Makes a physically remote file system accessible to local users.
Applies to: NFS client
Mounts a remote directory to local device DNFSn:. Similar in
function to the UNIX /etc/mount command, MOUNT gives a file
system a UNIX path name. (In format and style, MOUNT resembles
the DCL command MOUNT.)
You can mount either OpenVMS or UNIX file systems.
Related commands: DISMOUNT, SHOW MOUNT
Format
MOUNT mount_point [ volume_label ] [ logical_name ]
[ /HOST=host ]
[ /PATH="/path/name" ]
[ /ACP_PARAMS=options ]
[ /[NO]ADF[=option] ]
[ /AUTOMOUNT[=INACTIVITY:time] ]
[ /BACKGROUND[=options] ]
[ /CACHE_TIMEOUT[=options] ]
[ /[NO]CONVERT ]
[ /DATA=[options] ]
[ /FILEIDS[=options] ]
[ /[NO]FORCE ]
[ /GID=gid ]
[ /GROUP ]
[ /OWNER_UIC=n ]
[ /PROCESSOR=acp_option ]
[ /PROTECTION=protections ]
[ /RETRIES=n ]
[ /SERVER_TYPE=type ]
[ /SHARE ]
[ /STRUCTURE ]
[ /[NO]SUPERUSER=uid ]
[ /SYSTEM ]
[ /TIMEOUT=OpenVMS_delta_time ]
[ /TRANSPORT=protocol ]
[ /UID=uid ]
[ /USER=user ]
[ /[NO]WRITE ]
1 – Restrictions and Tips
If you mount remote OpenVMS directories where the NFS server is
running TCP/IP Services software, use the /NOADF qualifier on the
MOUNT command line unless you are using the OpenVMS-to-OpenVMS
integration feature.
The /NOADF requirement applies only if the remote NFS server is
running versions of TCP/IP Services earlier than Version 3.3 and
cannot participate in OpenVMS-to-OpenVMS mode operation. Other
tips include:
o For the qualifiers that require a time value, specify OpenVMS
delta time.
o Whenever you specify multiple options and values, use the
following syntax:
/qualifier=(option_a:value1,option_b:value2,value3)
2 – Parameters
mount_point
Required.
Local device (and optional directory tree) on which to mount the
remote NFS file system. Specify this mount point as one of the
following:
DNFSn:
DNFSn:[dir.subdir]
DNFSn:[dir.subdir]file
where:
n Specifies the unit number. Specify a value from 0 to
9999. Specifying 0 causes the client to choose the next
available unit number. (It does not mount a device named
DNFS0:.)
[dir] Specifies the directory to mount
or (up to eight subdirectories in addition to the [000000]
[dir.subdir]ectory).
file Specifies the individual file to mount.
volume_label
Optional. Default: First 12 characters of the combined values
of the /HOST and /PATH qualifiers. The default label is a
combination of /HOST and /PATH with a dollar sign ($) separating
the two.
Specifies the Files-11 (ODS-2 or ODS-5) volume label to be
associated with the remote path name.
You can use this parameter to provide a unique volume label on
a system where there is a label conflict. The client does the
following:
o Accepts only the first 12 characters for all other entries.
o Applies volume_label only on the first mount of a particular
disk.
o Ignores volume_label with subsequent mounts on that disk.
HP recommends that if you use the SET FILE /STATISTICS command
on a file mounted with DNFS, do not include any colons (:) in the
volume_label.
logical_name
Optional. Default: None.
Specifies the logical name associated with the volume.
The client creates the following logical definitions, depending
on what you specify:
o If you mount DNFSn:[000000], the client defines the logical
name as DNFSn:
o If you mount DNFSn:[dir.dir], the client defines the logical
name as DNFSn:[dir.dir.] The extra dot allows for relative
directory specifications. If you issue the following command:
$ SET DEFAULT logical:[subdir]
The full default definition becomes:
DNFSn:[dir.dir.subdir]
The client places the logical name in the SYSTEM logical name
table, unless you specify the /GROUP or /SHARE qualifier. The
client deletes the logical name from the SYSTEM table when you
dismount the volume. The process must have SYSNAM privilege to
mount a system mount point. Without SYSNAM or GRPNAM privilege,
the user must specify /SHARE for a JOB mount. (See the /SHARE
qualifier for more information.)
3 – Qualifiers
3.1 /ACP_PARAMS
/ACP_PARAMS=
{ BUFFER_LIMIT=n | DUMP | IO_DIRECT=n | IO_BUFFERED=n | MAX_
WORKSET=pages | PAGE_FILE=file | PRIORITY=base-priority |
WORKSET=pages }
Optional.
Specifies modifiable process parameters for the ancillary control
process (ACP).
These parameters are dynamic. The NFS client applies your
settings at each first start of an ACP.
For descriptions of these options, see the section on RUN
(PROCESS) in the OpenVMS DCL Dictionary.
3.2 /ADF
/ADF=CREATE
/NOADF
Optional. Default: /ADF=CREATE.
If attributes data files (ADFs) exist on the NFS server, the /ADF
qualifier lets you use them.
The server uses ADFs to store OpenVMS file attributes. These
files appear on the server as .$ADF$file files, but you cannot
view them directly on the local client system.
The option is:
o CREATE
The client uses and updates the ADFs, and creates ADFs for new
files.
/NOADF - No ADFs are created or used.
3.3 /AUTOMOUNT
/AUTOMOUNT[=INACTIVITY:time]
Optional. Defaults:
o If you omit this qualifier, automounting is not enabled for
this file system.
o If you include the /AUTOMOUNT qualifier but omit the
INACTIVITY keyword, file systems are automatically dismounted
after five minutes of inactivity.
This qualifier enables automounting for the file system. The file
system is automatically mounted when you access its path name.
You can include the optional INACTIVITY keyword to specify the
number of minutes of inactivity before automatically dismounting
the file system. Be sure to specify the time as hh:mm:ss. When
this inactive period expires, the NFS client dismounts the path
name.
3.4 /BACKGROUND
/BACKGROUND [= {DELAY:OpenVMS_delta_time | RETRY:n}]
Optional. Defaults:
o If you omit this qualifier, background mode mounting is not
attempted.
o If you omit the DELAY keyword, background mode mounting is set
up with /BACKGROUND=(DELAY:00:00:30,RETRY:10).
This qualifier enables background mode for mounting the file
system.
The optional DELAY time specifies amount of time to wait if
the mount attempt fails before trying again. Specify the
time as hh:mm:ss. The maximum delay period you can specify is
approximately 49 days. The default delay time is 30 seconds.
The optional RETRY keyword specifies the number of times to
repeat the attempt to mount the file system. RETRY:0 means that
the client uses the first try only. The default number of times
to retry is 10.
If you use the /BACKGROUND qualifier, you must also use the
/RETRIES qualifier and specify a nonzero value. For example:
$ TCPIP MOUNT DNFS0: /BACKGROUND=RETRY:9 /RETRIES=4 /HOST="robin" -
_$ /PATH="/USR/USERS/GEORGE"
In this example, you are asking for four data retries on each
mount attempt and nine mount attempts, for a total of 36 tries.
If you use the default value for /RETRIES, the first mount
attempt can never complete except by succeeding, and the process
doing the mount will hang until the server becomes available.
3.5 /CACHE_TIMEOUT
/CACHE_TIMEOUT=
[ DIRECTORY:OpenVMS_delta_time ]
[ ATTRIBUTE:OpenVMS_delta_time ]
[ READ_DIRECTORY ]
Optional. Defaults:
If you omit this qualifiier, the file system is set up with
caching timeouts as follows:
/CACHE_TIMEOUT=(DIRECTORY:00:00:30,ATTRIBUTE:00:00:15)
Specifies the following caching timeout information for the mount
point:
o DIRECTORY:OpenVMS_delta_time [00:00:30]
- Amount of time that the client waits between rereading a
directory's status or contents.
- Specify OpenVMS_delta_time as hh:mm:ss.
o ATTRIBUTE:OpenVMS_delta_time [00:00:15]
- Amount of time that the client waits between rereading a
file's attributes from the NFS server.
- Specify OpenVMS_delta_time as hh:mm:ss.
o READ_DIRECTORY
- Forces the client to read the contents of the directory
requested when the cache timeout occurs rather than rely on
the directory's modified time.
- By reading the directory contents, the client can be aware
of any changes to the number of files within the directory,
even if the directory's modify time was not updated.
3.6 /CONVERT
/CONVERT
/NOCONVERT
Optional. Default: /CONVERT.
Converts files with the following attributes to STREAM_LF files:
o Sequential
o Variable length
o Carriage return/carriage control (VAR-CR)
The convert feature works with some utilities and DCL commands
but not with others. For example, it works with the CREATE
command and with EDIT/TPU, but it does not work with COPY,
BACKUP, or EDIT/EDT. There is no simple way to identify what
works. However, for the feature to take effect, the following
conditions must be satisfied:
o The file attributes must be sequential, variable length, and
carriage return/carriage control.
o The file must be opened for exclusive write access (generally
true for newly created files).
o The file must be opened with the FAB$M_SQO bit set in the
FAB$L_FOP field.
o The file creation and the open for write must be done in one
step. That is, if the program first creates the file and
afterward opens it for write, the convert feature does not
work.
You can convert only those files that were opened using RMS
sequential access. For additional information, refer to the HP
TCP/IP Services for OpenVMS Management manual.
3.7 /DATA
/DATA [= {read_bytes | write_bytes}]
Optional. Default: /DATA=(8192,8192).
Largest amount of NFS data received or transmitted in one network
operation. The options mean:
o read_bytes - Data received. Minimum value = 512.
o write_bytes - Data transmitted. Minimum value = 512.
If you specify only one value, it applies to both READ and WRITE.
You do not need to use /DATA unless a remote NFS server imposes
a restriction on data size. If the server requests a smaller
transfer size than you specified, the server's requested value
overrides the one you set.
3.8 /FILEIDS
/FILEIDS [= {UNIQUE | NONUNIQUE}]
Optional. Default: /FILEIDS=UNIQUE.
With UNIQUE, the client uses the file name and 32-bit NFS file ID
when processing the directory information returned by the server
to determine whether cached information is valid.
With NONUNIQUE, the client uses the file handle instead of the
file ID. This can refresh directory entries in the client's cache
more quickly. However, this can degrade performance because the
client must issue additional RPC requests to get the file handle.
3.9 /FORCE
/FORCE
/NOFORCE
Optional. Default: /NOFORCE.
Performs an overmount or a mount that can cause file system
occlusion.
Required privileges:
o OPER
o SYSPRV (for overmounting a /SYSTEM mount)
o GRPNAM (for overmounting a /GROUP mount)
3.10 /GID
/GID=n
Optional. Default: -2.
Default GID if no GID mapping exists for file access.
Restriction: Requires OPER privilege.
3.11 /GROUP
Optional. Default: User mounted.
Adds the logical name to the group logical name table. If the
mount is the first one on the volume, /GROUP marks the volume as
being group-mounted.
Restrictions:
o Requires GRPNAM privilege.
o /GROUP and /SYSTEM are mutually exclusive.
3.12 /HOST
/HOST=host
Required.
Remote NFS server on which the physical files reside. Type either
domain-name or IP-address format.
3.13 /NLM_TRANSPORT
/NLM_TRANSPORT=protocol
Optional. Default: UDP.
Protocol used for the network lock manager communication. Specify one:
o UDP
o TCP
o SCTP
3.14 /OWNER_UIC
/OWNER_UIC=n
Optional. Default: Ownership recorded on the volume.
UIC-assigned ownership of the volume while you mount it.
Applied only on the first mount of an NFS disk.
3.15 /PATH
/PATH="/path/name"
Required.
Path name on the NFS server (specified by /HOST). Must match an
exported directory, subdirectory, or file of an exported file
system on the server.
The /path/name is mounted as the master file directory (MFD) of
the specified device.
If the TYPELESS-DIRECTORIES option is set and the path name is
mapped to an OpenVMS ODS-5 disk volume, omit the file extension
.DIR from the directory name used in the path name. For example,
specify the path name for directory dove.dir;1 as dove, as in the
following command line:
MOUNT ... /PATH="/white/dove"
If TYPELESS_DIRECTORIES is not set (the default is NOTYPELESS_
DIRECTORIES), then specify the path name for directory dove.dir;1
as dove.dir.
3.16 /PROCESSOR
/PROCESSOR= {UNIQUE | SAME:DNFSn: | FILE:file}
Optional. Default: New ACP for each mount.
Associates an ancillary control process (ACP) to process the
volume, overriding the default manner in which the client
associates ACPs with NFS devices (starting a new ACP for each
mount request). The options are:
o UNIQUE
- Creates a new ACP for the new NFS device.
- Requires OPER privilege.
o SAME:DNFSn:
- Uses the same ACP as the specified device.
- Requires OPER privilege.
o FILE:file
- Creates a new ACP running the image specified by file. Do
not use wildcards, host names, or directory names.
- Requires CMKRNL or OPER privilege.
3.17 /PROTECTION
/PROTECTION=protections
Optional. Default: /PROTECTION=(S:RWED,O:RWED,G:RWED,W:RWED).
Protection code for the volume. If you omit a category, the
client denies access to that category of user.
Applied only on the first mount of an NFS device and ignored with
subsequent mounts on that device.
Restrictions: Requires OPER privilege.
3.18 /RETRIES
/RETRIES=n
Optional. Default: /RETRIES=0 (the client retries the request
forever or until the server responds).
Maximum number of read or write retries if the NFS server fails
to respond.
/RETRIES=0 is a close equivalent to a UNIX hard mount. If you
attempt to abort a command or program while it is still retrying
the I/O operation on a client device, the process enters RWAST
state and remains in that state until the NFS server responds. A
process in RWAST state cannot be terminated. If the server does
not become available, the only way remove the process without
rebooting the client host is to dismount the client device with
the /ALL qualifier.
3.19 /SERVER_TYPE
/SERVER_TYPE=type
Optional. Default: UNIX.
Operating system of the host running NFS server. The values for
type are:
o UNIX
o IBM_VM
If the server is TCP/IP Services Version 3.3 or later, this
qualifier is ignored because the client and server always operate
in OpenVMS-to-OpenVMS mode.
3.20 /SHARE
Places the logical name in the job logical name table and
increments the volume mount count regardless of the number of
job mounts. When the job logs out, all job mounts are dismounted,
causing the volume mount count to be decremented. Refer to
the HP TCP/IP Services for OpenVMS Management manual for more
information.
3.21 /STRUCTURE
Optional. Default: /STRUCTURE=2
Specifies whether the volume should be formatted in Files-11 On-
Disk Structure Level 2 (ODS-2), which is the default, or Files-11
On-Disk Structure Level 5 (ODS-5).
For more information about ODS-5 disks, refer to the HP OpenVMS
System Manager's Manual: Essentials.
3.22 /SUPERUSER
/SUPERUSER=uid
/NOSUPERUSER
Optional. Default: /NOSUPERUSER.
Maps users with SYSPRV, BYPASS, or READALL privileges to the
superuser UID. The NFS server must allow superuser access.
The normal superuser UID is 0.
/NOSUPERUSER: No mapping.
3.23 /SYSTEM
Optional. Default: System mounted.
Places the logical name in the system logical name table unless
you specify the /GROUP or /SHARE qualifier. The client deletes
the logical name from the system table when you dismount the
volume.
Restrictions:
o Requires SYSNAM privilege.
o The /GROUP, /SYSTEM, and /SHARE qualifiers are mutually
exclusive.
o Without SYSNAM or GRPNAM privilege, you must use /SHARE for a
job mount.
3.24 /TIMEOUT
/TIMEOUT=OpenVMS_delta_time
Optional. Default: ::01 (1 second).
Minimum timeout period for initial remote procedure call (RPC)
request retransmissions.
Specify the timeout period as your estimate of the typical round-
trip time for RPC requests. For slower-speed links - for example,
NFS traffic over SLIP - specify a value that is larger than the
default.
Example: For a maximum read/write size of 8192 (see the /DATA
qualifier) over a 19,200-baud SLIP line, set the absolute minimum
timeout value as follows:
10240 bytes*8 bits per byte
--------------------------- = 4.27 seconds
19,200 bits per second
Here, the 10240 bytes is 8192 data bytes plus the worst-case RPC
overhead. Because 4.27 seconds is the absolute minimum, a more
realistic value for this link is 15 to 30 seconds to allow for
other traffic.
3.25 /TRANSPORT
/TRANSPORT=protocol
Optional. Default: TCP.
Protocol used for communicating with the NFS server. Specify one:
o UDP
o TCP
o SCTP
3.26 /UID
/UID=n
Optional. Default: -2.
Default UID if no UID mapping exists for file access.
Restriction: Requires OPER privilege.
Both the NFS server and NFS client use the proxy database for
access control. HP strongly recommends that you provide a proxy
with a unique UID for every NFS client user.
If you need to provide universal access to world-readable
files, you can use the default UID to avoid creating a proxy
for every NFS client user. You should avoid using the default
UID if clients require additional file access; otherwise, client
users may see unpredictable and confusing results when they try
to create files. Refer to the HP TCP/IP Services for OpenVMS
Management manual for a detailed discussion about using proxies,
the default user, and security considerations.
3.27 /USER
/USER=user
Optional. Default: USER account.
Existing OpenVMS account to which the NFS client maps unknown
UIDs.
If the client does not find the USER account, the DECnet account
becomes the default. If the client does not find the DECnet
account, [200,200] becomes the default.
3.28 /WRITE
/WRITE
/NOWRITE
Optional. Default: /WRITE.
Mounts files with WRITE privilege.
/NOWRITE mounts files as read only.
4 – Examples
1.TCPIP> MOUNT DNFS2: /HOST="loon" /PATH="/usr/users/curlew"
Mounts, on local device DNFS2:, the remote directory
/usr/users/curlew, which resides on NFS server loon.
2.TCPIP> MOUNT DNFS3: /HOST="sigma" -
_TCPIP> /PATH="/usr" /AUTO=(INACT:00:10:00)
Using automounting, this command mounts the /usr file system
from sigma onto the OpenVMS mount point when it references the
path name. The client keeps the path mounted for an inactive
period of 10 minutes, after which it dismounts the path name.
3.TCPIP> MOUNT DNFS4: /HOST="sigma" /PATH="/usr" -
_TCPIP> /BACKGROUND=(DEL:00:01:00,RET:20) /RETRIES=4
Attempts to mount the /usr file system. If it cannot, it waits
1 minute and retries the connection up to 20 times.
4.TCPIP> MOUNT DNFS5:[USERS.MNT] /HOST="sigma" /PATH="/usr"
%DNFSMOUNT-S-MOUNTED, /usr mounted on _DNFS5:[USERS.MNT]
TCPIP> MOUNT DNFS5:[USERS.MNT] /HOST="sigma" /PATH="/usr/users" /FORCE
%DNFSMOUNT-S-REMOUNTED, _DNFS5:[USERS.MNT] remounted as /usr/users on SIGMA
Specifies a lower level in the NFS server path with the second
mount. This constitutes another path name and qualifies for an
overmount.
5.TCPIP> MOUNT DNFS22:[USERS.SMITH.MNT] /HOST="sigma" /PATH="/usr"
%DNFSMOUNT-S-MOUNTED, /usr mounted on _DNFS22:[USERS.SMITH.MNT]
TCPIP> MOUNT DNFS22:[USERS.SMITH] /HOST="sigma" /PATH="/usr" /FORCE
%DNFSMOUNT-S-MOUNTED, /usr mounted on _DFS22:[USERS.SMITH]
%TCPIP-I-OCCLUDED, previous contents of _DNFS22:[USERS.SMITH] occluded
The /FORCE qualifier performs an occluded mount. If you issue
the DIRECTORY command, the NFS client occludes (hides from
view) the subdirectory dropped from the first MOUNT command.
To make the directory visible again, either issue the SHOW
MOUNT command (both mounts will be visible) or dismount
DNFS22:[USERS.SMITH].
6.TCPIP> MOUNT DNFS0: BOOK1 BEATRICE -
_TCPIP> /PATH="/INFERNO" -
_TCPIP> /HOST="FOO.BAR.EREWHON" -
_TCPIP> /STRUCTURE=5 -
_TCPIP> /SYSTEM
Mounts path INFERNO with label BOOK1 and logical name BEATRICE.
Specifies the volume structure as ODS-5.