but /var being ext2 is a matter of habit now). Ditto here. I've had nothing but reiser since on my machine since 7.2 and a few power outages and no problems. -- Kelley Terry [EMAIL PROTECTED] It said use windows 95 or better so I loaded linux! In a world without walls or fences who needs windows or gates?
Re: [Cooker] ReiserFS killed my /var
On Monday 06 August 2001 10:09 am, you wrote: On Mon Aug 06, 2001 at 04:07:18PM +0200, Stefan Siegel wrote: This is really odd... I've been using reiserfs almost exclusively since 7.2 on all my machines. The only partition that I don't use it for is /var because I run qmail on all my machines and there was an issue with qmail and reiserfs in the past (no idea if it's fixed, but /var being ext2 is a matter of habit now). Anyways, I've had a number of power outtages despite having UPS's with the systems all of a sudden going down. I may have lost the odd file occassionally, but certainly never an entire partition. Considering this is on about 5 machines all showing the same thing, I'm a little hesitant to think it's reiserfs *alone*. Maybe there is something with the files on /var that causes this problem (again, maybe if I had reiserfs /var partitions I would have experienced the same thing, I don't know). I do know that I have never lost more than a handful of files and I've got many reiserfs partitions across various systems. Same here. I did loose two partitions on completetly different hard drives but tracked it down to hard drive controler failure. The only thing I can say about the failure is that the ext2 partitions survived it but reiser didn't However I run reiser on almost everything and this is the first time I've had a problem.
Many success stories (was: RE: [Cooker] ReiserFS killed my /var)
Folks, you all miss the point. Nobody argues that ReiserFS has been used with success by many people. The problem is that if something goes wrong, Reiser gives very few chances to recover compared with other file systems (*especially* non-journaling). -andrej
Re: [Cooker] ReiserFS killed my /var
During the bombing raid on Mon, 6 Aug 2001 11:09:45 -0600, Vincent Danen was heard mumbling in fear: On Mon Aug 06, 2001 at 04:07:18PM +0200, Stefan Siegel wrote: thesis on it ... Since then I don't trust ReiserFS = I wouldn't install it currently on a critical system, only on playgounds for data loosing The question is whether it ameliorates mean crash recovery on critical systems. Nornally journaling would go in the right direction, but there are still potential problems. Any filesystem you're using, you'll still want to do periodic saves on tape devices, or even replication. Sure in theory you are right. But ReiserFS don't seem to be stable enough (atleast for me) as on the _FIRST_ crash (Power failure on the whole PC) I had, it shrederd my data. ext2 sometime eats data too, I know - but only files - at least on my side it never killed whole partitions. I am hoping for ext3 as it works on top of ext2 afaik. This is really odd... I've been using reiserfs almost exclusively since 7.2 on all my machines. The only partition that I don't use it Count me in as a happy reiserfs user, since somewhere in the 7.1 appearance (hand-patched into 2 7.0 servers and 3 workstations)...only had one problem, and that actually was due to the infamous VIA bug + the Reiserfs bug of a while back (chipset optimizations bug, I believe) that we tracked here in cooker...but no problems at all since nor before. And back then (it was during the road to 8.0) it was file corruption, never partition corruption...only thing that insisted on dieing was the rpm database whenever I installed 2 or more files at a time. Anyways, I've had a number of power outtages despite having UPS's with the systems all of a sudden going down. I may have lost the odd file My workstation has hadwell...enough power outtages...without any problems. occassionally, but certainly never an entire partition. Considering this is on about 5 machines all showing the same thing, I'm a little hesitant to think it's reiserfs *alone*. Maybe there is something with the files on /var that causes this problem (again, maybe if I had reiserfs /var partitions I would have experienced the same thing, I don't know). Nop, nothing to do with /var...I have everything under reiserfs and no problems at allso...I don't think it's a reiserfs+/var problem, but something else...maybe something like the qmail bug you mentioned. I do know that I have never lost more than a handful of files and I've got many reiserfs partitions across various systems. Counting 32 reiserfs partitions under my hand, most of them under stock mandrake kernels (have a suse box somewhere) Vox -- Pain is the gift of the gods, and I'm the one they chose as their messenger For info on safety in the BDSM lifestyle http://www.the-vox.com Think of the Linux community as a niche economy isolated by its beliefs. Kind of like the Amish, except that our religion requires us to use _higher_ technology than everyone else. -- Donald B. Marti Jr. Vox populi, vox deii