[reiserfs-list] Re: Kernel 2.4.7 Released! Any Updates?
On Sat, 21 Jul 2001, Manuel Krause wrote: > Did fix for "unlinking of opened files" > >(ftp.namesys.com/pub/reiserfs-for-2.4/2.4.6.pending/2.4.6-unlink-truncate-rename-rmdir.dif.bz2) Doesn't look like it, the patch mostly still applies, with just 1 reject. Probably not much work to fix up. > and Gergely Tamas' update make it in, too?! Don't know, haven't got that patch to hand. regards, Dave. -- | Dave Jones.http://www.suse.de/~davej | SuSE Labs
Re: [reiserfs-list] Kernel 2.4.7 Released! Any Updates?
On 07/21/2001 04:50 PM, Dirk Mueller wrote: > On Sam, 21 Jul 2001, Daniel wrote: > > >>Hey the subject says it all... ^ ^ ^ >> > > everything important except quota support is in. > > > Dirk > Did fix for "unlinking of opened files" (ftp.namesys.com/pub/reiserfs-for-2.4/2.4.6.pending/2.4.6-unlink-truncate-rename-rmdir.dif.bz2) and Gergely Tamas' update make it in, too?! That would be very very nice! Manuel
Re: [reiserfs-list] Kernel 2.4.7 Released! Any Updates?
On Sam, 21 Jul 2001, Daniel wrote: > Hey the subject says it all... ^ ^ ^ everything important except quota support is in. Dirk
Re: [reiserfs-list] Kernel 2.4.7 Released! Any Updates?
On Sat, 21 Jul 2001 19:00:25 +1000 "Daniel" <[EMAIL PROTECTED]> wrote: > Hey the subject says it all... ^ ^ ^ > ftp://ftp.us.kernel.org/pub/linux/kernel/v2.4/ChangeLog-2.4.7 <--snip--> -pre8: - Chris Mason: reiserfs update - Paul Mackerras: PPC updates (softirq) - Kai Germaschewski: ISDN updates - various: workaround for cpuid inline asm problem with egcs-2.91.66 <--snip--> -- OpenPGP key available from http://frodo.net.dhis.org/GnuPG/sjohnson.asc PGP signature
[reiserfs-list] Superblock damaged - How to calculate root block?
After a system crash i lost 2 drives. One could be recovered with reiserfsck (Suse 7.0 Versions), the other failed. Even --rebuild-tree failed. After this i tried the the newest versions: --rebuild-sb gives the following <-reiserfsck, 2001-> reiserfsprogs 3.x.0k-pre9 could not setrlimit: Invalid argument Will check su and rebuild if it is needed Will put log info to 'rs-rbtree4.log' Do you want to run this program?[N/Yes] (note need to type Yes):Yes 3276k will be used pass1.c 875 rebuild_sb Not ready Aborted debugreiserfs gives the following: Block 16 (0x700) contains super block of format 3.5 with standart journal Block count: 12485648 Bitmap number: 382 Blocksize: 4096 Free blocks: 12485648 Root block: 4294967295 Tree height: 65535 Hash function used to sort names: "tea" Objectid map size 2, max 1004 Journal parameters: Device [0x0] Magic [0x0] Size 8193 (including journal header) (first block 18) Max transaction length 0 Max batch size 0 Max commit age 0 Spase reserved by journal: 0 Correctness checked after mount 0 Fsck field 0xfaaf I would be glad to have a drive with that root block number ;-). The question is, how to find the correct value and what's the offset and value in sb to correct this manually? Some additional questions: As far as i understand, reiserfsck does not depend on a kernel with reiserfs. Is this correct? Do you have a documentation like XFS has about design and structures of ReiserFS? -- [EMAIL PROTECTED]
[reiserfs-list] My drive has been nulled-out after reiserfsck --rebuild-sb and --rebuild-tree !
Ok. Heres the story. Recently I reinstalled Slack 8 and upgraded my linuxbox to reiserfs. (installed new hdd, copied files over to it, and back after reformatting drives with reiser) and today my slightly failing harddrive caught up with me. The old ext2 format the drives had, actually checked the hdds for bad sectors, and marked the sectors bad when it found them. I thought it odd that even telling slack8's install to 'check for bad blocks' that it went on its merry way, did not read-verify the drive, and continued to make the superblock(s) and journal. That was 6 days ago. Now tonight I wake up to find my linuxbox in a state of 'drive cannot be read'. The procfs failing to show a 'top' or 'ps', and DMA Seek type errors all thru the dmesg. I think, ok, e2fsck can check for bad sectors, why not reiserfsck. I ran both reiserfsck --rebuild-tree and --rebuild-sb on the partition in question, and they kind of didnt work right. --rebuild-sb didnt complain (as far as my memory of it goes, i was slightly panic stricken at the time) but --rebuild-tree failed almost as soon as it started. (Turns out that sector 47205 or something close on the drive was dead, and that was around 25 megs into it.) I tried backing the drive up at this point (dd if=/dev/hdc of=/root/wd6.4g) but it errored reading and dd failed. any Mount attempt at this point was met with a kernel error, and the mount segfaulted. Since the system wouldnt cleanly shut down, more than once I was forced with hitting the power button to shut it off. This coupled with my hdparm'ing to turn on the write cache probably screwed something over. Yes, i know it was stupid, but when I did it 2 days ago I had no idea that the drive would give up its ghost as i assumed wrongly that any bad sectors were marked 'unusable' or the like. So after exhausting every option I could think of within linux I had to go find my WD data lifeguard disk. After I ran it on the drive and all 19 bad sectors were remapped, I ran reiserfsck with --rebuild-sb and --rebuild-tree options again. I didnt see any errors but it read every block and totally rebuilt everything. After that, it mounted fine. However I started noticing files that wouldnt work. Upon analysing them with a HEX editor, i found out that the files or parts of files were NULLed out. The drive in question is currently mounted as /home/ and a lot of important files are at risk. Is there any hope of recovery of the files? or am I truly 'fscked' Just looking at my personal homedir, a -ton- of my files have been replaced with null values. HELP Mike -- Mike Hodson <[EMAIL PROTECTED]>
[reiserfs-list] Kernel 2.4.7 Released! Any Updates?
Hey the subject says it all... ^ ^ ^