Hi, thanks again :) The server is shut down, will copy the two broken disks. I think (and hope) there has been a small amount of writing since the problems occurred. Unexpected problem with the names of ddrescue, but the apt-get package in Debian called gddrescue, the program is called ddrescue. I became uncertain because you mention that it works like dd, but it doesn't use if=foo of=bar like regular dd? Anyway the program ddrescue --help refers to the homepage http://www.gnu.org/software/ddrescue/ddrescue.html which I assume is right one..? And there are a lot of options, any tips on some special ones I should use? I also wonder if it is right to let mdadm try to recover from all four disks, the first one to stop working where three days before I discovered it. Isn't it better to just use three disks, the two disks that are ok, and the last disk that got too many write errors the night before I discovered everything? Otherwise, you have confirmed/clarified that everything seems to work the way I hoped. And I will read up on all the news in mdadm, and dm-integrity sounds interesting. Thanks! Cheers Rickard
