-
Ops, forget about block size suggestion, I had not readed carefully that topic: http://sourceforge.net/projects/myrescue/forums/forum/235860/topic/3284820 where is explained about the block size.
And another experience:
Now I'm experimenting with myrescue as a hard disk diagnostic and repairing tool (reading and writing to the same device) and seems it works well, but I wonder if it...
2009-11-17 16:13:32 UTC by tonikun
-
Hello, I want to be grateful to the developers for this great program and I will talk about my experiences with the program as they suggest in the man page.
1) The great things that make a great program:
Reverse direction, Skip mode and Jump mode. With a mix of this modes I have elaborated a general recommended procedure that I'm applying successfully (see below).
2) Some improvements:
2009-11-17 00:31:27 UTC by tonikun
-
Hi Kristof, many thanks for your explanation.
* What do you think about the O_DIRECT flag of open()?
I'm trying to play with that, however it's far from obvious... I always get an "Invalid parameters" error...
* What do you think about the "hdparm --read-sector" feature?
It could be useful for a recovery tool, isn't it?
(It requires kernel module ide-core to be...
2009-05-29 12:04:04 UTC by clohr
-
Kristof,
Thanks for the big explanation. I have a lot of things more clear now.... including that we still do not know how to get the hardware sector size :D
I have gone to see the specs for my disk:
Drive Specifications
WD model number WD5000AAKS
SATA Hard Drives 500 GB
Formatted capacity1 500,107 MB = 500107000000
User sectors per drive 976,773,168
Doing...
2009-05-29 00:37:15 UTC by sjmuniz
-
Funny, this is pretty similar to two tools I commited in the CVS but that have not yet made it into a release. In the CVS you find a primitive stat programm that basically just counts how often each value appears. The printout is not as elaborate as yours, I just printed value and count, not the interpretation.
And there is a visual bitmap generator, that has even been somewhat helpful. It...
2009-05-28 22:02:54 UTC by kristofk
-
I realize I should emphasize this: You should never, Never, NEVER EVER CHANGE THE BLOCK SIZE.
Here's why: In the block bitmap, the state of one block is stored in one byte. So when you started with a size of 4096 bytes and then switched to 512 bytes, the area on the disc from byte 0 to 4095 that before was block 0 now makes up blocks 0 to 7, what was once block 1 (4096 to 8191) is now block 8...
2009-05-28 22:00:36 UTC by kristofk
-
Oh yes, the block size. That has always bothered me.
The problem is, there are several figures around: The sizes of filesystem allocation units, the units used in the partition table, the block size the kernel uses when accessing the disc and others.
Which one you use is - in normal conditions - pretty much irrelevant. myrescue actually instructs the kernel to request specific byte ranges...
2009-05-28 21:58:59 UTC by kristofk
-
Hello,
I had the same problem recently.
I have searched for a while, and my conclusion is that new disks use a hardware sector size of 512 bytes.
The two places were I found that are:
1)
Mar 18 16:14:15 orion kernel: [224589.579883] sd 0:0:0:0: [sda] 976773168 512-byte hardware sectors (500108 MB)
Mar 18 16:14:15 orion kernel: [224589.579894] sd 0:0:0:0: [sda] Write Protect is off
Mar...
2009-05-28 21:50:12 UTC by sjmuniz
-
Well... Now I'm trying a run with -b 512
Messages are consistents between myrescue and kernels errors:
block 000026325 (000000000-059279787) ok 000000000 bad 000001263
May 28 14:13:57 pooki kernel: [69711.107938] hdg: dma_intr: status=0x51 {DriveReady SeekComplete Error}
May 28 14:13:57 pooki kernel: [69711.107938] hdg: dma_intr: error=0x40 {UncorrectableError}, LBAsect=26388...
2009-05-28 12:19:35 UTC by clohr
-
I am disappointed by the differences between sfdisk and fdisk.
#Â sfdisk -l /dev/hdg
Disk /dev/hdg: 59560 cylinders, 16 heads, 63 sectors/track
Warning: extended partition does not start at a cylinder boundary.
DOS and Linux will interpret the contents differently.
Warning: The partition table looks like it was made
for C/H/S=*/255/63 (instead of 59560/16/63).
For this listing I'll...
2009-05-28 12:04:45 UTC by clohr
