XFS file system mount parameter Translation

Installation Options:
Allocsize = size
Set the pre-allocated size of files on the buffer I/O end. When the delay is allocated for writing (the default size is
64kib ). Page size of this option (usually 4 kib) to 1 gib, including valid values, power 2
Incremental.

Attr2 | noattr2
This option is enabled/disabled (enabled by default) and the "opportunistic" inline method is improved.
The extended attributes are stored on the disk. When the new form is used immediately (by setting or deleting
Extended attributes) the feature bit field of the super block on the disk will be updated to reflect the format in
.

Barrier
The number of blocks allowed to be written to the log or unwritten level transfer layer write barriers. This
Allow the write cache of the drive to be enabled, write barrier device, and support.

Dmapi

Enable dmapi (data management API) event tagging. MTPT option used.

Grpid | bsdgroups and nogrpid sysvgroups
The Group ID defined by these options is obtained from the newly created file. The Group ID required by grpid
Directory, otherwise (default) The fsgid of the current process is required, unless
The directory has a setgid bit. In this case, you need to obtain
Setgid bit, if it is a directory.

Ihashsize = Value
Sets the number of inode hash buckets for the specified mount point in the memory. If
When the value is zero, the selected value is displayed in/proc/mounts by default algorithm.

IKeep | noikeep
When the inode is empty, the inode cluster puts them on the disk (IKeep)-this is the traditional XFS
Action, but it is still the default. With the noikeep option, the inode cluster is returned to idle
Space pool.

Inode64
Indicates that the inode that XFS allows to be created anywhere in the file system, including
The meaning of the inode number that occupies more than 32 places. This is to backward compatibil
So that the problem of backing up the application cannot process large inode numbers.

Largeio | nolargeio
If nolargeio is specified, the best I/O (2) reported by stat in st_blksize will be as small as possible
Allow user applications to avoid inefficient reading/modifying/writing I/O. If largeio is specified, a file system,
If there is a specified swidth st_blksize, the value of swidth is returned (in bytes ). If the file system does not have
Swidth, but allocsize is not specified, then allocsize (in bytes) will be changed to return. If the two
These two options, then the file system, if nolargeio specifies the behavior.

Logbufs = Value
Set the number of log buffers in the memory. Valid numbers range from 2 to 8 ). The default value is 8 buffers.
64kib of the block size of the file system, 32 kib of the file system with four buffers, and 3 Buffers
File Systems with 16 kib and 2 block sizes for all other configured buffers. Increase in number
The memory buffer used in the extra log buffer at the cost may increase the performance of some workloads.
And its related control structure.

Logbsize = Value
Set the size of the log buffer in each memory. You can specify the size (in bytes) (in kilobytes) or use the suffix "K.
Versions 1 and 2 are effective at 16384 (16 K) and 32768 (32 K ). The effective size of version 2nd is record
It also includes 65536 (64 K), 131072 (128 K), and 262144 (256 K ). The default value of the machine exceeds 32 MIB.
The memory is 32768. By default, 16384 is used for machines that use less memory.

Logdev = device and rtdev = Device
Use an External Log (metadata log) and/or a real-time device. The XFS file system has three parts: one data
Part, a log part, and a real-time section. The real-time part is optional, and the log part can be
Separate the data section or include it. See XFS (5 ).

MTPT = mount point
The dmapi option used. The value specified here will be included in the dmapi installation event. It should be
Path] is the actual installation point.

Noalign
Data distribution is not aligned at the strip cell boundary.

Noatime's
When an object is read, the access timestamp is not updated.

Norecovery
The file system will be installed without running log recovery. If the file system is not uninstalled cleanly, it is
The installation may be inconsistent in recovery mode. Some files or directories may not be accessible.
For this reason. Norecovery must install the read-only file system or the installation will fail.

Nouuid does not check whether the dual-mounted file system uses the uuid of the file system. This is very useful for installing LVM snapshot volumes
Ultrafine electrode.

Osyncisosync
O_sync write to implement real o_sync. Without this option, the XFS behavior of Linux, if it is the osyncisdsync Option
This will enable the write o_dsync flag, if the behavior of the file set opened with the o_sync flag
. This may lead to better performance without affecting data security. However, if this option is not
In fact, the timestamp Update written from o_sync may be lost if the system crashes. If the timestamp is updated
Key: the option to use osyncisosync.

Uquota | usrquota | uqnoenforce | quota
Enable the user's disk quota accounting and implement restrictions (optional ). For more information, see xfs_quota (8 ).

Gquota | maid | gqnoenforce
The group enables Disk Quota accounting and implementation restrictions (optional ). For more information, see xfs_quota (8 ).

Pquota | prjquota | pqnoenforce
Project Disk Quota accounting is enabled and restricted (optional. See xfs_quota (8) for more information.
Details.

Sunit = and swidth =
Used to specify the Strip unit and width of a RAID device or volume. Value, must be specified
The block size of 512 bytes is measured in units. If this option is not specified, strip the volume or File System
The stripe width or unit is specified as The mkfs time of the RAID device. Then, the disk is restored after the Mount System Call.
From super value. For file systems, directly on raid devices, these options can be used
If the file system's super block after the basic disk layout is changed has already overwritten Information
Create. The swidth option is required. If the Sunit option has been specified and must be a multiple
Sunit value.

Swalloc
When the data distribution rounds up to the current end of the file whose stripe width range is expanding,
The file size is larger than the width of the entry.