mount_samfs - Mounts a StorageTek QFS or Oracle HSM file system
mount
-F samfs
[generic_options]
[-o
FSType_specific_options]
special |\0 mount_point
mount
-F samfs
[generic_options]
[-o
FSType_specific_options]
special
mount_point
SUNWsamfs
The mount
command attaches
a file system to
the file
system hierarchy at the specified mount_point,
which is the path name of a directory.
This man page describes how to mount a StorageTek QFS
or Oracle HSM file system, and it explains the unique options
that can be used when mounting these file systems.
If the first form of the command is used, which specifies either
a special or a mount_point but not both,
the mount
command searches the ∕etc∕vfstab
file
and fills in missing arguments, including
the FSType_specific_options.
The mount
(1M) command also searches
the ∕etc∕opt∕SUNWsamfs∕samfs.cmd
file for mount options.
For more information on the mount
(1M) command,
see the mount
(1M) man page.
For more information on the ∕etc∕opt∕SUNWsamfs∕samfs.cmd
file,
see the samfs.cmd
(4) man page.
-F samfs
Specifies that the file system being mounted is of type samfs
.
This is a required option if you are mounting a StorageTek QFS
or an Oracle HSM file system. These file systems are all
type samfs
.
One or more generic Solaris file system options. For a list of
possible generic_options, see the mount
(1M) man page.
-o
FSType_specific_optionsA list of mount options specific to file systems of type samfs
.
If specifying multiple options, separate each option with a comma
and no intervening spaces.
For the list of possible -o
FSType_specific_options,
see one or more of the following headings on this man page:
Miscellaneous Tuning Options
I∕O Options
Storage and Archive Management Options
Shared File System Options
Multireader File System Options
StorageTek QFS and Oracle HSM Options
If no FSType_specific_options are specified, the the file system is mounted as a read∕write file system.
If invalid options are specified, a warning message is generated and the invalid options are disregarded.
NOTE: If running the StorageTek QFS software on a Linux client, the available mount options are very limited. The following mount options are the ONLY ones available on a Linux client system: rw, ro, retry, shared, rdlease, wrlease, aplease, minallocsz, maxallocsz, min_pool, meta_timeo, noauto, and auto.
The noauto and auto options are only recognized within the
∕etc∕fstab
file and min_pool only in the samfs.cmd
file.
The maximum value for meta_timeo is 60.
The Family Set Name from the StorageTek QFS
or Oracle HSM master configuration file (mcf
).
For more information on this file, see the
mcf
(4) man page.
The path name or directory at which the file system is to be mounted.
If the mount_point
has any contents prior to the mount
operation, these are hidden until the file system is unmounted.
The following options can be used when mounting a StorageTek QFS or Oracle HSM file system. These options can affect file system features and system performance.
nosam
|\0 sam
The nosam
option mounts a
Oracle HSM file system, but only the file system
functionality is enabled. The archiving, releasing, and
staging functionality is disabled. When a file system
is mounted with this option, the file system
returns ENOSPC
when it reaches 100% capacity.
Note that mounting a file system with the nosam
option offers no data protection for newly created files
or for previously archived files that have been modified.
The default is sam
.
noarscan
|\0 arscan
The noarscan
option disables file system scans,
typically performed by the sam-arfind
daemon, for
finding archive candidates on a mounted file system.
This mount option can be useful for file systems in which
new files are no longer being created yet staging and
releasing are still desired.
The default is arscan
.
nosuid
Mounts the file system with setuid
execution disallowed.
By default, the file system mounts with setuid
execution allowed.
nogfsid
|\0 gfsid
The nogfsid
option disables the setting of a global file system id,
and uses the historical setting of the root slice device type paired
with the file system type.
The gfsid
option enables the setting of a global file system id,
and uses the file system id that is stored in the superblock, which
consists of the file system creation time paired with the hostid.
The default is gfsid
.
nocdevid
|\0 cdevid
The nocdevid
option disables the setting of a global file system
device id, and uses the historical setting of the root slice device type.
The cdevid
option enables the setting of a global file system
device id that consists of the samioc module major number paired with
the file system equipment number as specified in the mcf
(4) file.
The default is cdevid
.
notrace
|\0 trace
The notrace
option disables file system tracing.
The trace
option enables file system tracing.
The default is trace
.
noquota
|\0 quota
The noquota
option disables file system quotas.
The quota
option enables file system quotas,
provided that at least one file system quota file is present.
The default is quota
.
For more information on quotas, see the
StorageTek QFS File System Configuration and Administration
Guide.
sync_meta=
nSpecifies whether or not the metadata is written to the disk every time it changes, as follows:
If sync_meta=0
, metadata is held in a buffer before being written
to disk. This delayed write delivers higher performance.
This is the default for
StorageTek QFS and Oracle HSM file systems that are
not mounted as multireader file systems or as StorageTek QFS shared file
systems.
If sync_meta=1
, metadata is written to disk every time it
changes. This slows performance, but it ensures data consistency.
This is the default for StorageTek QFS file systems that are
mounted as multireader file systems or as StorageTek QFS shared file
systems. In a StorageTek QFS shared file system,
this is the setting that must
be in effect if failover capability is required.
worm_capable
The worm_capable
option allows Write Once Read Many (WORM) files to
be stored in Oracle HSM filesystems. Enabling this feature allows the WORM flag
to be set on files and directories. Once the WORM flag is set, a file's
data and path are immutable and the file can not be deleted until its
retention period expires. In addition, the volume on which the WORM
file resides can not be deleted using sammkfs.
worm_lite
The worm_lite
option is similar to the worm_capable mount option but
eases the restrictions regarding actions that can be taken on WORM-enabled
volumes and retained files. WORM lite enabled volumes can be deleted
using sammkfs. Retained files can be removed before their retention
period expires and their retention period can be shortened (must have
root privileges). File data and path remain immutable.
worm_emul
The worm_emul
option is similar to the worm_capable mount option and
enables WORM "Emulation mode". The difference with this option is
the trigger used to retain files is the transition from a writable to
read-only file. File data and path are immutable after appying the WORM
trigger. A file retained in this mode can not be deleted until it's
retention period expires. Volumes containing WORM emulation mode files
can not be deleted using sammkfs.
emul_lite
The emul_lite
option is similar to the worm_capable mount option and
enables WORM "Emulation Lite mode". The trigger to retain files is
the transition from a writable to read-only file. Retained files can be
removed before their retention period expires and their retention period
can be shortened (must have root privileges). Data and path changes to a
file are immutable after applying the trigger. Emulation lite enabled
volumes can be deleted using sammkfs.
def_retention=
nThe def_retention
option sets the default retention period.
This option requires a WORM mount option enabled. This option sets
the default retention period for files which have the WORM feature
enabled with no supplied retention period. The retention period can
take three forms.
A value of permanent (or 0)specifies permanent retention.
A value of the form MyNdOhPm where
M, N, O, P are arbitrary non-negative integers;
y, d, h, m specify the number of years, days, hours, and minute(s)
respectively.
Note that combinations of this form are allowed, and specifiers may
be omitted, e.g., 5y, 3d1h, 4m.
The final form is a simple integer value in minutes
for n, an integer 1 ≤ n ≤ 2147483647
(231 - 1).
If this option is not supplied, a 30 day (43,200 minute) default
retention period is used.
rd_ino_buf_size=
nrd_ino_buf_size
sets the size of buffer to n. This is the
buffer which is used to read the .inodes file into buffer cache.
For n, specify an integer such that 1024 ≤ n ≤ 16384.
n is in units of bytes and rounded down to the nearest power of 2.
The default is 16384 bytes.
wr_ino_buf_size=
nwr_ino_buf_size
sets the size of the buffer to n. This is
the buffer which is used to synchronously write an inode through to the disk.
For n, specify an integer such that
512 ≤ n ≤ rd_ino_buf_size
.
n is in units of bytes and rounded down to the nearest power of 2.
The default is 512 bytes.
The following options are available for StorageTek QFS and Oracle HSM
file systems.
Also see the mcf
(4) man page.
stripe=
nSets the stripe width for the block-based file system to n disk
allocation units (DAUs). The stripe width means that n * DAU bytes
are written to one data device logical equipment number (LUN)
before switching to the next LUN.
The DAU size is set on the sammkfs
(1m) command's -a
option
when the file system is initialized.
For n, specify an integer such that 0 ≤ n ≤ 255.
If n=0, files are round robined on each slice.
The default n on file systems with an ms
Equipment Type and
on file systems with an ma Equipment Type with no
striped group (g
x) components is as follows:
128 kilobytes∕DAU for DAUs 128 kilobytes
1 for DAUs ≥ 128 kilobytes
By default, n=0 on a StorageTek QFS shared file system.
By default, n=0 on file systems with an ma
Equipment Type with any striped group (g
XXX) components.
NOTE: The system sets stripe=0
if mismatched
striped groups exist.
The following options are available for StorageTek QFS and Oracle HSM file systems. They allow changing the type of I∕O for a file based on I∕O size and history. Note that if direct I∕O is specified for a file, these options are ignored and all I∕O to regular files is direct, if possible. Well-aligned I∕O occurs when the file offset falls on a 512-byte boundary and when the length of the I∕O transfer is at least 512 bytes.
dio_rd_consec=
nSets the number of consecutive I∕O transfers with a buffer size greater
than the specified lower limit (which is dio_rd_form_min
for aligned reads or dio_rd_ill_min
for misaligned reads) to
n operations.
By default, n=0, which means that no default direct reads occur
based on I∕O sizes.
Also, by default, dio_rd_form_min
and dio_rd_ill_min
are ignored.
dio_rd_form_min=
nSets the read well-aligned lower limit to n 1024-byte blocks. By default, n=256, 1024-byte blocks. If n=0, automatic I∕O type switching for well-aligned reads is disabled.
dio_rd_ill_min=
nSets the read misaligned lower limit to n 1024-byte blocks. By default, n=0, which disables automatic I∕O type switching for misaligned reads.
dio_wr_consec=
nSets the number of consecutive I∕O transfers with a buffer size
above the specified lower limit (which is dio_wr_form_min
for aligned writes or dio_wr_ill_min
for misaligned writes)
to n operations.
By default, n=0, which means that no default direct writes occur
based on I∕O sizes.
Also, by default, dio_wr_form_min
and dio_wr_ill_min
are ignored.
dio_wr_form_min=
nSets the write well-aligned lower limit to n 1024-byte blocks. By default, n=256 1024-byte blocks. Setting n=0 disables automatic I∕O type switching for well-aligned writes.
dio_wr_ill_min=
nSets the write misaligned lower limit to n 1024-byte blocks. By default, n=0, which disables automatic I∕O type switching for misaligned writes.
atime=
\fI-1 |\0 0 |\01The file system is mounted by default with cached access time recording
(atime = 0
). This means access time updates to disk are
deferred for up to 1 minute after the file is last accesssed.
Note, the file access time is immediately updated on disk if SAM
is enabled and the space used is above the low water mark
or when the access time coincides with updates to the ctime or mtime.
See stat (2). The access time is also updated when the
the file system is unmounted.
If atime = 1
, the file system will always update access time on disk.
If atime = -1
, the file system will not update
access time except when it coincides with updates to the ctime or mtime.
See stat (2). The atime = -1
option reduces disk activity on file
systems where access times are unimportant (for example, a Usenet news spool).
Note, atime = -1
, should not be set when SAM is enabled.
The POSIX standard requires that access times be marked on files.
Note, for atime = 0
(the default),
the current access time may not be updated on disk in case
of an interruption.
noatime
The noatime
is added to be compatible with other file systems.
If noatime
is specified, atime = -1 will be set. This means
the file system will not update access time except when it coincides
with updates to the ctime or mtime. See stat (2).
Note, noatime
, should not be set when SAM is enabled.
forcedirectio
Specifies direct I∕O as the default I∕O mode.
This means that data is transferred directly between the user's
buffer and disk.
The forcedirectio
option should be specified only if
the file system is used for large block aligned sequential I∕O.
For more information, see the directio
(3C), setfa
(1),
sam_setfa
(3x), and sam_advise
(3) man pages.
The default I∕O mode is buffered (uses the page cache).
nodio_szero
|\0 dio_szero
The dio_szero
option causes uninitialized areas of sparse
files written with direct I∕O to be zeroed when the area is accessed.
This makes the sparse file behavior the same as that for paged I∕O.
By default, sparse files written by direct I∕O do not have the
uninitialized areas zeroed for performance reasons.
The default is nodio_szero
.
force_nfs_async
Causes the file system to cache nfs data written to the server even if nfs
has requested that the data be written synchronously through to disk.
The force_nfs_async
option is only useful if
the file system is mounted as a nfs server and the clients
have set the nfs mount option noac
. The default nfs noac
behavior without force_nfs_async
causes data to be
synchronously written through to disk. Caution, the force_nfs_async
option violates the nfs protocol and should be used with care.
Data may be lost in the event of a server interruption.
Also, data is cached on the server and will
not be immediately seen by all the clients if there are multiple nfs
servers. Multiple nfs servers can be enabled with Shared QFS.
sw_raid
Causes the file system to align the writebehind buffer. This option should be set if the software raid feature of packages such as Solstice DiskSuite is being used on this file system. This option is off by default.
readahead=
nSets the maximum readahead
value to n.
The readahead
option specifies the maximum number of bytes
that can be read ahead by the file system.
n is in units of kilobytes and must be a multiple of 8.
For n, specify an integer such that 0 ≤ n ≤ 16777216.
The default is 1024 (1,048,576 bytes).
writebehind=
nSets the maximum writebehind
value to n.
The writebehind
option specifies the maximum number of bytes
that can be written behind by the file system.
n is in units of kilobytes and must be a multiple of 8.
For n, specify an integer such that 8 ≤ n ≤ 16777216.
The default is 512 (524,288 bytes).
flush_behind=
nSets the maximum flush_behind
value to n.
When enabled, modified pages that are being written sequentially
are written to disk asynchronously to help the Solaris VM layer
keep the pages clean.
This option sets the maximum flush_behind
value to n.
n is in units of kilobytes.
For n, specify an integer such that 0 ≤ n ≤ 8192.
The default is 0, which disables flush behind.
wr_throttle=
nSets the maximum number of outstanding write bytes for one filesystem to n kilobytes. If n = 0, there is no limit.
The default is 5% of system memory. Using the 5% formula, and given the memory size on the left, the wr_throttle setting is on the right:
1 GB 51 MB 4 GB 205 MB 16 GB 819 MB 64 GB 3.2 GB
qwrite
Enables
simultaneous reads and writes to the same file from different threads.
Specify this option only if users of the
file system handle multiple simultaneous transactions to the same file.
For example, this is useful for database applications.
This option improves I∕O performance by queuing multiple requests at
the drive level.
By default, qwrite
is not enabled, and the file system
disables simultaneous reads and writes to the same file.
This is the mode
defined by the UNIX vnode interface standard that gives exclusive access
to only one writer and forces other writers and readers to wait.
The qwrite
option is disabled for NFS reads or writes of the
file system.
noabr
|\0 abr
For Oracle RAC with Oracle HSM AIO only. Disable (enable) Application Based Recovery of software mirrors. Applies only to Oracle HSM filesystems built on Solaris Volume Manager mirrored volumes that likewise support Application Based Recovery. Default is enabled.
nodmr
|\0 dmr
For Oracle RAC with Oracle HSM AIO only. Disable (enable) Directed Mirror Reads of software mirrors. Applies only to Oracle HSM filesystems built on Solaris Volume Manager mirrored volumes that likewise support directed mirror reads. Default is enabled.
The following options can be used when mounting an Oracle HSM file system. These options pertain to the storage and archive management facilities of these file systems.
nofsalog
|\0 fsalog
The nofsalog
option disables file system activity logging for this file
system.
The fsalog
option enables file system activity logging for this file
system.
The sam-fsd daemon starts sam-fsalogd which logs file system activity.
This option is automatically enabled if the sam_db
option is selected.
For more information, see the fsalogd.cmd
(4) man page.
The default is nofsalogd
.
nosam_db
|\0 sam_db
The nosam_db
option indicates there is no associated database with
this file system.
The sam_db
option indicates there is an associated database with
this file system and file system activity logging is enabled. The sam-fsd
daemon starts sam-fsalogd which logs file system activity and sam-dbupd
which updates the associated database.
For more information, see the fsalogd.cmd
(4) man page.
The default is nosam_db
.
high=
nSets the high-water mark for disk cache utilization to n
percent. When the amount of space used on the disk cache reaches
n percent,
the Oracle HSM file systems
start the releaser process. For more information, see the
sam-releaser
(1m) man page. If n is set to 100, releaser
is not started and ENOSPC
is returned.
The default is 80.
low=
nSets the low-water mark for disk cache utilization to n percent. When the amount of space used on the disk cache reaches n percent, the Oracle HSM file system starts the releaser process, which stops releasing disk space. The default is 70.
partial=
nSets the default partial release size for the file system to n
kilobytes. The partial release size is used to determine how many
bytes at the beginning of a file marked for partial release should be
retained on disk cache when the file is released. The user can
override the default on a file-by-file
basis by specifying a size when
marking a file for partial release. For more information, see
the release
(1) man page.
For n, specify an integer from 8 to whatever has been set
for the maxpartial
option.
For more information on maxpartial
, see the maxpartial
option in this list.
The default is 16.
maxpartial=
nSets the maximum partial release size for the file system to n
kilobytes. The partial release size cannot be set larger than
this maxpartial
setting.
For n, specify an integer such
that 0 ≤ n ≤ 2097152.
The default is 16.
partial_stage=
nSets the partial stage size for the file system to n
kilobytes.
For a partial release file, this value specifies the offset in the
file past which access results in the entire file being staged to disk.
For n, specify a integer from 0 to whatever has been set for
the maxpartial
option.
The default is equal to whatever has been set for
the partial
option.
stage_n_window=
nSets the stage -n
buffer size
for the file system to n kilobytes.
This option applies to files that are read directly from the archive media.
This attribute is set by using the
stage
(1) command's -n
option.
For a file with this attribute,
this is the size that is staged in to the
application's buffer at any one time.
For n, specify an integer such
that 64 ≤ n ≤ 2097152.
The default is 8192. If the total number of outstanding stage_n buffers is
less than physical memory, the access is not NFS, and the stage_n_window is
less than 1% physical memory, then the buffer
is allocated in pageable memory. Otherwise, blocks are allocated for
the buffer from the file system.
Note, the Oracle HSM shared file system does not support
stage -n
from a client.
stage_flush_behind=
nSets the maximum stage flush behind value to n kilobytes. Stage pages that are being staged are written to disk asynchronously to help the Solaris VM layer keep pages clean. For n, specify an integer such that 0 ≤ n ≤ 8192. The default is 0, which means that stage flush behind is disabled.
hwm_archive
Invokes the archiver when the amount of data in the file system increases above the high-water mark.
The following options are supported for StorageTek QFS and Oracle HSM shared file systems.
Both file system equipment types ms
and ma
are supported.
For a description of the ma
and ms
file systems,
see the mcf
(4) man page.
For a description of the StorageTek QFS shared file system, see
the StorageTek QFS Configuration and
Administration Guide.
The stripe width is set by default to round robin (using
the stripe=0
mount option).
shared
Specifies that the file system being mounted is a StorageTek QFS shared
file system. The shared
option must be specified in
the ∕etc∕vfstab
file because it is used in the boot initialization
sequence.
bg
Specifies that if the first mount attempt fails, the system should retry the mount in the background. If bg is not specified, the mount continues in the foreground.
retry=
nSpecifies the number of times to retry the mount operation. For n, specify an integer such that 0 ≤ n ≤ 20000. By default, n=10000.
minallocsz=
nSets the minimum block allocation value for the StorageTek QFS
shared file system to n.
Specify n in units of kilobytes and as a
multiple of 8 kilobytes.
The minallocsz
option specifies the minimum number of
bytes that are allocated
ahead of a write for a StorageTek QFS shared file system.
For n, specify an integer such that 16 ≤ n ≤ 2097152.
By default, n=8 * allocation_unit (DAU).
See sammkfs
(1m) command's -a
option.
maxallocsz=
nSets the maximum block allocation value for
the StorageTek QFS shared file system to n.
Specify n in units of kilobytes and as a multiple of 8
kilobytes.
The maxallocsz
option specifies the maximum number
of bytes that are allocated
ahead of a write for a StorageTek QFS shared file system.
For n, specify an integer such that 16 ≤ n ≤ 4194304.
By default, n=128 * allocation_unit (DAU).
See sammkfs
(1m) command's -a
option.
rdlease=
nSets the read lease time for the StorageTek QFS shared
file system to n seconds.
The rdlease
option specifies the maximum number of
seconds that a file can be read before reacquiring the read lease.
For n, specify an integer such that 15 ≤ n ≤ 600.
By default, n=30.
wrlease=
nSets the write lease time for the StorageTek QFS
shared file system to n seconds.
Only one host can write to a file at any one time unless
the mh_write
option is set on the metadata server.
If the mh_write
option is set on the metadata server,
multiple hosts can write to and read from the same file at the same time.
If multiple hosts are writing, the last write is the
one that is effective.
The wrlease
option specifies the maximum number of seconds
that a file can be written before reacquiring the write lease.
For n, specify an integer such that 15 ≤ n ≤ 600.
By default, n=30.
aplease=
nSets the append lease time for the StorageTek QFS
shared file system to n seconds.
Only one host can append to a file at any one time.
The aplease
option
specifies the maximum number of seconds that one host can append
to a file before reacquiring the append lease.
For n, specify an integer such that 15 ≤ n ≤ 600.
By default, n=30.
mh_write
Enables simultaneous reads and writes to the same file from multiple hosts. If mh_write is used, the StorageTek QFS shared file system switches all hosts into directio. The application must use page aligned memory buffers and well formed sector I∕O (512 bytes). Caution, if the application does not adhere to these alignment rules, data correctness is not guaranteed.
This option is effective only on the metadata server host.
If this option is specified when mounting the file system on
a client host, it is ignored. If the client host becomes the
metadata server in the future, however, this option becomes effective.
For this reason, it is recommended to use this mount option on the
metadata host and all potential metadata server hosts.
If the mh_write
option is not specified on the metadata server,
only one host can write at any one time to a single file.
min_pool=
nSets the minimum number of shared file system threads to keep around.
The number of threads grows and shrinks dynamically based on load.
This parameter tells the system to keep at least that many threads
in the active pool.
For n, specify an integer such that 8 ≤ n ≤ 2048.
The default n=64.
For Linux the default n=8.
NOTE: The min_pool
parameter must be set in samfs.cmd
file.
It is ignored if set in the ∕etc∕vfstab
file or on the mount
(1M)
command.
nstreams=
n* No longer used. *
meta_timeo=
nAllow attributes and directory data to be cached by a host system
for up to n seconds before checking for consistency with the
metadata server.
The default n=3.
Example 1. With the default setting of meta_timeo
=3, the
file system verifies attribute and directory consistency with the
metadata server at least every 3 seconds. For instance, a new file
created on one host may not be seen by an ls
(1) command on
another host for up to 3 seconds.
Example 2. If meta_timeo
=0, the file system verifies
attribute and directory consistency with the metadata server
before each use. The cattr
mount option can be used with
meta_timeo
=0 to ensure that changes made by other hosts currently
modifying a file are also immediately visible.
Example 3. If meta_timeo=3
, with the nocattr
mount option
(default), the file system verifies attribute consistency if it has not
been checked in the past 3 seconds; however, attribute changes made by a
client host which is currently modifying a file may not be detected until
the client lease time has expired.
Example 4. If meta_timeo=3
, with the cattr
mount option,
the file system verifies attribute consistency if it has not been checked
in the past 3 seconds, and also ensures that attribute changes made by
other hosts are detected within that time interval.
cattr
|\0 nocattr
Enable (disable) attribute consistency checking. If cattr
is set,
the file system ensures that attribute changes made by a host which is
modifying a file are visible to other hosts within the meta_timeo
interval. (Directories are not affected by cattr
; directory
modifications are always visible within the time interval set by
meta_timeo
.)
With the default setting of nocattr
, attribute changes made by a
host (in particular, file size and modification time) may not be visible
to other hosts until the write or append lease time has expired.
Note that enabling cattr
may adversely affect performance, as
additional network traffic is required.
lease_timeo=
nThe read, write, and∕or append lease for a single file is relinquished
if it is not being used after n seconds.
lease_timeo
varies from -1 to 15 seconds.
If lease_timeo
is =0, the lease is relinquished
if it is not being used after n seconds.
If lease_timeo
is set to -1, the lease is not relinquished
and the lease expires based on the lease time.
Note, the read and write lease is not relinquished
if mh_write
is set because multiple reader∕writer hosts are enabled.
The default n is 0.
The following options support the single-writer, multireader file system. This file system is mounted on one host system as a single-writer file system that updates the file system. In addition, this file system can be mounted on one or more host systems as a multireader file system.
These options can be specified only on StorageTek QFS
file systems. The writer
option
cannot be used if you are mounting the file system
as a StorageTek QFS shared file system, however, the reader
option
is supported. Note, sync_meta
should be set to 1 if the
reader
option is used in a StorageTek QFS shared file system.
A major difference between the multireader file system and StorageTek QFS shared file system is that the multireader host reads metadata from the disk, and the client hosts of a StorageTek QFS shared file system read metadata over the network.
The system administrator must ensure that only one host
in a multireader file system has the file system
mounted with the writer
mount option enabled.
writer
Sets the file system to type writer.
There can be only one host system that has the file system
mounted with the writer
option at any one time.
If writer
is specified, files are flushed to disk at
close and directories are always written through to disk.
The option atime = 1
is set for
writer
.
Prior to the 4.0 release, the writer
option was specified as
the shared_writer
option. The older syntax is supported
for backward compatibility.
reader
Sets the file system to type reader.
This mounts the file system as read only.
There is no limit to the number of host systems that can
have the same file system mounted with the reader
option.
By default, each lookup checks the inode and refreshes
the inode pages if the inode has been modified by the writer host.
If the invalid
option is set to a value greater than 0,
the inode is checked for
modification only after it has aged invalid
seconds after the last check; for more information, see
the invalid
option.
Prior to the 4.0 release, the reader
option was specified as
the shared_reader
option. The older syntax is supported
for backward compatibility.
invalid=
nWhen specified in conjunction with the reader
option,
holds cached attributes for the multireader file system
at least n
seconds after file modification. Caution, it is possible to
read stale data if invalid
is set to a nonzero value.
For n, specify an integer such that 0 ≤ n ≤ 60.
By default, n=0.
Example 1. If invalid=0
, which is the default,
the file system always checks
to see if the inode is stale. That is, it checks to
see if the inode has been
changed by the writer host.
Example 2. If invalid=30
, the
file system checks the inode 30 seconds after the last
check. This means that if you issue an ls
(1) command,
you might not see a new file for 30 seconds after it has been created
on the writer host. This also means that if you open an existing
file, for example with the cat
(1) command, you might not see
any changes made to the file on the writer host in the past 30 seconds.
refresh_at_eof
When specified in conjunction with the reader
option,
the current file size is refreshed when the read buffer exceeds the
end of file.
The following options are supported only for StorageTek QFS and Oracle HSM
file systems on ma Equipment Type file systems.
For more information on the ma
file system Equipment Type, see the
mcf
(4) man page.
mm_stripe=
nSets the metadata stripe width for the file system
to n 16-kilobyte disk allocation
units (DAUs).
By default, mm_stripe=1
,
which writes one DAU of metadata to one LUN before switching
to another LUN.
If mm_stripe=0
,
the metadata is round robined across all available metadata LUNs.
∕etc∕mnttab
Table of mounted file systems.
∕etc∕vfstab
List of default parameters for each file system.
∕etc∕opt∕SUNWsamfs∕samfs.cmd
List of default and global parameters for Oracle HSM
file systems. For more information, see the samfs.cmd
(4) man page.
release
(1),
setfa
(1),
ssum
(1).
mount
(1M),
mountall
(1M),
sam-fsalogd
(1m),
sam-releaser
(1m),
sammkfs
(1m),
umount_samfs
(1m).
mount
(2).
sam_setfa
(3x),
sam_advise
(3),
directio
(3C).
mcf
(4),
mnttab
(4),
samfs.cmd
(4),
vfstab
(4).
If the directory upon which a file system is to be mounted is a symbolic link, the file system is mounted on the directory to which the symbolic link refers, rather than on top of the symbolic link itself.
The mount parameters can be provided in the samfs.cmd
file,
in the ∕etc∕vfstab
file, and on the mount
(1M)
command. Specifications in the ∕etc∕vfstab
file
override the directives in the samfs.cmd
file,
and options to the mount
(1M)
command override specifications in the ∕etc∕vfstab
file.