mmdefragfs command

Availability

Available on all IBM Storage Scale editions.

Description

Use the mmdefragfs command to reduce fragmentation of a file system. The mmdefragfs command moves existing file system data within a disk to make more efficient use of disk blocks. The data is migrated to unused sub-blocks in partially allocated blocks, thereby increasing the number of free full blocks.

The mmdefragfs command can be run against a mounted or unmounted file system. However, best results are achieved when the file system is unmounted. When a file system is mounted, allocation status may change causing retries to find a suitable unused sub-block.

Parameters

Device

The device name of the file system to have fragmentation reduced. File system names need not be fully-qualified. fs0 is as acceptable as /dev/fs0.

This must be the first parameter.

-P PoolName

Specifies the pool name to use.

-N {Node[,Node...] | NodeFile | NodeClass}

Specifies the nodes that can be used in this disk defragmentation. This parameter supports all defined node classes. The default is all or the current value of the defaultHelperNodes parameter of the mmchconfig command.

For general information on how to specify node names, see Specifying nodes as input to GPFS commands.

--pit-continues-on-error

Allows the mmdefragfs command to continue defragmenting the remaining files, if errors are encountered in the parallel inode traverse (PIT) phase that performs defragmentation of subblocks. An output file is generated if an error occurs in the PIT phase. The location of the file that logs the errors is displayed in the command output.

Note:

• The mmdefragfs command continues to run only if the error is reported in the PIT phase of the command execution. If the error is reported in other phases of command execution, the command stops running.
• The --pit-continues-on-error option works only if the minimum release level of the IBM Storage Scale cluster is 5.1.1 or later.

The --pit-continues-on-error option allows the command to skip fatal errors encountered during the PIT phase and continue with the file system scan. If fatal errors are encountered, the command fails at the end of the scan and generates an output file that lists all the problematic files that had the fatal errors. Based on such output file, users can take additional actions to handle these problematic files before rerunning the command.

If --pit-continues-on-error is not used, the command run fails if a fatal error is encountered. Any error that is not included in the next table is considered as a fatal error.

Table 1. Non-fatal errors that can be encountered during the PIT phase

Non-fatal error	Code	Description
E_COMPRESS_NOT_ALLOWED	840	File compression is not allowed.
E_COMPRESS_HYPERALLOC	799	Compression of a file in hyperalloc mode is not allowed.
E_NOREPLGRP	219	Not enough replicas could be created because the desired degree of replication is larger than the number of failure groups.
E_NOREPLSPC	220	Not enough replicas could be created because there was not enough space left in one of the failure groups.
E_NOBALSPC	221	There was not enough space left on one of the disks to properly balance the file according to the current stripe method.
E_NOBALAVAIL	222	The file could not be properly balanced because one or more disks are unavailable.
E_ADDR_BROKEN	226	All replicas were on disks that have since been deleted from the stripe group.
E_ALL_UNAVAIL	218	A replicated read or write failed because none of the replicas were available.
E_ENC_CTX_NOT_READY	786	Encryption context is not ready.

--qos QosClass

Specifies the Quality of Service for I/O operations (QoS) class to which the instance of the command is assigned. If you do not specify this parameter, the instance of the command is assigned by default to the maintenance QoS class. This parameter has no effect unless the QoS service is enabled. For more information, see the topic mmchqos command. Specify one of the following QoS classes:

maintenance

This QoS class is typically configured to have a smaller share of file system IOPS. Use this class for I/O-intensive, potentially long-running IBM Storage Scale commands, so that they contribute less to reducing overall file system performance.

other

This QoS class is typically configured to have a larger share of file system IOPS. Use this class for administration commands that are not I/O-intensive.

For more information, see Setting the Quality of Service for I/O operations.

Options

-i

Specifies to query the current disk fragmentation state of the file system. Does not perform the actual defragmentation of the disks in the file system.

-u BlkUtilPct

The average block utilization goal for the disks in the file system. The mmdefragfs command reduces the number of allocated blocks by increasing the percent utilization of the remaining blocks. The command automatically goes through multiple iterations until BlkUtilPct is achieved on all of the disks in the file system or until no progress is made in achieving BlkUtilPct from one iteration to the next, at which point it exits.

Exit status

0

Successful completion.

nonzero

A failure has occurred.

Security

You must have root authority to run the mmdefragfs 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. For more information, see Requirements for administering a GPFS file system.

See also

Location