[Please CC me since I'm not subscribed to this list.] I've asked this question on stackexchange
http://superuser.com/questions/482447/quickly-skipping-sata-media-errors but I got no reply, so I'm trying here :-) I had a head crash on a SATA HDD. Zillions of blocks are damaged, and I try to save the contents to another HDD using ddrescue on a GNU/Linux rescue system booted from a DVD. The very problem is that accessing a single bad block takes up to 40 seconds. Obviously, the hard disk tries to re-read for some time until it gives up. Looking this up in the internet it seems that the HDD itself takes such a long time (it's a Toshiba HDD). Now my question: is there any method to make this faster? For example, telling the SATA controller to try reading only once and not, say 10 times until an I/O error gets returned? I know how to play with ddrescue options, BTW, so I can do a kind of binary search to find the start and the end of the large bad block areas, but having a full scan which doesn't take weeks would be nice... Thanks in advance for any pointers! Werner _______________________________________________ Bug-ddrescue mailing list [email protected] https://lists.gnu.org/mailman/listinfo/bug-ddrescue
