So then you get impossibly fast write times, and boot disks that dont work,
because they arent actually complete, if the user removes the disk when
ddrescue reports that the operation is complete.
$ sudo ddrescue --force bodhi-6.0.0-64.iso /dev/sdb
GNU ddrescue 1.23
Press Ctrl-C to interrupt
ipos: 872349 kB, non-trimmed: 0 B, current rate: 727 MB/s
opos: 872349 kB, non-scraped: 0 B, average rate: 436 MB/s
non-tried: 0 B, bad-sector: 0 B, error rate: 0 B/s
rescued: 872415 kB, bad areas: 0, run time: 1s
pct rescued: 100.00%, read errors: 0, remaining time: n/a
time since last successful read: n/a
Finished
Shown here, a 1 second run time, for a 700MB file to USB. But I already
know this is impossible on my hardware. I'm just a dumb construction
worker, but if i had to guess, ddrescue is listening to userspace
libraries, rather than kernel signal processes.
Here iostat reports:
iostat -d
Linux 5.15.0-56-generic (i) 01/10/2023 _x86_64_ (12 CPU)
Device tps kB_read/s kB_wrtn/s kB_dscd/s kB_read
kB_wrtn kB_dscd
loop0 0.01 0.09 0.00 0.00 1133
0 0
loop1 0.01 0.09 0.00 0.00 1133
0 0
loop2 0.00 0.03 0.00 0.00 363
0 0
loop3 0.01 0.09 0.00 0.00 1090
0 0
loop4 0.01 0.04 0.00 0.00 439
0 0
loop5 0.01 0.04 0.00 0.00 438
0 0
sda 47.56 465.18 313.64 0.00 5574816
3758761 0
sdb 0.47 0.16 36.52 0.00 1932
437704 0
It appears from iostat, that the disk sdb is still being written, despite
ddrescue reporting completion.
