mmvdisk filesystem command

Availability

Available on all IBM Storage Scale editions.

Description

Use the mmvdisk filesystem command to create mmvdisk file systems using IBM Storage Scale RAID vdisk sets. The mmvdisk filesystem command can also be used to add vdisk sets to an mmvdisk or IBM Storage Scale file system; to delete vdisk sets from an mmvdisk file system; to delete an mmvdisk file system in its entirety; to change an mmvdisk file system; and to list mmvdisk file systems and show the vdisk sets they contain.

An mmvdisk file system is an IBM Storage Scale file system where each vdisk NSD in the file system is from a vdisk set. The mmvdisk filesystem command applies IBM Storage Scale RAID best practices to the management of file systems constructed from vdisk sets. One important such best practice is that the member vdisk NSDs of a given vdisk set can belong only to one file system, but a file system can contain multiple vdisk sets.

The mmvdisk filesystem create command takes a file system device name and one or more vdisk sets, and creates an IBM Storage Scale file system from the member vdisk NSDs of the vdisk sets. The file system device name can not already be in use, and the member vdisk NSDs of the vdisk sets must be created. File system NSD usage, storage pools, and block sizes are set according to the vdisk set attributes. File systems built from vdisk NSDs always use the -j scatter option of the mmcrfs command. File system failure groups for the vdisk NSDs are automatically assigned appropriate values, depending on the data and metadata replication values and the node classes and recovery groups in the vdisk sets. If the vdisk sets and maximum replication values meet certain conditions, custom failure group values can be assigned on a per-node class basis using the --failure-groups option.

Additional IBM Storage Scale file system options supported by the mmcrfs command (for example, quota enablement or a custom mount point) can be specified using the --mmcrfs flag: Any parameters on the command line that appear after the --mmcrfs flag are not interpreted by mmvdisk, but are instead passed unchanged to the mmcrfs command. However, the mmvdisk filesystem create command will inspect these parameters to ensure that -j cluster is not specified, and the -m and -r default replication settings and the -M and -R maximum replication settings will, if present, be used to guide the assignment of automatic failure group values. If replication settings are not specified when creating a file system, the -m and -r default replication settings are both 1 and the -M and -R maximum replication settings are both 2. Note that the IBM Storage Scale maximum replication settings cannot be changed after a file system is created.

If the file system metadata and data maximum replication settings are both equal to 2 or are both equal to 3, it is possible to use the --failure-groups option to assign custom failure group values to the file system vdisk NSDs on a per-node class basis. The node class serves as an identifier for the single point of failure (a failure group) of the vdisk NSDs in the associated scale-out recovery group or in the two paired recovery groups of the node class. To assign custom failure group values, the number of node classes in use across the file system must be equal to or greater than the maximum replication setting, and the number of distinct failure group values assigned to the node classes must also be equal to or greater than the maximum replication setting. The failure group FG values must be a whole number from 1 to 65535.

Use the --trim auto option with the mmvdisk filesystem create command to enable vdisk NSD trim for capable file system vdisk set member NSDs. The default value is consistent with the corresponding declustered array for each vdisk NSD. See IBM Storage Scale RAID: Administration for a description of the vdisk NSD trim function.

The mmvdisk filesystem add command is an interface between vdisk sets, mmvdisk file systems, and the IBM Storage Scale mmadddisk command. Use the mmvdisk filesystem add command to add new or newly extended vdisk sets to an existing mmvdisk file system. The member vdisk NSDs of the vdisk sets must be created using mmvdisk vdiskset create before they can be added to a file system. When a vdisk set has been extended by adding recovery groups, the mmvdisk filesystem add command only adds those member vdisk NSDs that are not already in the file system. Custom failure group assignments can only be made for vdisk NSDs from node classes that are not already represented in the file system; if the node class for a new vdisk NSD is already present in the file system, new vdisk NSDs from that node class must use the existing failure group value. To change the failure group values of existing node classes, the mmvdisk filesystem change command can be used after vdisk sets are added.

The mmvdisk filesystem add command can also add vdisk sets to an existing non-vdisk IBM Storage Scale file system. In this case, the file system becomes an mmvdisk file system with both non-vdisk NSDs and vdisk set NSDs. Custom failure group assignments are permitted if the same conditions are met as required by mmvdisk filesystem create.

The mmvdisk filesystem add command can also add non-vdisk NSDs to an mmvdisk file system. In this case, the file system becomes an mmvdisk file system with both non-vdisk NSDs and vdisk set NSDs.

Use the --trim auto option with the mmvdisk filesystem add command to enable vdisk NSD or non-vdisk NSD trim for capable file system vdisk NSDs or non-vdisk NSD that are being added. If the file system already contains members of a specified vdisk set that have vdisk NSD trim enabled, new members from the same vdisk set will automatically have vdisk NSD trim enabled. Otherwise, the default value is consistent with the corresponding declustered array for each vdisk NSD. See IBM Storage Scale RAID: Administration for a description of the vdisk NSD trim function.

Use the --pools option with the mmvdisk filesystem add command to set the property of pool, which is introduced by newly added NSD.

Use the mmvdisk filesystem delete command to delete vdisk sets or recovery groups or an individual scale-out recovery group node from an mmvdisk file system, or to delete an mmvdisk file system in its entirety. When deleting vdisk sets or entire recovery groups from an mmvdisk file system, the mmvdisk filesystem delete command is an interface between vdisk sets, mmvdisk file systems, and the IBM Storage Scale mmdeldisk command is specified without the --confirm option. When deleting an mmvdisk file system in its entirety, the mmvdisk filesystem delete command is an interface between vdisk sets, mmvdisk file systems, and the IBM Storage Scale mmdelfs command is specified without the --confirm option.

Deleting an individual scale-out recovery group server node from an mmvdisk file system should only be performed as a prelude to deleting the entire node from its scale-out recovery group. For each specified vdisk set with members in the specified recovery group, mmdeldisk is used to remove the vdisk NSDs from the two highest numbered log groups of the recovery group (note that vdisk NSDs are not specific to server nodes, but instead belong to log groups of the recovery group, and therefore the specified node does need to be active to have "its" vdisk NSDs deleted). Until the node is itself entirely deleted from the recovery group, these vdisk NSDs will be orphaned from the file system that their vdisk set belongs to.

It is not always possible to delete a node from a scale-out recovery group or file system. The reallocated disk capacity when the server node's disks are removed must be able to accommodate the remaining log groups and vdisk set members, and the number of pdisks remaining in the declustered arrays must be sufficient to maintain vdisk RAID code width and VCD (vdisk configuration data) fault tolerance. The mmvdisk filesystem delete -N Node command will check first to see that these conditions will be satisfied before proceeding to delete a server node from a file system.

Deleting vdisk sets from a file system will always ask for confirmation, since the underlying mmdelvdisk command can take a very long time to migrate data off the vdisk NSDs that are being deleted. This prompt cannot be bypassed.

Deleting an entire mmvdisk file system will always ask for confirmation, since this will destroy all file system data. This prompt can be bypassed and answered in the affirmative with the --confirm flag.

Use the mmvdisk filesystem change command to reassign the failure group values for an mmvdisk file system. To have mmvdisk assign the failure group values automatically, use the --failure-groups flag without any arguments. Custom failure group assignments are permitted on a per-node class basis using the --failure-groups NodeClass=FG[,NodeClass=FG...] parameter, provided that the same conditions are met as required by mmvdisk filesystem create. In either case, changing the failure group assignments for a file system should be followed by mmrestripefs -r to ensure proper replication using the new failure group assignments.

Use the --trim {auto | no} option with the mmvdisk filesystem change command to change whether vdisk NSD trim is enabled or disabled for capable file system vdisk set member NSDs. See IBM Storage Scale RAID: Administration for a description of the vdisk NSD trim function.

Use the --trim {auto | no} option with the mmvdisk filesystem change command to change whether non-vdisk NSD trim is enabled or disabled also.

Use the --register option with the mmvdisk filesystem change command to make sure that the association between a file system and its vdisk sets is properly registered by the mmvdisk command. This is useful for certain unusual mixed vdisk and non-vdisk file systems, or if somehow the association between a file system and its vdisk sets is not registered.

Use the mmvdisk filesystem list command to list mmvdisk file systems and the file system NSD information for their vdisk sets.

The mmvdisk filesystem command manages only the vdisk NSD aspects of an mmvdisk file system. Other IBM Storage Scale file system attributes continue to be managed by other IBM Storage Scale file system commands; for example, the mmlsfs and mmchfs commands.

Parameters

mmvdisk filesystem create

Create mmvdisk file systems using IBM Storage Scale RAID vdisk sets.

mmvdisk filesystem add

Add IBM Storage Scale RAID vdisk sets to mmvdisk file systems.

mmvdisk filesystem delete

Delete IBM Storage Scale RAID vdisk sets from mmvdisk file systems, or delete an mmvdisk file system in its entirety.

mmvdisk filesystem change

Change the failure group assignments of an mmvdisk file system.

mmvdisk filesystem list

List mmvdisk file system or their component IBM Storage Scale RAID vdisk sets.

--file-system FsName

Specifies the name of the file system to be operated upon.

--vdisk-set VdiskSet[,VdiskSet...]

Specifies the vdisk sets to be affected by a file system operation.

--nsd nsdName[,nsdName...]

Specifies the non-vdisk NSD to be affected by a file system operation.

--mmcrfs

With the mmvdisk filesystem create command, the --mmcrfs flag indicates that all following command line parameters are not to be interpreted by mmvdisk, but are instead to be passed to the IBM Storage Scale mmcrfs command.

--recovery-group RgName

With the mmvdisk filesystem delete command, restricts vdisk set deletion to specified recovery group. This is required preparation for removing a recovery group from a vdisk set definition.

-N Node

With the mmvdisk filesystem delete --recovery-group RgName command, restricts vdisk set deletion to a single node of the specified scale-out recovery group. This is required preparation for removing the server node from a scale-out recovery group.

--failure-groups

With the mmvdisk filesystem change command, performs an automatic reassignment of the failure groups for a file system's vdisk NSDs.

--failure-groups NodeClass=FG[,NodeClass=FG...]

With the mmvdisk filesystem create, mmvdisk filesystem add, and mmvdisk filesystem change commands, assigns custom failure group values to the vdisk NSDs from the specified recovery group node classes. A failure group FG value must be a whole number from 1 to 65535.

--pools Pool,key=value[,key=value ...][;Pool,key=value[,key=value ...]...]]

With the mmvdisk filesystem add, assign the properties of pool.

--trim {auto | no}

With the mmvdisk filesystem create, mmvdisk filesystem add, and mmvdisk filesystem change commands, specify if the vdisk NSD trim function is enabled (auto) or disabled (no) for capable file system vdisk set member NSDs. The --trim auto option is ignored for any targeted vdisk set NSDs that are not capable of vdisk NSD trim. See IBM Storage Scale RAID: Administration for a description of the vdisk NSD trim function.

--register

With the mmvdisk filesystem change command, register any missing associations between a file system and its vdisk sets. If no change is required, the --register option will simply report that the associations between a file system and its vdisk sets are properly registered.

--confirm

Confirms the deletion of an entire or vdiskset of mmvdisk file system. This flag bypasses the interactive prompt that would otherwise be printed to confirm file system deletion.

-Y

Specifies that the mmvdisk filesystem list command produce colon-delimited raw output.

Exit status

0

Successful completion.

nonzero

A failure has occurred.

Security

You must have root authority to run the mmvdisk filesystem command.

The node on which the command is issued must be able to execute remote shell commands on any other node in the cluster without the use of a password and without producing any extraneous messages.

See also

Location