Paul L Daniels wrote:
I switched over to an eSATA (powered) connection and I was
able to get 120K/sec until the drive hits the 4th head on the
platters which is dead. Each time this happens the old
'quirk' that I've observed where the data rate doesn't recover
until you ctrl-C and restart is back.
I thought there were slow reads only, no errors. If there are errors
then ddrescue begins to skip, and the read rate drops, but only until
good data is found again.
Can you (Antonio) tell me where the data-rate/read-size
adjustment occurs [ when errors are returned ] in your C++ code,
Sure. Change the line 139 in rescuebook.cc from
const int alignment = ( skip_size ? hardbs() : softbs() );
to
const int alignment = softbs();
(Do the same to line 210 if you use reverse mode).
But (if it works for you) it is easier to make hardbs() == softbs() by
using --cluster-size=1 and adjusting the sector size.
I just want to first try disabling the rate-adjustment to see if
I can start to isolate where this quirk is coming from.
Please, try it. I am also interested in making ddrescue work around
drive problems like this.
Best regards,
Antonio.
_______________________________________________
Bug-ddrescue mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/bug-ddrescue
