Skip to main content
FlashGrid.io

Reducing ASM disk size (AWS)

  • Updated
  • Products: FlashGrid Cluster on AWS

Oracle requires that all disks in an ASM disk group should be of the same size (link). Therefore, in-place resize is not possible and the procedure below should be followed instead.

  1. Create and attach new EBS volumes of the desired size to the database node EC2 instances. Attach the volumes using disk names xvdba through xvdbz - these disks will be shared automatically by FlashGrid Storage Fabric.

  2. Confirm FlashGrid names of the new drives, e.g. rac2.xvdbb

    $ flashgrid-cluster drives

    If the new drives are not listed, then check that the corresponding devices (e.g. /dev/xvdbb) are visible in the OS. If they are visible in the OS, then as root user, run flashgrid-node reload-config and check the output of flashgrid-cluster drives again. If they are not visible in the OS, then double-check that you have attached them with the correct device names.

  3. Run the following command to replace all regular (non-quorum) disks in a disk group:

    $ flashgrid-dg replace-all-disks -G <diskgroup_name> -d <disk> [<disk> ...]

    For example, to replace all regular disks in the DATA disk group with xvdbc, xvdbd:

    $ flashgrid-dg replace-all-disks -G DATA -d /dev/flashgrid/rac[12].xvdb[cd]

    This command relocates data from the old disks to the new ones and runs rebalance. It can generate significant I/O load, so plan this procedure during periods of lower database activity.

    Wait till the Resync=No for the disk group in the flashgrid-cluster output. It may take a while to resync the disk group depending on the amount of data, disk performance characteristics, and instance parameters.

    Note: Starting with Storage Fabric version 23.04, the replace-all-disks operation executes data rebalance in the background. While rebalance is running in the background, the flashgrid-cluster and flashgrid-dg commands may show old disks reported as "LostDisks". This will be cleared after the background operation is completed.

  4. The old disks can be removed. Please run for each DB node:

    [fg@<db_node> ~] flashgrid-node stop-target <disk>

    For example, to stop disk rac1.xvdbb on rac1, run:

    [fg@rac1 ~] flashgrid-node stop-target /dev/flashgrid/rac1.xvdbb

    To stop disk rac2.xvdbb on rac2, run:

    [fg@rac2 ~] flashgrid-node stop-target /dev/flashgrid/rac2.xvdbb
  5. Detach the old EBS volumes from the EC2 instances.