zero-length partition - fsck.ext3
SUSE 9.0 100g WD IDE drive 133 bus /dev/hde IDE DMA 100 I have just encountered some file corruption, I think. SUSE9 hung up and, after a 20 minute wait, I rebooted. I got notice that /dev/hde8 needed manual run of fsck. wahoo:~ # fsck.ext3 -yv /dev/hde8 e2fsck 1.34 (25-Jul-2003) fsck.ext3: Attempt to read block from filesystem resulted in short read while trying to open /dev/hde8 Could this be a zero-length partition? wahoo:~ # fdisk -l /dev/hde Disk /dev/hde: 100.0 GB, 100030242816 bytes 255 heads, 63 sectors/track, 12161 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Device Boot Start End Blocks Id System /dev/hde1 1 6867 55158736+ 5 Extended /dev/hde5 1 1306 10486161 83 Linux /dev/hde6 1306 2611 10486192+ 83 Linux /dev/hde7 2611 5043 19535008+ 83 Linux /dev/hde8 5043 6867 14651248+ 83 Linux Expert command (m for help): m Command action e list extended partitions Expert command (m for help): e Disk /dev/hde: 255 heads, 63 sectors, 12161 cylinders Nr AF Hd Sec Cyl Hd Sec Cyl Start Size ID 1 00 1 1 0 254 63 1023 63110317473 05 5 00 254 63 1023 254 63 1023 20972385 20972448 05 6 00 254 63 1023 254 63 1023 41944833 39070080 05 7 00 254 63 1023 254 63 1023 81014913 29302560 05 8 00 0 0 0 0 0 0 0 0 00 Which appears to show that the extended partition information is missing. Google and suse db provide no (recent) answers, but many other similar experiences. Is there and answer or have I lost the partition? Would the "Expert command" to "f fix partition order" have any helpful result? Is there hope? -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org
The Friday 2004-02-06 at 13:36 -0500, Patrick Shanahan wrote:
Which appears to show that the extended partition information is missing.
I just tried on my system, and I get this results: |Command (m for help): p | |Disk /dev/hda: 60.0 GB, 60022480896 bytes |255 heads, 63 sectors/track, 7297 cylinders |Units = cylinders of 16065 * 512 = 8225280 bytes | | Device Boot Start End Blocks Id System |/dev/hda1 * 1 2805 22531131 c Win95 FAT32 (LBA) |/dev/hda2 2806 6630 30724312+ f Win95 Ext'd (LBA) |/dev/hda3 6631 6892 2104515 83 Linux |/dev/hda4 6893 7297 3253162+ 83 Linux |/dev/hda5 2806 5100 18434556 b Win95 FAT32 |/dev/hda6 5101 5360 2088418+ 6 FAT16 |/dev/hda7 5361 5620 2088418+ 6 FAT16 |/dev/hda8 5621 5880 2088418+ 6 FAT16 |/dev/hda9 5881 6011 1052226 82 Linux swap |/dev/hda10 6012 6241 1847443+ 83 Linux |/dev/hda11 6242 6630 3124611 fd Linux raid autodetect |Expert command (m for help): p | |Disk /dev/hda: 255 heads, 63 sectors, 7297 cylinders | |Nr AF Hd Sec Cyl Hd Sec Cyl Start Size ID | 1 80 1 1 0 254 63 756 63 45062262 0c | 2 00 0 63 1023 254 63 485 45062325 61448625 0f | 3 00 254 63 1023 254 63 1023106510950 4209030 83 | 4 00 254 63 1023 254 63 1023110719980 6506325 83 | 5 00 254 63 1023 254 63 1023 63 36869112 0b | 6 00 254 63 1023 254 63 1023 63 4176837 06 | 7 00 254 63 1023 254 63 1023 63 4176837 06 | 8 00 254 63 1023 254 63 1023 63 4176837 06 | 9 00 254 63 1023 254 63 1023 63 2104452 82 |10 00 254 63 1023 254 63 1023 63 3694887 83 |11 00 254 63 1023 254 63 1023 63 6249222 fd (note that from #5 onwards, they are identical) |Expert command (m for help): e | |Disk /dev/hda: 255 heads, 63 sectors, 7297 cylinders | |Nr AF Hd Sec Cyl Hd Sec Cyl Start Size ID | 2 00 0 63 1023 254 63 485 45062325 61448625 0f | 5 00 0 63 1023 254 63 1023 36869175 4176900 05 | 6 00 0 63 1023 254 63 1023 41046075 4176900 05 | 7 00 0 63 1023 254 63 1023 45222975 4176900 05 | 8 00 0 63 1023 254 63 1023 49399875 2104515 05 | 9 00 0 63 1023 254 63 1023 51504390 3694950 05 |10 00 0 63 1023 254 63 1023 55199340 6249285 05 |11 00 0 0 0 0 0 0 0 0 00 Some partitions are missing from that list of extended partitions, and #11 shows as empty. Results for my other disk are very similar. I use suse 8.2. I don't know how those "extended" commands work, so I think it is more a weirdness of fdisk thtan any thing "real". From man fdisk: |BUGS | There are several *fdisk programs around. Each has its | problems and strengths. Try them in the order cfdisk, | fdisk, sfdisk. (Indeed, cfdisk is a beautiful program | that has strict requirements on the partition tables it | accepts, and produces high quality partition tables. Use | it if you can. fdisk is a buggy program that does fuzzy | things - usually it happens to produce reasonable results. | Its single advantage is that it has some support for BSD | disk labels and other non-DOS partition tables. Avoid it | if you can. sfdisk is for hackers only - the user inter | face is terrible, but it is more correct than fdisk and | more powerful than both fdisk and cfdisk. Moreover, it | can be used noninteractively.) So... I go to cfdisk, dump the table to file, and I get: | ---Starting--- ----Ending---- Start Number of | # Flags Head Sect Cyl ID Head Sect Cyl Sector Sectors |-- ----- ---- ---- ---- ---- ---- ---- ---- -------- --------- | 1 0x80 1 1 0 0x0C 254 63 1023 63 45062262 | 2 0x00 254 63 1023 0x0F 254 63 1023 45062325 61448625 | 3 0x00 254 63 1023 0x83 254 63 1023 106510950 4209030 | 4 0x00 254 63 1023 0x83 254 63 1023 110719980 6506325 | 5 0x00 254 63 1023 0x0B 254 63 1023 63 36869112 | 6 0x00 254 63 1023 0x06 254 63 1023 63 4176837 | 7 0x00 254 63 1023 0x06 254 63 1023 63 4176837 | 8 0x00 254 63 1023 0x06 254 63 1023 63 4176837 | 9 0x00 254 63 1023 0x82 254 63 1023 63 2104452 |10 0x00 254 63 1023 0x83 254 63 1023 63 3694887 |11 0x00 254 63 1023 0xFD 254 63 1023 63 6249222 Huh? 1023 only? My disk is bigger. So I dump by sectors: | First Last | # Type Sector Sector Offset Length Filesystem Type (ID) Flags |-- ------- -------- --------- ---- --------- ---------------------- --------- | 1 Primary 0 45062324 63 45062325 Win95 FAT32 (LBA) (0C) Boot (80) | 2 Primary 45062325 106510949 0 61448625 Win95 Ext'd (LBA) (0F) None (00) | 5 Logical 45062325 81931499 63 36869175 Win95 FAT32 (0B) None (00) | 6 Logical 81931500 86108399 63 4176900 FAT16 (06) None (00) | 7 Logical 86108400 90285299 63 4176900 FAT16 (06) None (00) | 8 Logical 90285300 94462199 63 4176900 FAT16 (06) None (00) | 9 Logical 94462200 96566714 63 2104515 Linux swap (82) None (00) |10 Logical 96566715 100261664 63 3694950 Linux (83) None (00) |11 Logical 100261665 106510949 63 6249285 Linux raid autode (FD)None (00) | 3 Primary 106510950 110719979 0 4209030 Linux (83) None (00) | 4 Primary 110719980 117226304 0 6506325 Linux (83) None (00) See? They are different - Mmmm, interesting... this is the order in which I created my partitions, I think. Try " sfdisk -l -V /dev/hda", it checks for errors, and the man page above says sfdisk is the most correct of all programs. In fact, it says mine has errors - but as it works, I aint touching it :-)
Google and suse db provide no (recent) answers, but many other similar experiences. Is there and answer or have I lost the partition? Would the "Expert command" to "f fix partition order" have any helpful result?
Order? No, that's another thing.
Is there hope?
Always :-) -- Cheers, Carlos Robinson
* Carlos E. R.
Try " sfdisk -l -V /dev/hda", it checks for errors, and the man page above says sfdisk is the most correct of all programs. In fact, it says mine has errors - but as it works, I aint touching it :-)
This acknowledges that there is a _dirty_ file-system, and advises e2fsck which bombs with the same short/zero length block excuse.
Is there hope?
Always :-)
I'm not seeing it. I did not make the extended partition cover the entire disk when I first set it up, so I created another partition (rest of the drive) with parted and am copying the dirty file-system to the new partition. I don't know it this will accomplish anything but, it might leave me with a recoverable area if I finish trashing the problem area. Parted has a rescue command: rescue START END rescue a lost partition near START and END which accomplish exactly nothing. Thanks for the ideas, -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org
The Friday 2004-02-06 at 23:53 -0500, Patrick Shanahan wrote:
* Carlos E. R.
[02-06-04 22:29]: Try " sfdisk -l -V /dev/hda", it checks for errors, and the man page above says sfdisk is the most correct of all programs. In fact, it says mine has errors - but as it works, I aint touching it :-)
This acknowledges that there is a _dirty_ file-system, and advises e2fsck which bombs with the same short/zero length block excuse.
:-( In my case, the errors may come because some of my partitions were created with partition magic, and there are different opinions between both groups about partition definitions.
Is there hope?
Always :-)
I'm not seeing it.
I know... but there is still life.
I did not make the extended partition cover the entire disk when I first set it up, so I created another partition (rest of the drive) with parted and am copying the dirty file-system to the new partition. I don't know it this will accomplish anything but, it might leave me with a recoverable area if I finish trashing the problem area.
Parted has a rescue command: rescue START END rescue a lost partition near START and END which accomplish exactly nothing.
The easiest way would be to recreate the partition, and copy files back from the backup - providing you can at least mount it ro. It might be a size mismatch, and confusing error messages, triggered perhaps when you reached the end of the partition on a write operation. -- Cheers, Carlos Robinson
* Carlos E. R.
The Friday 2004-02-06 at 23:53 -0500, Patrick Shanahan wrote:
The easiest way would be to recreate the partition, and copy files back from the backup - providing you can at least mount it ro.
parted after 14 hours has accomplished nothing apparent, ie: backup attempt failed. There is no mounting of the defective partition or the attempted copy and fsck.ext3 or e2fsck bomb as before.
It might be a size mismatch, and confusing error messages, triggered perhaps when you reached the end of the partition on a write operation.
since the copy failed, I do not see this. thanks again, -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org
The Saturday 2004-02-07 at 14:24 -0500, Patrick Shanahan wrote:
The easiest way would be to recreate the partition, and copy files back from the backup - providing you can at least mount it ro.
parted after 14 hours has accomplished nothing apparent, ie: backup attempt failed. There is no mounting of the defective partition or the attempted copy and fsck.ext3 or e2fsck bomb as before.
:-( I'm sorry, I'm out of ideas. Perhaps if I had it, but maybe not even then. Maybe... I don't know, try mounting with the rescue CD of another version of SuSE. -- Cheers, Carlos Robinson
* Carlos E. R.
:-(
I'm sorry, I'm out of ideas. Perhaps if I had it, but maybe not even then. Maybe... I don't know, try mounting with the rescue CD of another version of SuSE.
I did and this morning I mkfs.ext3. There is nothing left to do. Thanks, -- Patrick Shanahan Registered Linux User #207535 http://wahoo.no-ip.org @ http://counter.li.org
participants (2)
-
Carlos E. R.
-
Patrick Shanahan