Please excuse my ultra newbie questions. I have a problem with two RAID 0 hard drives from a WD My Book which has failed. The drives have the ext3 file system and are Linux raid autodetect type. Here are the questions: Does the mkdir /media/testing
approach create the directories on the boot drive? Is that approach going to work for me since I have two physical drives?
3 Answers
The MyBook is just a Linux appliance. In the MyBook NAS with a single disk it's simply a broken RAID mirror, as it is missing the second disk but is already configured.
You can run mount
to see the currently mounted filesystems. cat /proc/mdstat
will show the status of the RAID.
What are you attempting to do? Further information is necessary to provide you a more concise answer.

- 23,756
- 2
- 59
- 69
-
First, thank you very much for helping. The (broken) MyBook had two 1 TB drives in a raid 0 configuration. It was used as NAS and had a whole bunch of really important files on it. I'm trying to attach the drives to a UBUNTU 10.10 box, recreate the RAID, mount the RAID, and get the files back. Two things, if there is a better LINUX dist, I can change the hard drives and install a new OS. Second, if there is a way to capture screen shots, I'll post them. – user68697 Jan 29 '11 at 23:03
-
2If you had a disk in a RAID0 fail with no backups, your data is probably toast without paying a data recovery service. I lost my notes for re-assembly so I will have to re-write from scratch. I'll give more details later but `mdadm --assemble --scan` should work in many cases. – Warner Jan 31 '11 at 22:12
-
Warner, thanks a bunch for the help. When I run the UBUNTU 10.10 Disk Utility I get a green dot and a "disk is healthy" message on the S.M.A.R.T. satus line for each drive. I dropped to a Terminal window and did the mdadm command you gave me and I got no response (and no error message) which I assume is good. Not knowing what else to do, I returned to the UBUNTU Disk Utility and tried to mount the volume. I got an error message of "error mounting: mount: wrong fs type, bad option, bad superblock on /dev/md4, missing codepage or helper program, none of which sounds good. – user68697 Feb 01 '11 at 03:35
-
What does `cat /proc/mdstat` show after you assemble? – Warner Feb 01 '11 at 15:12
-
personalities : [raid1] [raid0] [linear] [multipath] [raid6] [raid5] [raid4] [raid10] md4 : active raid0 sdc4[0] sdb4[1] 1945407104 blocks 64k chunks md2 : active raid1 sdb2[1] sdc2[0] 104320 blocks [2/2] [uu] md1 : active raid1 sdc1[0] sdb1[1] 2939776 blocks [2/2] [uu] md3 : active raid1 sdc3[0] sdb3[1] 987904 blocks [2/2] [uu] unused devices:
looks like the data is on md4?? thanks again – user68697 Feb 01 '11 at 23:35 -
Looks like it assembled without issue.. but that's difficult to read. You could edit your original post to retain formatting. – Warner Feb 03 '11 at 14:56
Is that approach going to work for me since I have two physical drives?
No, because data splits across two disks.
I test on MyBook Studio Edition 1TB, but RAID not failed:
# mknod /dev/md4 b 9 4
# mdadm -Cv /dev/md4 -l0 -n2 -c64 /dev/sdb4 /dev/sdc4
# mkdir /media/mybook
# mount /dev/md4 /media/mybook

- 8,437
- 25
- 22
-
First, thank you for the help. When I went to the WIKI link, the first command was mdprobe md. That came back with a Fatal: Module md not found error. I'm currently using UBUNTU 10.10 but can install something else if need be. Thanks. – user68697 Jan 29 '11 at 23:19
I would initially try mdadm --build ...
to test-build the array with params being tested, then mdadm -o
to make it read-only, then checking mounting read-only and possibly fsck. Only since the 2 last steps (as well as own files checking) would complete relatively successfully, I would re-create the array with the same params as used when '--build'ing. Otherwise data loss chances would be too high.

- 9,448
- 2
- 25
- 52
-
I would be glad to do those steps but I'm too much of a Linux newbie to know how. Can you give me simpler language? I got the mdadm --build part but I don't know what params to use. Thanks.. – user68697 Feb 01 '11 at 23:44