Using the esxcli stratus mpm Command

The esxcli stratus mpm command provides a control interface for creating and managing MPM volumes on internal NVMe disks for systems running VMware ESXi and Automated Uptime Layer with Smart Exchange.

To use the esxcli stratus mpm command, open a connection to the ESXi Shell on the ESXi host and run a command in the following format:

# esxcli stratus mpm subcommand [options]

Use the --help or -h argument to view the command usage.

# esxcli stratus mpm -h
Usage: esxcli stratus mpm {cmd} [cmd options]
 
Available Commands:
  add                   Add a member disk to a volume.
  assemble              Assemble a volume or volumes.
  create                Create a volume.
  examine               Examine a disk's raid superblock
  fail                  Set a member disk faulty.
  info                  Show disk or volume info.
  list                  List mirrored sets of internal disks
  remove                Remove a member disk from a volume.
  speedLimit            Show or set sync speed.
  stop                  Stop a volume.
  uuid                  Show a volume or disk's UUID, or --scan to see all
  zeroSuperblock        Zero a disk's raid superblock.

Some of the esxcli stratus mpm subcommands require you to specify a volume name (in the format -v|--volume=volume) or a disk name (in the format -d|--disk=disk), or both. For example, the following commands add the disk vmhba4:C0:T0:L0 to volume mpm1.

# esxcli stratus mpm add -v mpm1 -d vmhba4:C0:T0:L0
# esxcli stratus mpm add --volume=mpm1 --disk=vmhba4:C0:T0:L0

For information about determining the volume names and disk names to use with this command, see Disk Drives.

For some subcommands that require you to specify volumes and/or disks, you can instead specify the -s|--scan argument. This argument, which causes esxcli stratus mpm to apply the subcommand to all of the volumes or disks to which it applies, has the following format:

# esxcli stratus mpm subcommand --scan

For example, the following command scans for any internal disks that have MPM volume metadata but that are not currently running as MPM volumes, and if possible, assembles those disks into running MPM volumes:

# esxcli stratus mpm assemble --scan

To display a summary of the esxcli stratus mpm subcommands while you are logged in to a system, run the following command:

# esxcli stratus mpm --help

For additional information about each subcommand, specify the subcommand you want to use. For example, to learn the syntax for the assemble subcommand, run the following command:

# esxcli stratus mpm assemble --help

esxcli stratus mpm Subcommands

A description of the esxcli stratus mpm subcommands follows.

Adds an unused disk to a volume.

Assembles (starts) one or more existing volumes that are not currently running. You can optionally specify the -s argument instead of volume or disk arguments to automatically find and assemble all such volumes.

Note: Specify either the -d and -v argument or the -s argument; you cannot specify both.

Creates an MPM volume. Typically, you specify two disks, resulting in an MPM volume that is a RAID-1 protected (mirrored) volume. Specifying one disk creates a simplexed MPM volume that is not protected until you add a second disk.

Note: When you create an MPM volume, it is best practice to specify two disks in identical slots of each compute module and then name the volume mpmn, where n is the slot number of the disks. For example, create an MPM volume named mpm2 for the disks in slots A/2 and B/2.
Notice: Creating an MPM volume permanently deletes existing data on the target disks. Existing data is not preserved.

Displays a disk's RAID superblock, which contains information that identifies the disk as a member of an MPM volume.

Forces a member disk to be marked as faulty. To ensure data integrity, you can remove only a disk that is currently marked as faulty. If any disks fail, the system automatically marks them as faulty, but this subcommand allows you to safely remove disks that the system has not already marked as failed.

Displays one or more specific disk or volume information items:

  • The -d or --device argument displays information for one disk or one volume.

  • The -s or --scan argument displays information for all disks or all volumes.

Note: Specify either the -d argument or the -s argument; you cannot specify both.

  • The -t argument specifies the type of information to display for disk or volume:

  • endurance_remaining — A countdown of an NVMe disk's lifespan, ranging from 100% for a new disk to 0% for an expired disk.

  • product — The disk's product name (for example, MTFDKCC6T4TGQ-1BK1JABYY).

  • revision — The disk's revision number (for example, E3MQ003).

  • serial — The disk's serial number (for example, 41131021A72C).

  • size — The disk's size, in bytes (for example, 6401252745216).

  • state — The disk's state. Possible values are in_sync, syncing, and faulty.

  • temperature — The disk's temperature, in Celsius (for example, 27C).

  • vendor — The disk's vendor. This value is always NVMe.

  • volume — The volume's name (for example, mpm4).

  • disks — The names of the disks in the volume (for example, vmhba9:C0:T0:L0,vmhba10:C0:T0:L0).

  • total_disks — The number of disks in the volume (for example, 2).

  • raid_disks — The number of RAID disks in the volume (for example, 2).

  • volume_size — The volume's size, in bytes (for example, 6251223040).

Displays all volumes and disks present in the system, if you omit the -v or -d argument, or displays only the specified volume or disk.

Removes a member disk from a volume. Before you remove a disk, the disk to remove must be marked as faulty, either manually by running the esxcli stratus mpm fail command, or automatically by the system because an error occurred on the disk.

Displays or sets the global synchronization speed limits. The speeds range from 1 to 16. To set the default value, which is 8, specify either 8 or 0.

Stops a volume, removing it from usage and making its disks unused disks.

Note: Stopping a volume does not permanently remove the volume or its data. The volume and its data are restored either upon the next boot or when the disks are hot-unplugged and reinserted, unless a new volume is created or a zeroSuperBlock subcommand is specified on the volume's disks.

Displays a volume or disk's unique ID (UUID), which is used to identify disks that belong to the same volume. You can optionally specify the -s argument instead of a volume or disk argument to display the UUIDs for all volumes that are present in the system.

Note: Specify only one of the following: the -d argument, the -v argument, or the -s argument.

Erases a disk's RAID superblock to remove any indication that the disk belonged to a volume. The volume's data is not guaranteed to be preserved on the disk or securely erased from the disk.