Home server running Ubuntu Server edition went down last weekend. Power supply failure.
After installing a new power supply and reboot, I noticed I lost my RAID5 array (4x 1TB drives). Upon closer inspection, I noticed that two of the drives showed faulty. Checking the drives individually the drives did not appear to be faulty.
Admittedly, I am rusty working with RAID arrays. I don't use linux much any more at work and this is my home setup. I'm perfectly good at searching all over Google-dom, as I have, but it appears I may have made some mistakes here so this is my last ditch effort to find a solution for data recovery. Thankfully the *most* important data that was on the array was backed up, but half of it (non-critical but would like to have back) was not.
So when I went to re-mount the RAID array, it no longer existed.
I used mdadm to examine the drives, here is what I found:
As I checked two of the other drives, they claimed that two of the disks in the middle (/dev/sdc1 and /dev/sda1) were faulty.
I searched and searched. Ultimately I went with trying to re-assemble the array with the following command:
It then rebuilt the array. I should mention that while I re-installed the power supply, I had to unplug some of the SATA cords and reinstall them. They may have not been re-connected in the same original order. However, I assumed that mdadm was smart enough to recognize which drives went in which order. Maybe I was wrong? Anyhow, after I ran my command, I noticed that mdadm began assembling the array in a different order than I prescribed in the command.
I got this:
Since it took a while to rebuild, I went to bed. Then, the next morning I woke up and tried to see if I could remount the drive. No luck. No filesystem found (should have been ext4). Tried running fsck but nothing was found. Also tried e2fsck, etc.
Then I remembered something. I vaguely remembered when building the array several years ago (yes this array has been running for several years with no failures) that I think I might have used 128k chunk sizes. I noticed the newly re-assembled array was 64k.
Now I have the sinking feeling that I may have hosed the array. At any rate, I'm lost and don't know what else can be done, if anything.
Thanks to anyone who might be able to help!
After installing a new power supply and reboot, I noticed I lost my RAID5 array (4x 1TB drives). Upon closer inspection, I noticed that two of the drives showed faulty. Checking the drives individually the drives did not appear to be faulty.
Admittedly, I am rusty working with RAID arrays. I don't use linux much any more at work and this is my home setup. I'm perfectly good at searching all over Google-dom, as I have, but it appears I may have made some mistakes here so this is my last ditch effort to find a solution for data recovery. Thankfully the *most* important data that was on the array was backed up, but half of it (non-critical but would like to have back) was not.
So when I went to re-mount the RAID array, it no longer existed.
I used mdadm to examine the drives, here is what I found:
As I checked two of the other drives, they claimed that two of the disks in the middle (/dev/sdc1 and /dev/sda1) were faulty.
I searched and searched. Ultimately I went with trying to re-assemble the array with the following command:
Code:
sudo mdadm --create /dev/md0 -v -l 5 -n 4 /dev/sd[bcae]1I got this:
Since it took a while to rebuild, I went to bed. Then, the next morning I woke up and tried to see if I could remount the drive. No luck. No filesystem found (should have been ext4). Tried running fsck but nothing was found. Also tried e2fsck, etc.
Then I remembered something. I vaguely remembered when building the array several years ago (yes this array has been running for several years with no failures) that I think I might have used 128k chunk sizes. I noticed the newly re-assembled array was 64k.
Now I have the sinking feeling that I may have hosed the array. At any rate, I'm lost and don't know what else can be done, if anything.
Thanks to anyone who might be able to help!
