> Software > Storage for Backups > HyperScale X > HyperScale 1.5 > Reference Architecture > Operations

Replacing a Metadata Drive

When an NVME drive in a HyperScale 1.5 storage pool fails due to disk issues, you can replace it with a new empty disk. The NVME drive contains the metadata that is associated with the index cache and the deduplication database (DDB).

Before You Begin

  • Verify and ensure that the capacity of the new disk is the same as the disk you want to replace.

  • Enable root access, if root access is disabled in the storage pool. For more information, see Enabling or Disabling Root Access.

Procedure

  1. Shut down the node and remove the defective disk containing the metadata.

  2. Attach the new replacement disk to the MediaAgent that has the failed disk.

  3. Reboot the node.

    Note

    Do not mount the disk on the MediaAgent after the reboot.

  4. Navigate to the following folder:

    /opt/commvault/Base/

  5. Access the node using the iRMC management console in your environment and navigate to the following folder:

    /opt/commvault/Base/

  6. Run the following command to identify the unused drives that were replaced:

    ./CVSDS -S

    The sample output list the unused drives.

    Unused Disks: [metadata]
============================
   Disk Name | Disk Size(GB)
============================
/dev/nvme0n1 |   5961.631348
============================

  7. To configure the new metadata drive, run the following command:

    ./CVSDS -M -I /dev/new disk -V CV

    For example:

    ./CVSDS -M -I /dev/nvme0n1 -V CV

    To replace multiple disks, include all the names in the <new disk> list, using a comma (,) as the delimiter.

    For example:

    ./CVSDS -M -I /dev/nvme0n1,/dev/nvme0n2 -V CV

    The following prompt will be displayed:

    Please confirm that the node has been rebooted after replacing the damaged metadata drive. Press 'y' to confirm and continue with the process [y/n]: y
Unused Disks: [metadata]
============================
   Disk Name | Disk Size(GB)
============================
/dev/nvme0n1 |   5961.631348
============================
Following disks will be formatted and used for metadata:
        /dev/nvme0n1
Press 'y' to confirm and continue with the process [y/n]: y
Step: Pre-processing

Result

The command will mount the drive, and the Index Cache and/or the deduplicated database (DDB) will be automatically reconstructed in the new disk, if it is a control node. The automatic reconstruction of DDB may take some time. To speed up the DDB reconstruction, you can manually recover the DDB. For more information on manually recovering the DDB, see Manually Recovering the Deduplication Database.

What to Do Next

Disable root access, if root access was originally disabled in the storage pool. For more information, see Enabling or Disabling Root Access.

×

Loading...