Mount Removed Raid Disk Back On Same Machine As? Original Raid

Home » CentOS » Mount Removed Raid Disk Back On Same Machine As? Original Raid
CentOS No Comments

Unfortunately, no.  The system would not mount the drive without some other changes.  Listing the process here for anyone else who comes across this thread.

Trying to start the array (with or without –force), fails with the message:

    mdadm: Found some drive for an array that is already active: /dev/md127
    mdadm: giving up

I found with some digging that I needed to change the UUID of the drive to be able to mount it separately from the existing array.  I used
“sgdisk -G /dev/sdg1” to do this.  It worked, but gave quite a few scary warning messages in the process.  A better idea would have been to use uuidgen to generate a random uuid and then start the array like this:

    mdadm –assemble /dev/md99 –update=uuid –uuid= /dev/sdg1
(might require –force for a broken array, I’m not sure since I didn’t actually do it this way)

Once the array is running, there is another problem.  Attempting to mount the array gives another error:

    mount: wrong fs type, bad option, bad superblock on /dev/md99,
        missing codepage or helper program, or other error

Useless error message.  You have to look in dmesg to see the actual problem:

    XFS (md99): Filesystem has duplicate UUID
7e237dbd-6c24-4781-98d1-a1ae80a3ed13 – can’t mount

I would assume you would have a similar issue with any other filesystem.  In my case, since it is XFS, I used uuidgen to generate another random uuid and then updated it like this:

    xfs_admin -U /dev/md99

After that, the filesystem mounted normally.

Hopefully that is helpful for anyone else who finds themselves in this situation.

—
Bowie

Mount Removed Raid Disk Back On Same Machine As Original Raid

Home » CentOS » Mount Removed Raid Disk Back On Same Machine As Original Raid
CentOS 2 Comments

I have a CentOS 7 system with an mdraid array (raid 1).  I removed a drive from it a couple of months ago and replaced it with a new drive. 
Now I want to recover some information from that old drive.

I know how to mount the drive, and have done so on another system to confirm that the information I want is there.

My question is this:

What is going to happen when I try to mount a drive that the system thinks is part of an existing array?

To put it another way:  I had two drives in md127.  I removed one (call it drive1), and replaced it with a new drive.  Some files were accidentally deleted from md127, so now I want to connect drive1 back to the same machine and mount it as a separate array from md127 so I can copy the files from drive1 back to md127.  What do I need to do to make that happen?

Thanks,

Bowie

2 thoughts on - Mount Removed Raid Disk Back On Same Machine As Original Raid

  • Once upon a time, Bowie Bailey said:

    I don’t _think_ anything special will happen – md RAID doesn’t go actively looking for drives like that AFAIK. And RAID 1 means you should be able to ignore RAID and just access the contents directly.

    However, the contents could still be a problem. If LVM was in use on it, that will be a problem, because LVM does auto-probe and will react when it sees the same UUID (IIRC LVM will only block access to the newly seen drive). I don’t think any filesystems care (I know I’ve mounted snapshots of ext4 and IIRC xfs on the same system, haven’t touched btrfs).

  • I’m not using LVM on this drive, so that won’t be an issue.

    My concern is that since the raid info on the drive will identify itself as part of the active raid, the system will try to add it to the raid
    (probably as a spare) when it comes online.  I don’t think that would be destructive, but I would have to figure out how to separate it out if that happens.  I’m hoping that it won’t be an issue since there are no missing drives in the existing raid.

    I know I will have to bring the drive online as a broken array, but I’ve done that from other systems.  The only question there is can I simply rebuild it with a different name.  I assume I can just do “mdadm -A
    –run /dev/md0 /dev/sdc1” (possibly with “–force” due to the broken array) even if sdc1 was originally part of the existing md127 array?

    The system in question is in a data center, so I’m trying to get ahead of any possible problems to avoid having to deal with unexpected issues while I’m there.

    —
    Bowie