[Bug-ddrescue] Strange behaviour at the endof the rescue

[David Morrison]

Hello

I have been using ddrescue 1.6 very extensively over the last month 
or so to try to recover files from a failed disk. Apart from other 
problems, the disk seems to have a temperature sensitivity. It will 
only work for a few minutes until it heats up, then it will go 
offline. I have to stop ddrescue, turn off the disk, wait for a while 
to let it cool down, turn it on again, then start ddrescue again. 
It's a slow process, but it has allowed me to got a large number of 
my files back. So thank you very much for writing ddrescue.

I am writing now because of what seems like incorrect behaviour right 
at the end of the scan.

The ddrescue command was:

/usr/local/bin/ddrescue -r 2 -v -i137563369472 -s112495972352 -o0 
/dev/disk3 /Volumes/127GB/ddr-104GB.dmg /ddr-disk2-104GBsecond.log

This was grabbing one partition from the bad disk and creating a disk 
image file on Mac OS X.

The logfile at the time looked like this:

# Rescue Logfile. Created by GNU ddrescue version 1.6
# current_pos  current_status
0x3345FC7A00     /
#      pos        size  status
0x00000000  0x20076A7000  ?
0x20076A7000  0x6244AE000  +
0x262BB55000  0x00001000  -
0x262BB56000  0x6121C000  +
0x268CD72000  0x00001000  -
0x268CD73000  0x6BCB74200  +
0x2D498E7200  0x00007200  -
0x2D498EE400  0x00006C00  +
0x2D498F5000  0x00000800  -
0x2D498F5800  0x00000800  +
0x2D498F6000  0x00001200  -
0x2D498F7200  0x00000200  +
0x2D498F7400  0x0000FE00  -
0x2D49907200  0x00000200  +
0x2D49907400  0x0000FE00  -
0x2D49917200  0x5FC6B0000  +
0x3345FC7200  0x00000800  -
0x3345FC7A00  0x0000C600  /
0x3345FD4000  0x34EAB3200  +
0x3694A87200  0x00001E00  /
0x3694A89000  0xA350E200  +
0x3737F97200  0x00009E00  /
0x3737FA1000  0x11F0D6000  +
0x3857077000  0x00001000  /
0x3857078000  0x1738F200  +
0x386E407200  0x0000DE00  /
0x386E415000  0x1CA717000  +
0x3A38B2C000          -1  ?

I ran the ddrescue command again. It reported having finished, and 
the logfile was this:

# Rescue Logfile. Created by GNU ddrescue version 1.6
# current_pos  current_status
0x3345FD2600     +
#      pos        size  status
0x00000000  0x20076A7000  ?
0x20076A7000  0x1A31485000  +
0x3A38B2C000          -1  ?

This would seem to indicate that the whole partition has been read 
correctly. However, the previous logfile indicates a lot of errors, 
some of which had been bad enough to cause the disk to go offline 
while trying to read the blocks in earlier runs.

This has actually happened to me twice. The first time I had not been 
tracking what was happening so well, so I assumed that somehow the 
disk had come good to allow all the blocks to be recovered. But when 
it happened a second time, I thought it was too strange. Apart from 
anything else, it would have been useful to know which blocks on the 
disk were truly bad, so as to be able to identify whether specific 
files had been damaged.

Thanks

David
--
David Morrison
PO Box 105, Cardiff, NSW 2285, Australia
David_Morrison at internode.on.net
Ph +61 2 49546164
IT Consulting and Support