mount, umount - Mounts and unmounts file systems
/usr/sbin/mount [-d] [-r | -u | -w] [-o argument,...]
[-t [no]type] file-system directory
/usr/sbin/mount [-el] [-t [no]type]
/usr/sbin/mount -a [-fv] [-t [no]type]
/usr/sbin/mount [-d] [-r | -u | -w] [-o argument,...]
[-t [no]type] file-system | directory
/usr/sbin/umount -a | -A -b [-fv] [-t type] [-h host]
/usr/sbin/umount [-fv] file-system... | directory...
There are options for the mount and umount commands.
Options for mount:
Attempts to mount all the file systems described in the
/etc/fstab file. In this case, file-system and directory
are taken from the /etc/fstab file. If -t type is specified,
all of the file systems in the /etc/fstab file with
that type will be mounted. Alternatively, if type is prefixed
with no, all the file systems in the /etc/fstab file
that do not have that type will be mounted. File systems
are not necessarily mounted in the order listed in the
/etc/fstab file. Mounts a UNIX file system (UFS) even if
it has not been unmounted cleanly or checked by fsck for
consistency. Also used to mount a CD-ROM UFS file system.
Caution
Do not employ the -d option to mount an AdvFS fileset.
When an AdvFS fileset is mounted with the -d
option, AdvFS initializes the domain transaction
log. As a result, no domain recovery will occur for
previously incomplete operations (which could cause
data corruption). If you cannot mount a fileset,
use the verify command. Lists all mount points.
Without this option, mount does not list mount
points served by either Automount or AutoFS. Performs
a fake mount and actually does not mount the
file system. This option is used to verify the
arguments you plan to use with the mount command.
Displays the value of all the file system options.
Specifies a list of comma-separated arguments.
Every argument specified is used. Some arguments
are valid for all file system types, while others
apply only to a specific type. See the Options for
mount -o Arguments section that is specific to your
file system type for a description of the arguments
supported by that file system. Queries the device
to determine which file system or file systems are
found on DVD or CD media, and thus, how that device
can be mounted. Mounts the specified file system
with read-only access. This option is the equivalent
of the following command: mount -o ro filesystem
directory
Physically write-protected and magnetic tape file
systems must be mounted with read-only access or
errors will occur when access times are updated,
whether or not any explicit write is attempted.
Note that -r and -w are paired; the default is -w.
Specifies the file system type. The supported file
systems are as follows:
advfs - Advanced File System (AdvFS)
ufs - UNIX File System (UFS)
nfs - Network File System (NFS)
mfs - Memory file system (RAM Disk) See mfs(8)
cdfs - ISO 9660 CD-ROM (Compact Disc Read Only Memory)
File System. See cdfs(4).
dvdfs - DVD-ROM (Digital Versatile Disk, Read-only)
File System. See dvdfs(4). This option mounts
DVD-ROM disks formatted in the Universal Disk Format
(UDF).
dfs - DCE Distributed File System
efs - DCE Episode File System
fdfs - File Descriptor File System (used by
streams)
ffm - File on File Mounting File System (used by
streams)
procfs - Process File System (used by debuggers)
pcfs - PC File System
sysv - System V File System
See fstab(4) for a description of the valid file
system types. If the no prefix is used, all file
types except the one specified are mounted.
Requests that the system remount a file system so
that it can update any incore data blocks for UFS
and AdvFS type file systems.
If you mount a UFS file system in a cluster for
read-only access, you cannot use the -u option to
update the file system to read-write access. A UFS
file system mounted for read-only access is accessible
by all cluster members; a UFS file system
mounted for read-write access is accessible only by
the cluster member on which the mount request is
issued.
For CDFS, this option is used to change the
attributes of a mount, such as the version
attribute. For example, CDFS is mounted noversion
by default. The following use of the -u option
shows how you can change the default: # mount -u -o
version /cdmntpnt Displays a message indicating
which file system is being mounted (verbose).
Mounts the specified file system with read/write
access. This option is equivalent to the -o rw
option. Read/write is the default access.
Options for umount:
Attempts to unmount all the file systems currently
mounted. Attempts to unmount all the file systems listed
in the /etc/fstab file. Broadcasts a message to all
server machines in the subnetwork to remove the client
host's name from their NFS mountdtab files. Performs a
fast unmount operation that causes remote file systems to
be unmounted without notifying the server. Unmounts all
file systems listed in the /etc/fstab file that are
remotely mounted from host. Unmounts all file systems
listed in the /etc/fstab file that are of the specified
type. The -a option must be used together with the -t
option. Displays a message indicating the file system is
being unmounted (verbose).
Options for mount -o Arguments [Toc] [Back]
There are many arguments for the -o option; they are discussed
in the following paragraphs.
TruCluster mount - o Arguments [Toc] [Back]
The following arguments are valid for TruCluster systems:
Enables you to specify the cluster member, identified by
cluster_member_name, that serves a given file system at
startup.
This mount command option determines where the file
system is first mounted; it does not limit or
determine the cluster members to which the file
system might later be relocated or fail over.
The -o server=cluster_member_name option is valid
only in a cluster, and only for the AdvFS, UFS,
MFS, NFS, CDFS, and DVDFS file systems. MFS file
systems support this option in a limited fashion:
the file system is mounted only if the specified
server is the local node.
For information on using this option, refer to the
Distributing File Systems Via the mount -o Command
section of the Cluster Administration manual.
Enables cluster file system partitioning. Use this
option only in a cluster. For example: # mount -o
server_only file-system
If a file system is already mounted, you cannot use
this option to update the mount status to
server_only. You must first unmount the file system
and then remount it with the server_only option.
When you mount a UFS file system in a cluster for
read/write access, or when you mount an MFS file
system in a cluster for read-only or read/write
access, the server_only option is used by default.
These file systems are treated as partitioned file
systems. That is, the file system is accessible for
both read-only and read/write access only by the
member that mounts it. Other cluster members cannot
read from, or write to, the MFS or UFS file system.
There is no remote access; there is no failover.
If you want to mount a UFS file system for readonly
access by all cluster members, you must
explicitly mount it read-only.
For information on using this option, refer to the
Partitioning File Systems section of the Cluster
Administration manual.
AdvFS mount - o Arguments [Toc] [Back]
The following arguments are valid for the Advanced File
System (AdvFS): Causes all files in the mounted fileset to
use atomic-write data logging for the duration of the
mount. Unlike chfile, which activates data logging on a
file in a manner that persists across mounts and unmounts,
the data logging provided by the -o adl mount option is
temporary and lasts only for the duration of the mount.
Additionally, files using temporary -o adl data logging
may be mmaped(), unlike files that have had persistent
data logging activated on them. The temporary data logging
is suspended until the last thread using the mmapped file
unmaps it. Finally, the use of chfile on a file that is
using temporary data logging causes the chfile command to
override the temporary data logging provided by the new
mount and the file's I/O mode is changed persistently
according to the arguments given to the chfile command.
For more information see chfile(8). Enables an AdvFS
fileset to be mounted as a domain volume even though it
has the same AdvFS domain ID as a fileset that is already
mounted.
AdvFS and UFS mount - o Arguments [Toc] [Back]
The following arguments are valid for the AdvFS and UFS
file systems: Flushes to disk file access time changes for
reads of regular files. (Default behavior when neither
atimes or noatimes is specified.) Marks file access time
changes made for reads of regular files in memory, but
does not flush them to disk until other file modifications
occur. This behavior does not comply with industry standards
and is used to reduce disk writes for applications
with no dependencies on file access times. Allows the
file system to make use of expanded space on the underlying
storage device. The -o option can be used on the initial
mount of a file system or by way of the -u update
option if the file system is already mounted. For UFS, if
you do not want to mount the file system, or if you want
to take only part of the available storage space, you can
use the extendfs command. See the extendfs(8) reference
page for more information.
There are several steps needed in order to expand a
file system: Make additional storage space available
on the underlying storage device (that is, a
LSM volume or hardware RAID LUN). For non-LSM volumes,
modify the disk label to include additional
storage. See the disklabel reference page for more
information. Use the mount command with the extend
option to allow the file system to use the additional
storage.
For an unmounted AdvFS fileset, the following is an
example of volume expansion:
# mount -o extend domain#fileset /ausr1
For a mounted AdvFS fileset, the following is an
example of volume expansion:
# mount -u -o extend domain#fileset /ausr2
For an an unmounted UFS file system, the following
is an example of volume expansion:
# mount -o extend /dev/disk/dsk0g /useref
For a mounted UFS file system, the following is an
example of volume expansion:
# mount -u -o extend /dev/disk/dsk0h /useracct
Refer to the System Administration manual for more
information. Allows read/write access. Allows
read-only access. Allows read/write access.
Allows file system to be used as swap space.
Allows a file system to be mounted even if it was
not cleanly unmounted. Allows access to block and
character-special devices. Disallows access from
the file system to either block or character-special
devices. Allows set-user-ID execution. Prohibits
set-user-ID execution. Causes all writes to
be written immediately to disk as well as to the
buffer cache. Specifies that writes may return
before data is written to disk. Enables the alternate
smooth sync policy, in which modified pages
are not written to disk until they have been dirty
and idle for the smoothsync_age time period.
By default, modified pages are flushed after they
have been dirty for the smoothsync_age time period,
regardless of continued modifications to the page.
Pages that have been mapped into virtual memory
will always use this default policy, regardless of
the smsync2 setting. The default smoothsync_age
period is 30 seconds, and can be modified by editing
the inittab file.
If you enable the smsync2 option on a mount point
in an AdvFS domain, the alternate smooth sync policy
goes into effect for all of the filesets in the
domain. Allows binary execution. Prohibits binary
execution. Enables new files to inherit the parent
directory's group ID. This is the default and
matches BSD semantics. Applies SVID 3 semantics.
For example, if the parent directory's mode bits
include IS_GID, then the new file will inherit the
parent's group ID. If IS_GID is off, then it
inherits the process group ID.
UFS mount - o Arguments [Toc] [Back]
The following mount -o arguments are valid only for the
UFS file system. Delays synchronously flushing metadata
updates to disk. Instead, metadata (such as inode,
directory, and indirect blocks) is flushed by the sync
daemon. This mount option improves performance in the
following ways: Multiple updates to a block are accomplished
with a single write instead of with multiple
writes of the same block, which can occur during synchronous
metadata updates. System responsiveness improves
when running metadata intensive applications. Metadata
writes to disk do not occur immediately.
CAUTION
Data might be lost if you use this option and your
system crashes before the sync daemon flushes the
metadata to disk. Do not use this option for the
root (/) or /usr file systems.
You can use this option for a temporary file system,
such as /tmp, in which applications cache
temporary data that is expendable. Refer to the
nodelayed option for information on disabling
delayed metadata updates. Synchronously flushes
metadata updates to disk. This is the default
behavior.
By default, to maintain file system consistency,
UFS metadata (such as inode, directory, and indirect
blocks) is updated synchronously, which
ensures that the UFS file system is consistent at
all times and no data is lost if your system
crashes. However, it can affect file system performance.
Refer to the delayed option for information
on disabling synchronous metadata updates to
improve performance. Prevents excessive asynchronous
I/O from overloading the device queue,
which can affect response time for processes waiting
for I/O operations to complete. To use this
argument, you must enable smooth sync.
See the EXAMPLES section for usage examples.
NFS mount - o Arguments [Toc] [Back]
The following arguments are valid for NFS file system:
Allows access to block and character-special devices.
Disallows access from the file system to either block or
character-special devices. Allows read/write access.
Allows read-only access. Allows set-user-ID execution.
Prohibits set-user-ID execution. Causes all writes to be
written immediately to disk as well as to the buffer
cache. Specifies that writes may return before data is
written to disk. Allows binary execution. Prohibits
binary execution. New files inherit the parent directory's
group ID. This is the default and matches BSD's
semantics. SVID 3 semantics applied. For example, if the
parent directory's mode bits include IS_GID, then the new
file will inherit the parent's group ID. If IS_GID is
off, then it inherits the process group ID. Retries in
the background, if the first mount attempt fails. Retries
in the foreground. Sets the number of mount failure
retries to n. Sets the read buffer size to n bytes. Sets
the write buffer size to n bytes. Sets the initial NFS
timeout period for UDP mounts to n tenths of a second.
NFS continually adjusts the timing as a function of network
response time. Sets the maximum value, in seconds,
that is allowed between request transmissions. (UDP mounts
only) Sets the number of NFS retransmissions to n. Allows
hard mounted file system operations to be interrupted.
Prevents hard mounted file system operations from being
interrupted, unless the thread is terminated (for example
by a SIGKILL or an AST). Returns an error if the server
does not respond. Retries the request until the server
responds. Usually, the mount command tries to use Version
3 of the NFS protocol. If the server does not support
Version 3, then the mount command retries the mount using
Version 2. Specifying -o nfsv2 forces the mount command
to use NFS Version 2. NFS Version 3 is an enhanced version
of the NFS protocol that provides 64-bit file access, as
well as features designed to improve performance and correctness.
Alternatively, you can use the vers=2 argument.
Tries to use Version 3 of the NFS protocol. If the
server does not support it, Version 2 is used.
This is the default.
Alternatively, you can use the vers=3 argument.
Specifies the network transport: udp or tcp.
Specify udp to use UDP as the network transport.
This is supported by all known NFS servers. UDP
works best in local, fast, and reliable environments.
The mount will fail if the server does not
support NFS over UDP. The proto=udp syntax is the
default.
Specify tcp to use TCP as the network transport.
This is supported by some vendors, but not all.
TCP works better than UDP in high-loss, congested
networks, and is the only way to use NFS over the
Internet. The mount will fail if the server does
not support NFS over TCP.
The -o tcp syntax is compatible with 4.4 BSD syntax,
while the proto=tcp syntax is compatible with
Solaris 2.4 syntax. Sets the server IP port number
to the value of n. The default is to query the
portmap daemon on the server for the port number
(which is almost always 2049). This argument is
useful only when the server is not running the
portmap daemon or is running multiple NFS servers.
Both of these situations are very rare. Allows the
use of extended attributes (property list) including
access control lists (ACLs) on this file system.
The NFS server exporting this file system must
be running the proplistd daemon. See proplist(4),
acl(4), and proplistd(8) . Specifies the version
of the NFS protocol. You can specify either Version
3 or Version 2.
Usually, the mount command tries to use Version 3
of the NFS protocol. If the server does not support
Version 3, then the mount command retries the
mount using Version 2. Specifying vers=2 forces
the mount command to use NFS Version 2. NFS Version
3 is an enhanced version of the NFS protocol that
provides 64-bit file access, as well as features
designed to improve performance and correctness.
Alternatively, you can use the nfsv2 or nfsv3 argument.
For NFS, the defaults are fg, retry=10000, timeo=11, maxtimo=20,
retrans=4, hard, and intr. Defaults for rsize
and wsize are set by the kernel.
The bg argument causes mount to run in the background if
the server's mountd does not respond. The mount command
attempts each request retry times before giving up. Once
the file system is mounted, each NFS request made in the
kernel waits timeo tenths of a second for a response. If
no response arrives, the timeout period is multiplied by 2
and the request is retransmitted.
When retrans retransmissions have been sent with no reply,
a soft mounted file system returns an error on the request
and a hard mounted file system retries the request at maxtimo
intervals. File systems that are mounted rw
(read/write) should use the hard argument. The number of
bytes in a read or write request can be set with the rsize
and wsize arguments.
Note
Using the mount command with the -t nfs option may cause
it to touch the /etc/exports file. If the/etc/exports file
has been manually created, you should ensure that it has
bin:bin owner:group ownership.
NFS Update Visibility mount - o Arguments [Toc] [Back]
These arguments control how quickly you see updates to a
file or directory that has been modified by another host.
Increasing these values gives you slightly better performance.
Decreasing the values decreases the time it takes
for you to see modifications made on the other host. If
you are the only person modifying files under this mount
point, you should increase these values. Holds cached
directory attributes for at least n seconds. Holds cached
directory attributes for no more than n seconds. The maximum
value you can specify is 3600. Holds cached file
attributes for at least n seconds. Holds cached file
attributes for no more than n seconds. The maximum value
you can specify is 3600. Sets all four attributes' cache
timeout values to n. Sets no attribute caching. This
argument is equivalent to actimeo=0. Sets no fresh
attribute when opening a file.
The NFS update visibility argument defaults are
acdirmin=30, acdirmax=60, acregmin=3, and acregmax=60.
CDFS mount - o Arguments [Toc] [Back]
The following arguments are valid for the CD-ROM File System
(CDFS): Ignores the permission bits, if present, and
defaults all file and directory permissions to the value
0555, with a zero User ID (UID, owned by root). Files and
directories recorded on an ISO 9660-formatted file system
might or might not have permission bits. This setting is
a default argument because the permissions on most existing
ISO 9660-formatted CD-ROMs do not map to the UID
scheme that is used. Uses the on-disk permission bits, if
present. If a file or directory is not recorded with permission
bits, the default 0555 is used. Strips off the
extension (;#) from the version string if a file recorded
on an ISO 9660-formatted file system or a file system formatted
by the High Sierra Group contains a version string.
File and directory names are displayed in lowercase letters
and case-insensitive name matching is performed. Use
this argument if you are mounting a CD-ROM containing
MS-DOS applications. Uses the Rock Ridge Interchange Protocol
(RRIP) extensions to ISO 9660 (if present on the
file system) to provide mixed-case file names, device special
files, and other attributes for files on the file
system. This setting is a default argument. If there are
no RRIP extensions on the file system, the file system
will be mounted and the argument will be ignored. Turns
off the Rock Ridge Interchange Protocol (RRIP) extensions
to ISO 9660 for files on the file system. If there are
RRIP extensions on the file system, the file system will
be mounted and the extensions will be ignored. Uses the
Microsoft Joliet formatted CD-ROM media, which provides
long, mixed-case file names. Turns off Microsoft Joliet
formatted CD-ROM media. Uses the ISO 9660 uppercase 8.3
formatted file system. This is the default if no other
file formats are found. Uses verbose messages in the output.
The defaults for CDFS are ro, nodev, defperm, and rrip.
CD-ROMs can contain several formats to support different
platforms and operating systems. If you know which format
you require (RRIP, Joliet, or ISO9660) specify the appropriate
qualifier to the -o option.
If you do not specify options for file name formats on the
command line, the mount command automatically tests for
the presence of formats and mounts it by default, according
to the following rules of precedence: Check if Rock
Ridge Interchange Protocol (RRIP) extensions are found on
the CD-ROM, if yes, mount as -t cdfs -o rrip. If RRIP
extensions are not found, check if the media has Microsoft
Joliet formatted file names. If yes, mount as -t cdfs -o
joliet. If neither of the above were found, the mount
command defaults to ISO 9660 format.
If you specify one or more exclusive qualifiers, such as
-o norrip, the mount command does not test for the presence
of that format, and defaults to the next highest
precedent.
If a specifically-requested format is not found and other
formats are not excluded, the mount command will attempt
to mount the next highest precedent. For example, you
attempt to mount a CD-ROM specifying -o joliet format but
the CD-ROM does not contain that format. Unless you
specifically requested -o norrip, the mount command will
attempt to mount RRIP. If RRIP is not found, the mount
command defaults to ISO 9660 format.
FFM mount - o Arguments [Toc] [Back]
The following arguments are valid for the File-on-FileMounting
(FFM) file system: Allows two separate files to
have identical contents, separate names, and separate file
descriptors. (Do not confuse this clone with an AdvFS
clone fileset.)
Specifies one or more file systems. How you specify a
file system depends on whether it is UFS or NFS or AdvFS.
To specify a UFS file system, enter the name of its
block device special file. For example:
/dev/disk/dsk3c. The mount command returns an error
if you try to mount the file system on a partition
that is already in use.
To specify an NFS file system, specify the host and
path name in either of these formats: host:path or
path@host.
To specify an AdvFS fileset, enter the name of the
file domain, a pound-sign(#) character, and the
name of the fileset. For example:
root_domain#root. Specifies one or more directories.
The directory must exist before you use the
mount command. When the command is successful, the
directory becomes the name of the newly mounted
root directory, its mount point.
When specified with the umount command, the directory
must not be in use. Use the pwd command to
check your present working directory. If you or
another user is in the mounted directory or in any
directory in its hierarchy, you must switch to a
different directory. Likewise, if you are using
files in the mounted directory, you must close the
files to successfully unmount the directory.
Use the mount command to make a file system available for
use, or mounted. Use the umount command to make a file
system unavailable for use, or unmounted.
Use the mount command with the -q option to determine
which file system or file systems are found on an
unmounted CD or DVD device. You can either read the command
output or interpret the exit code to determine which
file system is found, and thus, how it should be mounted.
The format used in the mount command determines the format
returned by the getfsstat and getmntinfo commands.
If the mount command is invoked with only a file-system or
directory operand specified, the command searches the
/etc/fstab file for an entry whose file-system or directory
field matches the argument specified with the command.
For example, if the line /dev/disk/dsk0g /usr ufs rw 1 1
is specified in the /etc/fstab file, both of these two
commands, mount /usr and mount /dev/disk/dsk0g are equivalent
to the following command: # mount /dev/disk/dsk0g
/usr
The umount command announces to the system that the file
system file-system previously mounted on directory should
be removed. Either the file system name or the directory
mount point can be specified in the command line.
To use the mount and unmount commands, you must be the
root user, with the following exceptions: If NFS file systems
have been explicitly exported to allow nonroot users
to mount the file system. Refer to the -n option of
mountd(8) for more information. If a CD-ROM is mounted
(by specifying the -t cdfs option) and the user owns the
mount point.
The mount command also lets you mount an ISO 9660- or HSGformatted
file system onto a directory.
No more than one user at a time should mount a disk partition
with read/write access or the file system might
become corrupted.
If the directory on 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.
When you boot to single-user mode, the root file system is
mounted with read-only access. If you want to modify a
file, you must change the options on the root file system
to read/write. You can do this with the following command:
# mount -u /
If your /etc/fstab file is corrupted, you can mount the
root file system with the following command: # mount -u
/dev/disk/dsk??/
You must be the root user to mount a UFS file system. By
default, the maximum number of UFS mounts is 1,000. However,
you can modify this value by using the sysconfig
command. For example: # sysconfig -r vfs max-ufsmounts=1100
The default for CDFS is to not allow access to device special
files (argument nodev) because the device numbers
recorded on a disc using RRIP extensions might not match
the device numbers used by the operating system. If you
want to allow device access, mount the file system with
the dev argument and use the cddevsuppl command to map the
device numbers of the device special files on the disc to
new device numbers used by the operating system.
The mount command attempts to dynamically load the cdfs
kernel modules if they are not statically built into the
running kernel. However, you must be the root user to
dynamically load the cdfs kernel modules. Other users
receive the following error should they attempt the operation:
mount: super user privileges required to load cdfs
module
All other errors that could occur as the cdfs kernel modules
are being dynamically loaded produce the following
error message: mount: Can't load cdfs module
Refer to cdfs(4) for information on the correct system
configuration options to set before using CDFS.
NFS mounts can fail due to authentication requirements on
the server. For example, a "Client credential too weak"
message is returned if a user attempts to mount and the
server only allows root user mounting. A "Server rejected
credential" message is returned if the server is not able
to resolve the client's IP address.
If your workstation has multiple network interfaces, the
server must be able to resolve all IP addresses from which
it might receive mount requests. See mountd(8) or the
Network Administration: Services manual for more information.
When you mount the first fileset in an AdvFS domain, AdvFS
determines whether or not it can access all data in all
volumes of that domain. If AdvFS determines that the size
of any volume in the domain is actually smaller than the
size recorded for that volume in the domain's metadata,
there are two possible outcomes: The mount succeeds, but
in read-only mode. In this case, AdvFS is able to read
the last currently in-use block on the volume. A message
similar to the following is displayed: Actual size of virtual
disk /dev/vol/vol01 is 100352 blocks but recorded
size is 102400 blocks. Mounting fileset staff#grads in
read-only mode. The mount fails. In this case, AdvFS
cannot read the last currently in-use block on the volume.
A message similar to the following is displayed: Actual
size of virtual disk /dev/vol/vol01 is 100352 blocks but
recorded size is 102400 blocks. Cannot read essential
data on /dev/vol/vol01. Corrupted volume found; failing
mount of staff#grads. staff#grads on /grads: I/O error
When you attempt to mount an AdvFS fileset in an AdvFS
domain, the number of volumes pointed to by the
/etc/fdmns/dmn_name links must equal the number of volumes
in the domain. If you attempt to mount an AdvFS file system
with an incorrect number of volumes, the following
message will appear on the console: # Volume count mismatch
for domain dmn_name. dmn_name expects 2 volumes,
/etc/fdmns/dmn_name has 1 links.
To correct the problem, you must match the number of volumes
and then mount them. See advscan(8) for more information.
Smoothsync [Toc] [Back]
Smoothsync increases efficiency in the part of the file
system that uses the disks for writing dirty pages. Prior
to smoothsync, dirty pages were scheduled for writing
every 30 seconds by the update daemon. The smoothsync
model schedules each page for writing after that page has
been dirty for the smoothsync_age period (default 30 seconds).
This allows all buffers to age the full smoothsync_age
period, versus an average of 15 seconds with the
update daemon model. This approach also distributes the
requests made of the disk subsystem evenly across the
smoothsync_age period. The update daemon model submits
all the I/O requests together.
The smoothsync_age period can be set using sysconfig. A
value of 0 disables smoothsync.
An alternate smoothsync policy can be enabled on a file
system basis by mounting with the smsync2 flag. With this
policy, a page is not scheduled for writing until it is
dirty and unmodified for the last smoothsync_age period.
For example, suppose you have an application that keeps
updating the same page repeatedly. With smsync2 enabled,
until the page has been idle (unchanged) in memory for the
entire smoothsync_age period, it will not be written to
disk. Therefore, if the smoothsync_age is 30 seconds, and
your application updates the page in memory every 10 seconds,
the page might not be written to disk for a very
long time.
While this policy might further decrease I/O load, it is
appropriate only for file systems or applications in which
additional data loss is acceptable if the system crashes.
The mount and umount commands support mount point argument
pathnames of up to MNAMELEN, which includes the null terminating
character. MNAMELEN can be up to 90 characters
long, including the null terminating character.
Before you can use the FFM file system, you must configure
the kernel option FFM_FS into the kernel.
Standard Usage
Success. An error occurred.
Query Option [Toc] [Back]
Use the following chart to determine which file systems
are on the device from the exit code when you invoke the
mount command with the -q option.
------------------------------
UFS CDFS DVDFS
------------------------------
0
1 X
2 X
3 X X
4 X
5 X X
6 X X
7 X X X
------------------------------
The following sections describe some warnings and errors
produced by the command.
Overlapping Partitions Warnings [Toc] [Back]
The following warning messages about overlapping partitions
are displayed only if you use the -v option. Warning:
partition special-device and overlapping partition(s)
are marked in use in the disklabel.
Explanation:
The specified partition overlaps with another partition
or partitions that have the fstype field
set. Warning: partition(s) which overlap specialdevice
are marked in use in the disklabel.
Explanation:
The partition overlaps another partition or partitions
that have the fstype field set. Warning: the
disklabel for special-device does not exist or is
corrupted.
Explanation:
The device specified either does not have a disk
label or the disk label has been corrupted. Warning:
unable to check special-device against active
AdvFS domains because the directory /etc/fdmns
seems to be missing or wrong.
Explanation:
There was a failure when checking the overlap with
AdvFS domains. The failure is with /etc/fdmns or
/etc/fdmns/dom, or an active domain does not exist.
Warning: unable to check special-device against
active swap devices because special swap files are
missing.
Explanation:
A failure occurred when checking the overlap with
active swap devices. The special device files
associated with active swap devices are invalid.
Warning: unknown overlap condition errno encountered
for partitionspecial-device.
Explanation:
An unknown overlap condition was encountered for
the specified device. Error: partition specialdevice
is marked 'unused'
Explanation:
The fstype in the disk label temporarily is set and
will revert with the following messages when you
unmount the file using umount: Warning: partition
/dev/disk/dsk5c was detected as marked unused.
Warning: partition /dev/disk/dsk5c temporarily set
to /
'FS_BSDFFS' 4.2BSD Fast File System. Warning:
Please use disklabel to correct this condition.
Overlapping Partitions Errors [Toc] [Back]
The following are fatal error messages associated with
overlapping partitions. Error: File system type fstype is
invalid or not installed.
Explanation:
The file system type specified is not resident in
the kernel or is otherwise inaccessible. Error: an
overlapping partition is open.
Explanation:
A partition that overlaps the specified partition
is open. Error: special-device is an invalid
device or cannot be opened.
Explanation:
The specified device is invalid and an overlapping
partition is open. Error: special-device contains
a fstype file system.
Explanation:
The specified partition and overlapping partitions
have the fstype field set. Error: Unknown severe
error errno encountered for partitionspecialdevice.
Explanation:
An unknown overlap condition was encountered for
the specified device.
To mount a local disk, enter: % mount /dev/disk/dsk0g /usr
To mount an AdvFS fileset, enter either of the following
commands: % mount -t advfs usr_dmn#user1 /usr/user1
% mount usr_dmn#user1 /usr/user1 To mount all ufs
file systems, enter: % mount -at ufs To mount a
remote file system, enter either of the following
commands: % mount -t nfs serv:/usr/src /usr/src
% mount -t nfs /usr/src@serv /usr/src To mount a
remote file system with a hard mount, enter: %
mount -o hard serv:/usr/src /usr/src
To mount an ISO 9660-formatted or HSG-formatted
file system from block device /dev/disk/cdrom3c
onto the local directory /cdfs with the file version
strings stripped off, enter either of the following
commands: % mount -t cdfs -o noversion
/dev/disk/cdrom3c /cdfs
% mount -o noversion /dev/disk/cdrom3c /cdfs To
mount a UFS CD-ROM (for example, the installation
CD-ROM) from block device /dev/disk/cdrom3c onto
the local directory cdrom, enter either of the following
commands: % mount -r /dev/disk/cdrom3c
/cdrom
% mount -o ro /dev/disk/cdrom3c /cdrom To mount the
joliet-formatted file system on a multi-formatted
file system from block device /dev/disk/cdrom3c
onto the local directory /cdfs enter the following:
% mount -t cdfs -o joliet /dev/disk/cdrom3c /cdfs
To unmount the file system mounted on the /mnt
local directory, enter the following command: %
umount /mnt To unmount all NFS file systems, enter
the following command: % umount -A -t nfs To
unmount all file systems exported from host2, enter
the following command: % umount -h host2 To use the
delayed metadata option, use commands similar to
the following examples: To enable delayed metadata
updates and improve performance (at the risk of
data loss), use a command similar to the following:
# mount -o delayed /dev/disk/dsk3c /tmp_files
To enable delayed metadata update on a file system
that is already mounted, use a command similar to
the following: # mount -u -o delayed /tmp_files
Any options that were in force are turned off by
this command. Therefore, you must also reenter all
required mount options when you use the -o delayed
option on a mounted file system To disable the
delayed metadata update option, use a command similar
to the following: # mount -u -o nodelayed
/tmp_files
Any options that were in force are turned off by
this command. Therefore, you must also reenter all
required mount options when you use the -o nodelayed
option on a mounted file system. To view
which mount option is in operation for a given file
system, use the mount command without arguments, as
follows: # mount /dev/disk/dsk3c on /tmp_files type
ufs (rw, delayed)
Note that the word delayed appears in the mount
options list at the end of the output from the
mount command. To determine how the device
/dev/rdisk/cdrom0c can be mounted, enter the following
command: # mount -q /dev/rdisk/cdrom0c
/dev/rdisk/cdrom0c can be mounted as: CDFS DVDFS
ENVIRONMENT VARIABLES [Toc] [Back]FILES
Specifies the command path. Specifies the command path.
Contains static information about file systems.
Commands: cddevsuppl(8), extendfs(8), mfs(8), mountd(8),
nfsd(8), proplistd(8)
Functions: mount(2), mount(2sv), umount(2), umount(2sv),
umount(3)
Files: advfs(4), cdfs(4), fstab(4), mountdtab(4)
System Administration, Network Administration: Services,
Cluster Administration
mount(8)
[ Back ] |
