[opensuse] ReiserFS partitions recognized as Ext3. How to recover?
Hi, I have strange problem with my home OpenSUSE 11.0 machine. Machine has dual boot configuration - Win2000 (not used for long time) and OpenSuSE 11.0. HD is Maxtore, 60 GB. The first partition is primary FAT, then I have Extended partition, and inside this I have another FAT partition, then linux swap (/dev/sda6) and 2 reiserfs partitions - / (dev/sda7) and /home (/dev/sda8). Something happened yesterday. One thing was power failure, but I can hardly explain observed behavior by physical problem. The disk was OK, smart monitor never complained. When I turned computer on, the system reported that the sda7 is Ext3 FS, tried to check file system, reported inconsistency and dropped to maintenance mode. sda8 is also recognized as Ext3. In maintenance mode, / looks normal. I can check files and everything is there.Commands and utilities from this disk work. When I manually mount /home - all files are there and look OK. Windows is working OK. Using some rescue tools I was able to copy home directory and all important information from / to different computer (no errors reported while copying files). Now I tried reiserfsck on sda8 and it reported no superblock present. I tried to re-create it and re-check and reiserfsck now complained that there is bad root block and recommended rebuilding tree. But rebuilding tree failed and the message was that I probably moved the beginning of the partition (I did not). Partition table looks OK: Disk /dev/sda: 61.4 GB, 61492838400 bytes 255 heads, 63 sectors/track, 7476 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Disk identifier: 0x000e77df Device Boot Start End Blocks Id System /dev/sda1 * 1 2422 19454683+ c W95 FAT32 (LBA) Partition 1 does not end on cylinder boundary. /dev/sda2 2423 7476 40596255 f W95 Ext'd (LBA) /dev/sda5 2423 4407 15936480 b W95 FAT32 /dev/sda6 4407 4537 1052194+ 82 Linux swap / Solaris /dev/sda7 4538 6007 11807743+ 83 Linux /dev/sda8 6008 7476 11799711 83 Linux Could somebody explain how it could be that files themselves are accessible and correct, while the type of file system is wrong? How the system recognizes it as Ext3, while it was Reiser? reiserfsck recommented to zero the newly created superblock and to try moving the start of partition and then search for superblock. But I definitely had not changed the partitions... Of course I could go and re-install, but I would prefer to investigate further. Any advice? -- Mark Goldstein -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Friday, 2010-02-05 at 21:36 +0200, Mark Goldstein wrote:
I have strange problem with my home OpenSUSE 11.0 machine.
Machine has dual boot configuration - Win2000 (not used for long time) and OpenSuSE 11.0. HD is Maxtore, 60 GB. The first partition is primary FAT, then I have Extended partition, and inside this I have another FAT partition, then linux swap (/dev/sda6) and 2 reiserfs partitions - / (dev/sda7) and /home (/dev/sda8).
Something happened yesterday. One thing was power failure, but I can hardly explain observed behavior by physical problem. The disk was OK, smart monitor never complained. When I turned computer on, the system reported that the sda7 is Ext3 FS, tried to check file system, reported inconsistency and dropped to maintenance mode. sda8 is also recognized as Ext3.
In maintenance mode, / looks normal. I can check files and everything is there.Commands and utilities from this disk work. When I manually mount /home - all files are there and look OK. Windows is working OK.
Using some rescue tools I was able to copy home directory and all important information from / to different computer (no errors reported while copying files).
I assume that dev/sda7 and dev/sda8 are mounted at this moment in time. If you issue the command "mount" and you see those partitions reported as ext3, and you can copy the files, and they are not corrupted, there is no doubt: those filesystems are ext3, not reiserfs - regardless what your memory says >:-)
Now I tried reiserfsck on sda8 and it reported no superblock present. I tried to re-create it and re-check and reiserfsck now complained
If you try to repair them as reiserfs, you will utterly destroy them, as they are not reiserfs. If you want them as reiser, reformat, and copy the data back. Unless you have more disks or more partitions and you didn't know. Or a filesystem loop mounted on top of another, but that needs some doing. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) iEUEARECAAYFAktse7wACgkQtTMYHG2NR9W2FgCXboVwGDqrnlaRbMFyuwe9b5wQ cACfcoBE4Q+/sr5G+YddX9PZymakiOc= =CaLu -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Fri, Feb 5, 2010 at 10:12 PM, Carlos E. R.
On Friday, 2010-02-05 at 21:36 +0200, Mark Goldstein wrote:
...
Something happened yesterday. One thing was power failure, but I can hardly explain observed behavior by physical problem. The disk was OK, smart monitor never complained. When I turned computer on, the system reported that the sda7 is Ext3 FS, tried to check file system, reported inconsistency and dropped to maintenance mode. sda8 is also recognized as Ext3.
In maintenance mode, / looks normal. I can check files and everything is there.Commands and utilities from this disk work. When I manually mount /home - all files are there and look OK. Windows is working OK.
Using some rescue tools I was able to copy home directory and all important information from / to different computer (no errors reported while copying files).
I assume that dev/sda7 and dev/sda8 are mounted at this moment in time. If you issue the command "mount" and you see those partitions reported as ext3, and you can copy the files, and they are not corrupted, there is no doubt: those filesystems are ext3, not reiserfs - regardless what your memory says
:-)
I would not argue about my memory, it fails sometimes :-(. But I never use ext3 after some years ago (while I was using RH9.0 with ext3) my HD (the same one BTW) got corrupted. I switched then to ReiserFS and use only it on all my Linux boxes since.
Now I tried reiserfsck on sda8 and it reported no superblock present. I tried to re-create it and re-check and reiserfsck now complained
If you try to repair them as reiserfs, you will utterly destroy them, as they are not reiserfs. If you want them as reiser, reformat, and copy the data back.
It's a fact that even after I tried to re-create reiser superblock on /dev/sda8 and Gparted now shows this partition as reiser, mount indicates it as ext3 and I can still see the files. I'm puzzled, to say the least... -- Mark Goldstein -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Friday, 2010-02-05 at 22:36 +0200, Mark Goldstein wrote:
On Fri, Feb 5, 2010 at 10:12 PM, Carlos E. R.> <> wrote:
I assume that dev/sda7 and dev/sda8 are mounted at this moment in time. If you issue the command "mount" and you see those partitions reported as ext3, and you can copy the files, and they are not corrupted, there is no doubt: those filesystems are ext3, not reiserfs - regardless what your memory says
:-)
I would not argue about my memory, it fails sometimes :-(. But I never use ext3 after some years ago (while I was using RH9.0 with ext3) my HD (the same one BTW) got corrupted. I switched then to ReiserFS and use only it on all my Linux boxes since.
I like reiserfs, but I have been bitten by all types of filesystems. None is fully reliable. Currently I tend to use ext3 for the root partition, and reiserfs or xfs for data partitions. On single partition systems I use reiserfs with a separate boot as ext2.
Now I tried reiserfsck on sda8 and it reported no superblock present. I tried to re-create it and re-check and reiserfsck now complained
If you try to repair them as reiserfs, you will utterly destroy them, as they are not reiserfs. If you want them as reiser, reformat, and copy the data back.
It's a fact that even after I tried to re-create reiser superblock on /dev/sda8 and Gparted now shows this partition as reiser, mount indicates it as ext3 and I can still see the files. I'm puzzled, to say the least...
I don't know why gparted says that. You could also try: file -s /dev/sda7 Perhaps there is something at the start of the partition that "look" like reiserfs. I would backup all the data, then recreate the filesystem entirely (on those two partitions), by using "mkreiserfs" (don't forget to add a label). - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) iEYEARECAAYFAktskxcACgkQtTMYHG2NR9UDhgCcDfCGWoSoRVWpGU611YgpLbUW uDoAn1oOsdc2r/XfaW24fdqnFo5R6u/v =xLSE -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Fri, Feb 5, 2010 at 11:52 PM, Carlos E. R.
On Friday, 2010-02-05 at 22:36 +0200, Mark Goldstein wrote:
On Fri, Feb 5, 2010 at 10:12 PM, Carlos E. R.> <> wrote:
I assume that dev/sda7 and dev/sda8 are mounted at this moment in time. If you issue the command "mount" and you see those partitions reported as ext3, and you can copy the files, and they are not corrupted, there is no doubt: those filesystems are ext3, not reiserfs - regardless what your memory says
:-)
I would not argue about my memory, it fails sometimes :-(. But I never use ext3 after some years ago (while I was using RH9.0 with ext3) my HD (the same one BTW) got corrupted. I switched then to ReiserFS and use only it on all my Linux boxes since.
I like reiserfs, but I have been bitten by all types of filesystems. None is fully reliable. Currently I tend to use ext3 for the root partition, and reiserfs or xfs for data partitions. On single partition systems I use reiserfs with a separate boot as ext2.
Are you assuming that root partition is changing less then data partition? Or vise versa? That is, do you want ext3 to do more journaling or less? My impression from ext3 corruption was that it happened during one of ext3 hidden activities - disk has been working hard, though I did nothing at the time. Of course it does not prove anything, just my gut feeling. I think I should anyway reconsider my partitioning approach. ...
I don't know why gparted says that. You could also try:
file -s /dev/sda7
Perhaps there is something at the start of the partition that "look" like reiserfs. I would backup all the data, then recreate the filesystem entirely (on those two partitions), by using "mkreiserfs" (don't forget to add a label).
I guess, this is because superblock of ReiserFS has kind of signature (ReIsEr2FS). When I tried re-creation of superblock, the signature was written there. file -s on this partition also indicates "ReiserFR V3.6 now. But mount still considers it as ext3. My impression is that during power down something got changed, but very locally, probably some offset. So that real reiser fs superblock is not where it should be (64K from the beginning of partition) and for that reason file system is not correctly recognized. Where is my old diskette with Norton Disk Editor :( ? Regards, -- Mark Goldstein -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 02/06/2010 10:32 AM, Mark Goldstein wrote:
On Fri, Feb 5, 2010 at 11:52 PM, Carlos E. R. <> wrote:
I like reiserfs, but I have been bitten by all types of filesystems. None is fully reliable. Currently I tend to use ext3 for the root partition, and reiserfs or xfs for data partitions. On single partition systems I use reiserfs with a separate boot as ext2.
Are you assuming that root partition is changing less then data partition? Or vise versa? That is, do you want ext3 to do more journaling or less? My impression from ext3 corruption was that it happened during one of ext3 hidden activities - disk has been working hard, though I did nothing at the time. Of course it does not prove anything, just my gut feeling. I think I should anyway reconsider my partitioning approach. ...
A root partition containing the system, not home, usually changes less, yes. Usually, because a lot of things get written to /var. The thing is, I trust a bit more ext3 recovery tools than reiserfs. I prefer having my system up, then recover data with system up. And, bugs related to reiserfs are not solved or very slowly (for instance, hibernation and restoring from reiser fails, blame grub)
I don't know why gparted says that. You could also try:
file -s /dev/sda7
Perhaps there is something at the start of the partition that "look" like reiserfs. I would backup all the data, then recreate the filesystem entirely (on those two partitions), by using "mkreiserfs" (don't forget to add a label).
I guess, this is because superblock of ReiserFS has kind of signature (ReIsEr2FS). When I tried re-creation of superblock, the signature was written there. file -s on this partition also indicates "ReiserFR V3.6 now. But mount still considers it as ext3.
My impression is that during power down something got changed, but very locally, probably some offset. So that real reiser fs superblock is not where it should be (64K from the beginning of partition) and for that reason file system is not correctly recognized. Where is my old diskette with Norton Disk Editor :( ?
No, the structure of a reiserfs partition is radically different from an ext3 partition. There is no way that changing a few blocks can change one to the other: the entire disk has to be rewritten. There is no way a power failure or any glitch can do the conversion. Actually, there is no tool to convert partition type from ext3 to reiser or viceversa except an entire rewrite. A backup restore using dd, for example, would do it, if the backup was ext3. Some tools may thinkk it is reiserfs because they only look at the beginning of the partition. - -- Cheers / Saludos, Carlos E. R. (from 11.2 x86_64 "Emerald" GM (Elessar)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/ iEYEARECAAYFAkttWyAACgkQU92UU+smfQXZBwCeNIhb/wNBPVlM/V1xYY+S8UfK bsIAn15OiY3985ffY5zzuAugxfp8IJt2 =s6Sm -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 02/06/2010 10:32 AM, Mark Goldstein wrote:
On Fri, Feb 5, 2010 at 11:52 PM, Carlos E. R. <> wrote:
I like reiserfs, but I have been bitten by all types of filesystems. None is fully reliable. Currently I tend to use ext3 for the root partition, and reiserfs or xfs for data partitions. On single partition systems I use reiserfs with a separate boot as ext2.
Are you assuming that root partition is changing less then data partition? Or vise versa? That is, do you want ext3 to do more journaling or less? My impression from ext3 corruption was that it happened during one of ext3 hidden activities - disk has been working hard, though I did nothing at the time. Of course it does not prove anything, just my gut feeling. I think I should anyway reconsider my partitioning approach. ...
A root partition containing the system, not home, usually changes less, yes. Usually, because a lot of things get written to /var. The thing is, I trust a bit more ext3 recovery tools than reiserfs. I prefer having my system up, then recover data with system up. And, bugs related to reiserfs are not solved or very slowly (for instance, hibernation and restoring from reiser fails, blame grub)
I don't know why gparted says that. You could also try:
file -s /dev/sda7
Perhaps there is something at the start of the partition that "look" like reiserfs. I would backup all the data, then recreate the filesystem entirely (on those two partitions), by using "mkreiserfs" (don't forget to add a label).
I guess, this is because superblock of ReiserFS has kind of signature (ReIsEr2FS). When I tried re-creation of superblock, the signature was written there. file -s on this partition also indicates "ReiserFR V3.6 now. But mount still considers it as ext3.
My impression is that during power down something got changed, but very locally, probably some offset. So that real reiser fs superblock is not where it should be (64K from the beginning of partition) and for that reason file system is not correctly recognized. Where is my old diskette with Norton Disk Editor :( ?
No, the structure of a reiserfs partition is radically different from an ext3 partition. There is no way that changing a few blocks can change one to the other: the entire disk has to be rewritten. There is no way a power failure or any glitch can do the conversion. Actually, there is no tool to convert partition type from ext3 to reiser or viceversa except an entire rewrite. A backup restore using dd, for example, would do it, if the backup was ext3. Some tools may thinkk it is reiserfs because they only look at the beginning of the partition. - -- Cheers / Saludos, Carlos E. R. (from 11.2 x86_64 "Emerald" GM (Elessar)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/ iEYEARECAAYFAkttWykACgkQU92UU+smfQXP4QCfRgD+a2/6Rqb1cWC5ydVEC2vC BeEAnRFKqI46jYK8Sa6xhslzkqSgi7CP =i9O6 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Fri, 2010-02-05 at 21:12 +0100, Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Friday, 2010-02-05 at 21:36 +0200, Mark Goldstein wrote:
I have strange problem with my home OpenSUSE 11.0 machine.
Machine has dual boot configuration - Win2000 (not used for long time) and OpenSuSE 11.0. HD is Maxtore, 60 GB. The first partition is primary FAT, then I have Extended partition, and inside this I have another FAT partition, then linux swap (/dev/sda6) and 2 reiserfs partitions - / (dev/sda7) and /home (/dev/sda8).
Something happened yesterday. One thing was power failure, but I can hardly explain observed behavior by physical problem. The disk was OK, smart monitor never complained. When I turned computer on, the system reported that the sda7 is Ext3 FS, tried to check file system, reported inconsistency and dropped to maintenance mode. sda8 is also recognized as Ext3.
In maintenance mode, / looks normal. I can check files and everything is there.Commands and utilities from this disk work. When I manually mount /home - all files are there and look OK. Windows is working OK.
Using some rescue tools I was able to copy home directory and all important information from / to different computer (no errors reported while copying files).
I assume that dev/sda7 and dev/sda8 are mounted at this moment in time. If you issue the command "mount" and you see those partitions reported as ext3, and you can copy the files, and they are not corrupted, there is no doubt: those filesystems are ext3, not reiserfs - regardless what your memory says >:-)
Now I tried reiserfsck on sda8 and it reported no superblock present. I tried to re-create it and re-check and reiserfsck now complained
If you try to repair them as reiserfs, you will utterly destroy them, as they are not reiserfs. If you want them as reiser, reformat, and copy the data back.
Unless you have more disks or more partitions and you didn't know. Or a filesystem loop mounted on top of another, but that needs some doing.
Wonder what's in /etc/fstab. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, Feb 6, 2010 at 1:20 PM, Mike McMullin
On Fri, 2010-02-05 at 21:12 +0100, Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Friday, 2010-02-05 at 21:36 +0200, Mark Goldstein wrote:
...
Now I tried reiserfsck on sda8 and it reported no superblock present. I tried to re-create it and re-check and reiserfsck now complained
If you try to repair them as reiserfs, you will utterly destroy them, as they are not reiserfs. If you want them as reiser, reformat, and copy the data back.
Unless you have more disks or more partitions and you didn't know. Or a filesystem loop mounted on top of another, but that needs some doing.
Wonder what's in /etc/fstab.
Now you've got me ashamed... /etc/fstab says it's ext3. And it is dated by June 2008 - when I installed 11.0. It looks like not just memory blackout, but rather sclerosis... I also looked through /var/log/messages and saw this fragment: Feb 3 18:36:37 dadpc kernel: Pid: 15, comm: kswapd0 Tainted: P N (2.6.25.20-0.5-pae #1) Feb 3 18:36:37 dadpc kernel: EIP: 0060:[<c019622d>] EFLAGS: 00010283 CPU: 0 Feb 3 18:36:37 dadpc kernel: EIP is at free_buffer_head+0xd/0x31 Feb 3 18:36:37 dadpc kernel: EAX: d94851f8 EBX: d94851d0 ECX: c1323480 EDX: d94851d0 Feb 3 18:36:37 dadpc kernel: ESI: 00000001 EDI: df025564 EBP: de45fd74 ESP: de45fd74 Feb 3 18:36:37 dadpc kernel: DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068 Feb 3 18:36:37 dadpc kernel: Process kswapd0 (pid: 15, ti=de45e000 task=de440120 task.ti=de45e000) Feb 3 18:36:37 dadpc kernel: Stack: de45fd8c c01966d9 d94851d0 d94851d0 df92a400 df92a400 de45fdb4 e15846ef Feb 3 18:36:37 dadpc kernel: 000000d0 c1323480 d94851d0 00000000 df92a4b8 c1323480 df92a400 000000d0 Feb 3 18:36:37 dadpc kernel: de45fdc8 e15c9b69 c1323480 000000d0 00000001 de45fdd8 c01598a7 c1323480 Feb 3 18:36:37 dadpc kernel: Call Trace: Feb 3 18:36:37 dadpc kernel: [<c01966d9>] try_to_free_buffers+0x70/0x7f Feb 3 18:36:37 dadpc kernel: [<e15846ef>] journal_try_to_free_buffers+0xfb/0x144 [jbd] Feb 3 18:36:37 dadpc kernel: [<e15c9b69>] ext3_releasepage+0x4b/0x54 [ext3] Feb 3 18:36:37 dadpc kernel: [<c01598a7>] try_to_release_page+0x32/0x3f Feb 3 18:36:37 dadpc kernel: [<c01627ac>] shrink_page_list+0x46f/0x556 Feb 3 18:36:37 dadpc kernel: [<c01629bf>] shrink_inactive_list+0x12c/0x33f Feb 3 18:36:37 dadpc kernel: [<c0162c9d>] shrink_zone+0xcb/0xea Feb 3 18:36:37 dadpc kernel: [<c016316e>] kswapd+0x380/0x4ba Feb 3 18:36:37 dadpc kernel: [<c0135fdf>] kthread+0x3b/0x61 Feb 3 18:36:37 dadpc kernel: [<c0106d37>] kernel_thread_helper+03,x7/0x10 Feb 3 18:36:37 dadpc kernel: ======================= So this is when it happened, most probable! How I made it ext3 - I do not know. Apologies for misleading (If I misled anybody except myself), Regards, -- Mark Goldstein -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 02/06/2010 01:25 PM, Mark Goldstein wrote:
On Sat, Feb 6, 2010 at 1:20 PM, Mike McMullin
wrote:
Unless you have more disks or more partitions and you didn't know. Or a filesystem loop mounted on top of another, but that needs some doing.
Wonder what's in /etc/fstab.
Now you've got me ashamed... /etc/fstab says it's ext3. And it is dated by June 2008 - when I installed 11.0. It looks like not just memory blackout, but rather sclerosis...
Don't worry. It was interesting :-) But now, you should really reformat that partition (as reiserfs if you like), and restore the data from a backup. Because with all this fiddling the partition could be damaged in the structure and give you a surprise one day - so better prevent :-) Though... grub is a problem when restoring. I have a problem with mine just now... - -- Cheers / Saludos, Carlos E. R. (from 11.2 x86_64 "Emerald" GM (Elessar)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/ iEYEARECAAYFAkttbxoACgkQU92UU+smfQXfagCgjgOq/1bXJ9i+2C/kR9q9bbOg +b0AoIGs/UD8+Y6v8CTyDvILNNr7l7Dt =azT6 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, Feb 6, 2010 at 3:31 PM, Carlos E. R.
On 02/06/2010 01:25 PM, Mark Goldstein wrote:
On Sat, Feb 6, 2010 at 1:20 PM, Mike McMullin
wrote: Unless you have more disks or more partitions and you didn't know. Or a filesystem loop mounted on top of another, but that needs some doing.
Wonder what's in /etc/fstab.
Now you've got me ashamed... /etc/fstab says it's ext3. And it is dated by June 2008 - when I installed 11.0. It looks like not just memory blackout, but rather sclerosis...
Don't worry. It was interesting :-)
But now, you should really reformat that partition (as reiserfs if you like), and restore the data from a backup. Because with all this fiddling the partition could be damaged in the structure and give you a surprise one day - so better prevent :-)
Unfortunately, you are right again :-( fsck for ext2 recovered the errors in partitions, and the system runs again ... mostly. zypper and yast crash now (and I did not play with reiserfschk on root partition, only on home). When it comes to "reading installed packages", zypper throws a number of "can't decompress" and then segfaults. rpm can read its db (at least rpm -qai works correctly). Looks like I need to at least new HD (and probably upgrade the computer - they are both "veterans" from Dec 2002). Regards -- Mark Goldstein -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, Feb 6, 2010 at 9:11 PM, Mark Goldstein
On Sat, Feb 6, 2010 at 3:31 PM, Carlos E. R.
wrote: On 02/06/2010 01:25 PM, Mark Goldstein wrote:
On Sat, Feb 6, 2010 at 1:20 PM, Mike McMullin
wrote: Unless you have more disks or more partitions and you didn't know. Or a filesystem loop mounted on top of another, but that needs some doing.
Wonder what's in /etc/fstab.
Now you've got me ashamed... /etc/fstab says it's ext3. And it is dated by June 2008 - when I installed 11.0. It looks like not just memory blackout, but rather sclerosis...
Don't worry. It was interesting :-)
But now, you should really reformat that partition (as reiserfs if you like), and restore the data from a backup. Because with all this fiddling the partition could be damaged in the structure and give you a surprise one day - so better prevent :-)
Unfortunately, you are right again :-( fsck for ext2 recovered the errors in partitions, and the system runs again ... mostly. zypper and yast crash now (and I did not play with reiserfschk on root partition, only on home). When it comes to "reading installed packages", zypper throws a number of "can't decompress" and then segfaults. rpm can read its db (at least rpm -qai works correctly).
Removing of one of subdirectories in /vat/cache/zypp/solv resolved it (it was the last one mentioned in zypper log). Now updates seems working. -- Mark Goldstein -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sat, 2010-02-06 at 14:25 +0200, Mark Goldstein wrote:
On Sat, Feb 6, 2010 at 1:20 PM, Mike McMullin
wrote: On Fri, 2010-02-05 at 21:12 +0100, Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Friday, 2010-02-05 at 21:36 +0200, Mark Goldstein wrote:
...
Now I tried reiserfsck on sda8 and it reported no superblock present. I tried to re-create it and re-check and reiserfsck now complained
If you try to repair them as reiserfs, you will utterly destroy them, as they are not reiserfs. If you want them as reiser, reformat, and copy the data back.
Unless you have more disks or more partitions and you didn't know. Or a filesystem loop mounted on top of another, but that needs some doing.
Wonder what's in /etc/fstab.
Now you've got me ashamed... /etc/fstab says it's ext3. And it is dated by June 2008 - when I installed 11.0. It looks like not just memory blackout, but rather sclerosis...
I also looked through /var/log/messages and saw this fragment:
Feb 3 18:36:37 dadpc kernel: Pid: 15, comm: kswapd0 Tainted: P N (2.6.25.20-0.5-pae #1) Feb 3 18:36:37 dadpc kernel: EIP: 0060:[<c019622d>] EFLAGS: 00010283 CPU: 0 Feb 3 18:36:37 dadpc kernel: EIP is at free_buffer_head+0xd/0x31 Feb 3 18:36:37 dadpc kernel: EAX: d94851f8 EBX: d94851d0 ECX: c1323480 EDX: d94851d0 Feb 3 18:36:37 dadpc kernel: ESI: 00000001 EDI: df025564 EBP: de45fd74 ESP: de45fd74 Feb 3 18:36:37 dadpc kernel: DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068 Feb 3 18:36:37 dadpc kernel: Process kswapd0 (pid: 15, ti=de45e000 task=de440120 task.ti=de45e000) Feb 3 18:36:37 dadpc kernel: Stack: de45fd8c c01966d9 d94851d0 d94851d0 df92a400 df92a400 de45fdb4 e15846ef Feb 3 18:36:37 dadpc kernel: 000000d0 c1323480 d94851d0 00000000 df92a4b8 c1323480 df92a400 000000d0 Feb 3 18:36:37 dadpc kernel: de45fdc8 e15c9b69 c1323480 000000d0 00000001 de45fdd8 c01598a7 c1323480 Feb 3 18:36:37 dadpc kernel: Call Trace: Feb 3 18:36:37 dadpc kernel: [<c01966d9>] try_to_free_buffers+0x70/0x7f Feb 3 18:36:37 dadpc kernel: [<e15846ef>] journal_try_to_free_buffers+0xfb/0x144 [jbd] Feb 3 18:36:37 dadpc kernel: [<e15c9b69>] ext3_releasepage+0x4b/0x54 [ext3] Feb 3 18:36:37 dadpc kernel: [<c01598a7>] try_to_release_page+0x32/0x3f Feb 3 18:36:37 dadpc kernel: [<c01627ac>] shrink_page_list+0x46f/0x556 Feb 3 18:36:37 dadpc kernel: [<c01629bf>] shrink_inactive_list+0x12c/0x33f Feb 3 18:36:37 dadpc kernel: [<c0162c9d>] shrink_zone+0xcb/0xea Feb 3 18:36:37 dadpc kernel: [<c016316e>] kswapd+0x380/0x4ba Feb 3 18:36:37 dadpc kernel: [<c0135fdf>] kthread+0x3b/0x61 Feb 3 18:36:37 dadpc kernel: [<c0106d37>] kernel_thread_helper+03,x7/0x10 Feb 3 18:36:37 dadpc kernel: =======================
So this is when it happened, most probable! How I made it ext3 - I do not know.
Apologies for misleading (If I misled anybody except myself),
I have found out the hard way, that looking at the config/reference files tells me things about my system that I thought I knew but didn't, /etc/fstab is the most logical indicator of how things were setup/ought to be. So I asked. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sun, Feb 7, 2010 at 1:27 AM, Mike McMullin
On Sat, 2010-02-06 at 14:25 +0200, Mark Goldstein wrote:
On Sat, Feb 6, 2010 at 1:20 PM, Mike McMullin
wrote: On Fri, 2010-02-05 at 21:12 +0100, Carlos E. R. wrote: ... How I made it ext3 - I do not know.
Apologies for misleading (If I misled anybody except myself),
I have found out the hard way, that looking at the config/reference files tells me things about my system that I thought I knew but didn't, /etc/fstab is the most logical indicator of how things were setup/ought to be. So I asked.
Thank you Mike. Actually Carlos told me from the very beginning that it is not possible, but I was so absolute sure I did not use ext3... And indeed all other computers I was using were reiserfs. But not this one. Stupid me. Will check and re-check myself next time :-(. -- Mark Goldstein -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 02/07/2010 07:42 AM, Mark Goldstein wrote:
On Sun, Feb 7, 2010 at 1:27 AM, Mike McMullin
wrote: On Sat, 2010-02-06 at 14:25 +0200, Mark Goldstein wrote:
On Sat, Feb 6, 2010 at 1:20 PM, Mike McMullin
wrote: On Fri, 2010-02-05 at 21:12 +0100, Carlos E. R. wrote: ... How I made it ext3 - I do not know.
Apologies for misleading (If I misled anybody except myself),
I have found out the hard way, that looking at the config/reference files tells me things about my system that I thought I knew but didn't, /etc/fstab is the most logical indicator of how things were setup/ought to be. So I asked.
Thank you Mike. Actually Carlos told me from the very beginning that it is not possible, but I was so absolute sure I did not use ext3... And indeed all other computers I was using were reiserfs. But not this one. Stupid me. Will check and re-check myself next time :-(.
Well... it just happened to me. :-O I had just installed 11.2_x64 on another partition, formatted as reiserfs - I'm definitely sure of this. I recovered some data from a backup of a previous install. Then I filled the fstab from another copy, and, big mistake, rewrote the root entry as _ext3_. Rebooted. No error was reported. But... there were strange things. I copied them: mount said: /dev/sda9 on / type ext3 (rw,acl,user_xattr) fstab said: LABEL=a_test2 / ext3 acl,user_xattr 1 1 "file -s /dev/sda9" said: /dev/sda9: ReiserFS V3.6 boot.msg log said: <5>[ 3.104641] REISERFS (device sda9): found reiserfs format "3.6" with standard journal <5>[ 3.104652] REISERFS (device sda9): using ordered data mode <4>[ 3.104653] reiserfs: using flush barriers <5>[ 3.105080] REISERFS (device sda9): journal params: device sda9, size 8192, journal first block 18, max trans len 1024, max batch 900, <5>[ 3.105360] REISERFS (device sda9): checking transaction log (sda9) <5>[ 3.146242] REISERFS (device sda9): Using r5 hash to sort names Ie, it is and was a reiserfs partition, but mount wrongly reported that it was ext3. There is definitely a bug here. (I corrected the fstab and rebooted; things are correct now and no damage seems to have come out of it... (I did not try to fsck). I'll keep my fingers crossed) And I'm very, very surprissed :-( - -- Cheers / Saludos, Carlos E. R. (from 11.2 x86_64 "Emerald" GM (Elessar)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/ iEYEARECAAYFAktu0roACgkQU92UU+smfQWm2gCggas7MqQRT4ZWTX8ypAdbLs3c N6MAmgN+WDe1phHTzo795kQhdkeGHdBm =PLRP -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sun, Feb 7, 2010 at 4:48 PM, Carlos E. R.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 02/07/2010 07:42 AM, Mark Goldstein wrote:
On Sun, Feb 7, 2010 at 1:27 AM, Mike McMullin
wrote: On Sat, 2010-02-06 at 14:25 +0200, Mark Goldstein wrote:
On Sat, Feb 6, 2010 at 1:20 PM, Mike McMullin
wrote: On Fri, 2010-02-05 at 21:12 +0100, Carlos E. R. wrote: ... How I made it ext3 - I do not know.
Apologies for misleading (If I misled anybody except myself),
I have found out the hard way, that looking at the config/reference files tells me things about my system that I thought I knew but didn't, /etc/fstab is the most logical indicator of how things were setup/ought to be. So I asked.
Thank you Mike. Actually Carlos told me from the very beginning that it is not possible, but I was so absolute sure I did not use ext3... And indeed all other computers I was using were reiserfs. But not this one. Stupid me. Will check and re-check myself next time :-(.
Well... it just happened to me. :-O
I had just installed 11.2_x64 on another partition, formatted as reiserfs - I'm definitely sure of this. I recovered some data from a backup of a previous install. Then I filled the fstab from another copy, and, big mistake, rewrote the root entry as _ext3_. Rebooted.
No error was reported. But... there were strange things. I copied them:
mount said:
/dev/sda9 on / type ext3 (rw,acl,user_xattr)
fstab said:
LABEL=a_test2 / ext3 acl,user_xattr 1 1
"file -s /dev/sda9" said:
/dev/sda9: ReiserFS V3.6
boot.msg log said:
<5>[ 3.104641] REISERFS (device sda9): found reiserfs format "3.6" with standard journal <5>[ 3.104652] REISERFS (device sda9): using ordered data mode <4>[ 3.104653] reiserfs: using flush barriers <5>[ 3.105080] REISERFS (device sda9): journal params: device sda9, size 8192, journal first block 18, max trans len 1024, max batch 900, <5>[ 3.105360] REISERFS (device sda9): checking transaction log (sda9) <5>[ 3.146242] REISERFS (device sda9): Using r5 hash to sort names
Ie, it is and was a reiserfs partition, but mount wrongly reported that it was ext3. There is definitely a bug here.
(I corrected the fstab and rebooted; things are correct now and no damage seems to have come out of it... (I did not try to fsck). I'll keep my fingers crossed)
And I'm very, very surprissed :-(
So quick... I hope it is not an e-mail communicable infection :-( It looks like though there is a signature in the superblock and probably some other ways to find out the partition type, the mount takes the type specified in fstab without further checking. It should at least compare what is specified in fstab with the result of file -s. Regards, -- Mark Goldstein -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 02/07/2010 04:07 PM, Mark Goldstein wrote:
On Sun, Feb 7, 2010 at 4:48 PM, Carlos E. R. <> wrote:
...
And I'm very, very surprissed :-(
So quick... I hope it is not an e-mail communicable infection :-(
X'-) No, I'm tired, I have a headache, so I make mistakes on my computer when trying to correct others. It was my fault that I changed reiserfs by ext3 in the fstab file... thick thumbs.
It looks like though there is a signature in the superblock and probably some other ways to find out the partition type, the mount takes the type specified in fstab without further checking. It should at least compare what is specified in fstab with the result of file -s.
Well, mount actually mounted the partition correctly, as reiserfs. The problem is that it reports incorrectly that it was mounted as ext3, which can cause all sorts of problems if the user then fscks the partition using the wrong checker. - -- Cheers / Saludos, Carlos E. R. (from 11.2 x86_64 "Emerald" GM (Elessar)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/ iEYEARECAAYFAktu5x0ACgkQU92UU+smfQVb6gCeLLx06KYEWIVt3MLE9ngR5lYt eqIAoIlOh/ZfVChi/0kZ0wtqS7xsyuuF =gVpb -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Sun, 2010-02-07 at 17:15 +0100, Carlos E. R. wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 02/07/2010 04:07 PM, Mark Goldstein wrote:
On Sun, Feb 7, 2010 at 4:48 PM, Carlos E. R. <> wrote:
...
And I'm very, very surprissed :-(
So quick... I hope it is not an e-mail communicable infection :-(
X'-)
No, I'm tired, I have a headache, so I make mistakes on my computer when trying to correct others. It was my fault that I changed reiserfs by ext3 in the fstab file... thick thumbs.
It looks like though there is a signature in the superblock and probably some other ways to find out the partition type, the mount takes the type specified in fstab without further checking. It should at least compare what is specified in fstab with the result of file -s.
Well, mount actually mounted the partition correctly, as reiserfs. The problem is that it reports incorrectly that it was mounted as ext3, which can cause all sorts of problems if the user then fscks the partition using the wrong checker.
Curious Carlos, why ext3 over ReiserFS? I've been dodging ext? for ages. (My wife would shut the system down using the power bar on/off switch, and ext2 kept canning on me, ReiserFS tolerated her abuse quite well.) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2010-02-08 00:13, Mike McMullin wrote:
On Sun, 2010-02-07 at 17:15 +0100, Carlos E. R. wrote:
Curious Carlos, why ext3 over ReiserFS? I've been dodging ext? for ages. (My wife would shut the system down using the power bar on/off switch, and ext2 kept canning on me, ReiserFS tolerated her abuse quite well.)
No, the other way round. I had to reinstall (grub was broken), and used the chance to reformat as reiserfs instead. I like reiserfs. I don't know for how long I'll be able to keep using it, but... as I say, I like it. - -- Cheers / Saludos, Carlos E. R. (from 11.2 x86_64 "Emerald" GM (Elessar)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/ iEUEARECAAYFAktvVHAACgkQU92UU+smfQUDswCYnA2L8Wof0yD2Kjp0+UWn1x8a UgCff67LjxXsSo0Lizd6DyMM9KERfUo= =dr0Y -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2010-02-07 15:48, Carlos E. R. wrote:
Ie, it is and was a reiserfs partition, but mount wrongly reported that it was ext3. There is definitely a bug here.
Reported as "Bug 577798" - -- Cheers / Saludos, Carlos E. R. (from 11.2 x86_64 "Emerald" GM (Elessar)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.12 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/ iEYEARECAAYFAktvIZAACgkQU92UU+smfQX0GACcC9XqLTIZ1++cgzFjKpu7sr1H Aj4An3ghAB4KZO5DQNDbsNtr2AGZGkpx =DTyX -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (3)
-
Carlos E. R.
-
Mark Goldstein
-
Mike McMullin