sammkfs, samfsinfo - Constructs or displays information for a StorageTek QFS Software or Oracle HSM file system
∕opt∕SUNWsamfs∕sbin∕sammkfs
[-a
allocation_unit]
[-i
inodes]
[-A
]
[-P
]
[-p
]
[-S
]
[-V
]
fs_name
∕opt∕SUNWsamfs∕sbin∕samfsinfo
[-l
]
fs_name
SUNWqfs
SUNWsamfs
The sammkfs
command
creates a StorageTek QFS or Oracle HSM
file system from the disk partitions
that belong to the family set fs_name, where
fs_name
is the family set name as defined in the mcf
file.
Up to 252 disk partitions can be specified in the mcf
file for a StorageTek QFS or Oracle HSM file system.
The sammkfs
command can also be used to recreate a file system after
a disaster.
The sammkfs
command can create either a file system that
is backwards compatible with previous releases, or a file system
that has new features, but is not compatible with previous releases,
which is the default. See -P
parameter
below for details on the new features, and how to create an older version
file system.
The sammkfs
command can create either NFSv4 or POSIX style ACL
file system in release 11 of Solaris or beyond. The NFSv4 ACL is
default for 5.4 or higher release, use the -p option a POSIX style
ACL file system.
The sammkfs
command aligns the block allocation bit maps and
round robins them on the metadata devices for improved performance.
This behavior is backwards compatible with previous releases.
The option feature Aligned Maps
is set.
The samfsinfo
command displays the structure of an
existing StorageTek QFS or Oracle HSM file system.
The output is similar to that obtained
by using the -V
option to the sammkfs
command.
These commands accept the following options:
-a
allocation_unitSpecifies the disk allocation unit (DAU). The DAU is the basic unit of online storage. When you specify a DAU size, you specify the number of 1024-byte (1 kilobyte) blocks to be allocated for a file.
The DAU size you can specify depends on the type of file system being initialized, as follows:
The Oracle HSM file system is an ms
file system.
The disk devices in it are all md
devices.
Both data and metadata are written to the md
devices.
The allocation_unit specifies the DAU to be
used for the md
devices. Possible allocation_unit
specifications are 16
, 32
, or 64
(the default).
The StorageTek QFS or Oracle HSM file systems are ma
file systems.
The metadata in these file systems is written to mm
devices.
The disk devices in these file systems are specified as
either md
, mr
, or g
XXX devices, as follows:
For the md
devices, possible allocation_unit
specifications are 16
, 32
, or 64
(the default).
A single file system cannot have md
devices mixed among
the mr
and g
XXX devices.
For mr
devices, the DAU is fully adjustable.
Specify an allocation_unit that is a multiple of 8 in the
following range for mr
devices:
8 ≤ allocation_unit ≤ 65528. The default is 64
.
For g
XXX devices, which specify striped groups, the
DAU is fully adjustable.
If the file system contains striped groups,
the minimum unit of disk space allocated
is the DAU multiplied by the number of members in the striped group.
Specify an allocation_unit that
is a multiple of 8 in the following range for g
XXX devices:
8 ≤ allocation_unit ≤ 65528. The default is 256
.
You can mix mr
and g
XXX devices
in a single StorageTek QFS or Oracle HSM file system.
If these device types are mixed, the allocation_unit specified is
used for both device types. If no allocation_unit is specified,
the DAU size used for each type of device is 256
.
-i
inodesSpecifies the number of inodes to be allocated for this file system. This is the total number of user inodes that can be used for the life of this file system. In StorageTek QFS and Oracle HSM file systems, a number of inodes are reserved for file system usage, and are unavailable to the user. This number is in addition to the specified number of user inodes. The actual number of inodes available vary from that specified, due to rounding to metadata DAU size.
NOTE: By specifying this option, you eliminate the possibility of ever increasing the number of inodes for the file system. Therefore, Oracle does not recommend the use of this option.
When this option is specified, later use
of the samgrowfs
(1m) command increases the size of
the file system, but it cannot increase the number of allowable
inodes.
For more information on enlarging file systems, see the WARNINGS section
of this man page and the samgrowfs
(1m) man page.
-A
Uses NFSv4 ACL style for the filesystem ACLs instead of POSIX ACL style. This feature is available only in releases of Solaris beyond Solaris 10.
This option is obsolete and will be removed in the future release since the default ACL style is NFSv4 for 5.4 and higher.
-P
Specifies that a previous version of the file system be created.
This version creates a file system which is compatible with
Oracle HSM version 6.0. This version cannot use the following feature
however: extended attributes on metadata partition.
Without the -P
parameter (default), this feature is available.
-p
Uses POSIX ACL style for the filesystem ACLs instead of NFSv4 ACL style.
-S
Indicates that this file system is shared.
In order to mount the file system as a StorageTek QFS shared file system,
you must also create a hosts.
fs_name configuration file.
For more information on this configuration file and other
aspects of the StorageTek QFS shared file system,
see the Oracle HSM Documentation Library.
For information on configuring a hosts
file, see the hosts.fs
(4) man page.
-V
Writes configuration information to standard output but does not
execute the sammkfs
command.
This information can be used to create a new file system.
The samfsinfo
command should be used to generate
configuration information for an existing file system.
-l
Prints the FUID mappings that exist in this file system. This option is only valid for mounted file systems on Solaris 11.
Example 1. The following command creates Oracle HSM file system with a DAU size of 128 kilobytes:
server# sammkfs -a 128 samfs1
∕etc∕opt∕SUNWsamfs∕mcf
The configuration file for a StorageTek QFS or Oracle HSM file system
As with creating any type of file system, if you specify the wrong partition names, you risk damaging user or system data. Be sure to specify partitions that are otherwise unused on your system. Do not use overlapping partitions.
Be careful when using the -i
inodes option for this command.
By using this option, you dictate the maximum number of inodes allowed
for the life of this file system. This eliminates the possibility
of ever using the samgrowfs
(1m) command to increase the number
of files in this file system. After a file system is made
with -i
specified, the samgrowfs
(1m) command can only be
used to increase the size of the file system in terms of bytes.
Data alignment refers to matching the allocation unit of the RAID controller with the allocation_unit of the file system. A mismatched alignment causes a read-modify-write operation for I∕O that is less than the block size. The optimal alignment formula is as follows:
allocation_unit = RAID_stripe_width * number_of_data_disks
For example, if a RAID-5 unit has a total of 8 disks with 1 of the 8 being the parity disk, the number of data disks is 7. If the RAID stripe width is 64 kilobytes, then the optimal allocation_unit is 64 * 7 = 448.