These things come in 3's well 4's I just got home after replacing 3 failing drives at a clients place, 2x in a raid array and one in a desktop. You don't want to use the standard DD as it barfs when you have a disk error. You can manually futz around with it to try and recover it or... Use the thing what saved my bacon this day AKA gnu dd recovery. In the following sda and sdb are the failed disks from the raid array, sdc is the good disk that you want to stick your stuff on. Easiest way to do it is boot off an ubuntu live CD. then open a terminal *sudo su (everything your doing is adminy anyway so no point typing sudo 400 times) apt-get install gddrescue fdisk -l * This will list the disks/partitions available in your system and their sizes. once you have determined from there what your source and destination drives are *ddrescue -v /dev/sda /dev/sdc logfile.log* syntax for that is ddrescue [options] infile outfile [logfile] I'd recommend running the rescue outside of the raid controller. IE run it on another computer or plug it into the mbo rather than the raid controller, otherwise your probably going to "unraid" the disk by erasing all the magic mojo the controller puts on there to keep track of it. 40GB drives copied in about 20-30 minutes or so. Assuming that the bad blocks haven't changed you can stick the other drive in and merge the two drives together by running it again *ddrescue -v /dev/sdb /dev/sdc logfile.log *Note that to do this you either need to save logfile.log on a usb stick (or whatever) or have all the drives plugged into the system at the same time as you loose the file system when you reboot. If you then reboot and plug your disk into the raid controller you can use gparted to resize the ntfs partition to the full size of your new disks. Now to login to my clients server and start the rebuild while nobody's using it ;-> Any problems let me know and we can arrange a phone call or something, I'm at GMT +10 and available from 9:00AM till about midnight our time. additional resources http://www.gnu.org/software/ddrescue/ddrescue.html https://help.ubuntu.com/community/DataRecovery#head-2720985ba3c025c9e3f9e487f7cfa7daf03dab24 James Newton wrote: > A few of you may have noticed that piclist.com was down for a few hours this > am. > > Short version: Does anyone have an old copy of a drive-imaging program that > will work on an NT 4.0 server? It would speed up the rebuild. The site may > be down for a while, but it will continue, one way or the other. This IN NO > WAY affects the list itself as it is hosted by MIT (thank you Jory and our > anonymous sponsor at MIT). > > At about 8:30PST my web server rebooted for no apparent reason, then came up > to a RAID array controller error screen saying that the RAID 1 set was off > line. A heart attack later, it was found that one of the drives had failed, > but the other was still reported as ok. After a comedy of errors which > included: Pulling the wrong drive, forgetting that I had a spare drive and > trying to purchase one locally (every computer store in a 5 mile radius was > CLOSED on Monday at 9:30 am!), remembering the spare drive, wondering why it > still wouldn't rebuild, finding the documentation for the controller that > says it requires two IDENTICAL drives (it doesn't), tracking down an > identical MAX6L040J2 on the net, ordering it and paying for overnight AM > delivery before figuring out that I had pulled the good drive from the > original array and that the spare I had would work just fine (the AM > delivery is costing me more than the drive), starting the rebuild only to > have it error out due to some error on the GOOD drive (!), eventually I gave > up and just booted it on the one drive so I could update my backup from > Friday. > > As of now, it's just up and running on the one drive. I'm not going to > chkdsk it until I can image the remaining good (?) drive. I have a pretty > good backup, including ALL the data, but not a drive image of the system. So > if it fails, we will be down for a few days while I try to rebuild it. Given > that the system is running Windows NT 4.0 on a PIII 800 and the drives, > power supply, and processor are more than 6 years old, it's well past time > to upgrade or move to a hosted service. > > I think the problem may be the power supply as it wouldn't boot on the good > drive when the spare was still in the box even without the data cable > attached. Mother board monitor shows the voltages as rock steady, but that > wouldn't catch spikes or other noise. > > I've opened a minimal account at godaddy just to see what can be done with a > shared host: I'll upgrade the account as I run into roadblocks. Piclist.org > will be pointed there sometime soon. Does anyone else host at godaddy? Do > they host DNS or do I have to keep my register.com account? > > It would be really nice to be able to image the drive before I try to chkdsk > it and rebuild the array again tomorrow or Thursday, but I'll be darned if I > can find a program that will run on NT 4.0. All the ones I can find are XP > and higher. If anyone has an old copy of Ghost pre 2003 they don't mind > donating, I would really appreciate it. > > One way or the other, the site will continue, but there may be some down > time while I build a new box or while I figure out how to migrate things to > a host. > > James Newton: PICList webmaster/Admin > mailto:jamesnewton@piclist.com 1-619-652-0593 phone > http://www.piclist.com/member/JMN-EFP-786 > PIC/PICList FAQ: http://www.piclist.com > > > -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist