For guidance information about selecting values for the statements,
see the section on customizing z/OS UNIX in z/OS UNIX System Services Planning.
- ALTROOT FILESYSTEM('fsname') PARM('parameter') MOUNTPOINT('pathname')
- Specifies the alternate sysplex root file system that is to automatically
replace the current sysplex root file system when the current one
becomes unowned.
When the replacement occurs, the alternate sysplex
root file system is unmounted and then mounted again as the current
sysplex root on all systems in the shared file system configuration.
If the replacement is successful, the failed sysplex root is unmounted.
The
alternate sysplex root file system is always mounted in the READ mode
and designated as AUTOMOVE=YES when it becomes the current sysplex
root file system. The previous mount mode and AUTOMOVE settings of
the alternate sysplex root and the failed sysplex root are ignored.
The mount parameters are preserved when the alternate sysplex root
file system becomes the current sysplex root file system.
To
manually replace the failed or failing sysplex root file system, use
the MODIFY OMVS,NEWROOT operator command with the COND=FORCE option.
To validate the syntax of the ALTROOT statement, use the SETOMVS SYNTAXCHECK
operator command. See the steps for dynamically
replacing the sysplex root file system topic in z/OS UNIX System Services Planning for
step-by-step instructions of manually replacing the failed or failing
sysplex root. For more information about the mentioned system commands,
see z/OS MVS System Commands.
- FILESYSTEM('fsname')
- The name of the alternate sysplex root file system.
- PARM('parameter')
- The parameter to be passed directly to the file system type.
- MOUNTPOINT('pathname')
- The absolute path name, or a symbolic link that resolves to the
path name of the directory onto which the file system is to be mounted.
MOUNTPOINT must be specified in the ALTROOT statement. The mount point
for the alternate sysplex root file system must reside in the root
directory of the current sysplex root file system, otherwise mount
of the ALTROOT will fail. The absolute path name can be up to 64 characters.
- AUTHPGMLIST('/etc/authfile')|NONE
- Specifies
the path name of a hierarchical file system (HFS) file that contains
the lists of APF-authorized path names and program names. If you do
not specify a value for AUTHPGMLIST, or if you specify NONE, invocations
of APF-authorized and program controlled programs will not be checked
against a list of authorized programs or authorized path names. If
you specify a path name for the AUTHPGMLIST parameter, the system
checks this list during hfsload, exec and spawn processing. If the
target program of an exec or spawn has an authorization code of 1
(AC=1), then that program name must appear in the authorized program
list.
Use the SETOMVS or SET OMVS command to dynamically change
the value of AUTHPGMLIST. To make a permanent change, edit the BPXPRMxx
member that will be used for IPLs.
For more information, see z/OS UNIX System Services Planning.
- AUTOCVT(ALL|ON|OFF)
- Activates
and deactivates automatic conversion of I/O data using coded character
sets for the program and its associated files. The coded character
set identifiers (CCSIDs) are specified by the program or by setting
the appropriate environment variables at run time. The system AUTOCVT
indicator can be overridden by individual programs at a thread or
file descriptor level; AUTOCVT is a controlling global switch only
for existing programs that do not explicitly establish their own conversion
environment.
Default: OFF
You
can use the SETOMVS or SET OMVS commands to change the value of AUTOCVT
between ALL, ON, and OFF. Changing this conversion mode does not affect
conversion of opened files for which I/O has already started.
AUTOCVT(ALL) enables Unicode Services conversion of data
during read and write I/O operations.
AUTOCVT(ON)
enables Enhanced ASCII conversion of data during read and write I/O
operations. Conversion is limited to the EBCDIC (1047) and ASCII (819)
code pages.
Guideline: When AUTOCVT(ALL)
or AUTOCVT(ON) is set, every read and write operation for a file must
be checked to see if conversion is necessary, which incurs a performance
penalty involved, even if no conversion occurs. If possible, keep
AUTOCVT(OFF) and have each program enabled for conversion. To do this,
set the compile or runtime environment variables that control conversion
or use programmable function, as described in
z/OS UNIX System Services Planning.
Automatic
conversion can also be controlled individually by a program with one
of the following flags in the thread Thli control block (BPXYTHLI):
ThliCvtOn - Activates automatic conversion for this thread.
ThliCvtOff - Deactivates automatic conversion for this thread.
Both
bits must not be on at the same time.
Automatic conversion is accomplished between programs
and files that are tagged with different CCSIDs when a conversion
table exists for that CCSID pair in the system. CCSID values are defined
in Character Data Representation Architecture. For information about supported Unicode Services CCSIDs,
see z/OS Unicode Services User's Guide and Reference.
- AUTOMOVE|NOAUTOMOVE
- The AUTOMOVE and NOAUTOMOVE parameters apply only
in a sysplex where systems are participating in shared file system.
They indicate what happens to the ownership of the file system when
a shutdown, PFS termination, dead system takeover, or file system
move occurs. AUTOMOVE indicates that ownership of the file system
automatically changes to another system that is participating in the
shared file system. NOAUTOMOVE indicates that ownership of the file
system is not moved if the owning system goes down; as a result, the
file system becomes inaccessible.
Note: When you specify NOAUTOMOVE,
the file system becomes inaccessible when the owning system goes down.
But it still exists in the file system hierarchy. The file system
remains unowned until the original owing system re-IPLs.
Use AUTOMOVE for the sysplex root file system and the
version file system. For file systems that are associated with a single
system, specify UNMOUNT; this includes /etc, /tmp, /var, /dev,
and other system-specific file systems. For descriptions of the sysplex
root, system-specific, and version file systems, see Sharing
file systems in a sysplex in z/OS UNIX System Services Planning.
To
ensure that the root is always available, use the default.
Default: AUTOMOVE
- CTRACE(parmlib_member_name)
- Specifies the
parmlib member that contains the initial tracing options to be used
for the z/OS UNIX component. Use this statement
to provide tracing while the kernel is starting and to avoid having
to issue a TRACE operator command to set tracing options.
Default:
CTIBPX00
- FILESYSTYPE TYPE(type_name) ENTRYPOINT(entry_name) PARM('parm')
ASNAME(proc_name[,'start_parms'])
- Specifies
the type of file system that is to be started. BPXPRMxx can contain
more than one FILESYSTYPE statement.
When SYSPLEX(YES) is specified,
each FILESYSTYPE in use within the participating shared file system
group must be defined for all systems that are participating in shared
file system. The easiest way to accomplish this is by having a single
BPXPRMxx member that contains file system information for each system
that is participating in shared file system. If you decide to define
a BPXPRMxx for each system, the FILESYSTYPE statements must be identical
on each system. For more information about shared file system, see z/OS UNIX System Services Planning.
Any
facilities that are required for a particular FILESYSTYPE must be
initiated on all systems that are participating in shared file system.
For example, NFS requires TCP/IP, so if you specify an NFS FILESYSTYPE,
you must also initialize TCP/IP on NFS initialization.
The
SETOMVS RESET command can be used to dynamically specify new FILESYSTYPE
statements. To make a permanent change, edit the BPXPRMxx member that
is used for IPLs. For more information, see Dynamically adding FILESYSTYPE statements in BPXPRMxx in z/OS UNIX System Services Planning.
The
parameters are as follows:
- ASNAME(proc_name[,'start_parms'])
- proc_name specifies
the name of a procedure in SYS1.PROCLIB that is to be used to start
the address space that is initialized by the physical file system
(PFS). Specify ASNAME for any PFS that does not run in the kernel
address space. The name that you specify is also used for the name
of the address space.
start_parms is an optional
quoted string that is to be appended to the proc_name when
the address space is started. The string can be up to 100 characters
long. The start_parms are not validated; they
are just passed to the system when the address space is started with
an internal command. Refer to the START command in z/OS MVS System Commands or
the ASCRE macro in z/OS MVS Programming: Authorized Assembler Services Reference ALE-DYN.
By
default the address space started with ASNAME is started under JES,
but this can be changed by including the additional start_parms SUB=MSTR.
ASNAME
is an optional parameter. proc_name is 1 to 8 characters;
the system converts the name to uppercase. If you do not specify
ASNAME, or specify proc_name as the name of the
kernel address space, the PFS is initialized in the kernel address
space.
Refer to the documentation for the specific physical
file system for valid ASNAME operands.
- PARM('parm')
- Provides a parameter
to be passed directly to the file system type. The parameter format
and content are specified by the file system type.
PARM is an
optional parameter. The parameter is up to 1024 characters long;
the characters can be in uppercase, lowercase, or both. The parameter
must be enclosed in single quotation marks.
Note: For
information about specifying the TFS configuration parameters, see
the section on parameter key options for the FILESYSTYPE statement
in
z/OS UNIX System Services Planning. For
example:
- -ea count
- Allows the TFS file system to automatically grow count times.
- -em count
- Allows the TFS file system to manually grow count times.
If the physical file system specified
does not expect a PARM operand, it ignores all PARM operands.
PRM=(aa,
bb, ..., zz) is valid for the z/OS® File
System (zFS).
SYNCDEFAULT(
t),
VIRTUAL(
max), and FIXED(
min)
are valid only when ENTRYPOINT is GFUAINIT. FSFULL(
threshold,
increment)
is supported by HFS,
TFS, and zFS.
Note: If
a syntax error is found in any of these four parameters (SYNCDEFAULT,
VIRTUAL, FIXED, or FSFULL), the system issues an error message issued
and set all four parameters to the default values.
- PRM=(aa, bb, ..., zz) is
used to define the IOEPRMyy members of the logical parmlib concatenation
for specifying zFS configuration parameters, where:
- aa, bb, ..., zz specify the suffixes of the
IOEPRMxx parmlib members in the order in which they are to be concatenated.
Up to 32 suffixes may be specified. Suffixes must be in uppercase.
Defining an IOEFSPRM configuration file with an IOEZPRM DD card
in the zFS procedure will override the PRM specification (that is,
cause it to be ignored). See Specifying the IOEPRMxx concatenation for
examples and additional syntax information.
- SYNCDEFAULT(t)
- t specifies the number of seconds used as
a default for the sync daemon interval. When the sync daemon is active,
the metadata for a file system is hardened. Setting t to
0 indicates that the file system should harden metadata synchronously
with syscall requests.
- Sync interval values are rounded up to the next 30-second value.
For example, specifying 31 seconds results in a sync interval of 60
seconds.
- The maximum value that can be specified for t is 65534.
Values between 65534 and 99999 are rejected.
- A value of 99999 specifies that no sync daemon intervals are
specified, and thus, the metadata is not hardened.
- Default: 60 seconds
- VIRTUAL(max)
- max specifies the maximum amount of virtual
storage (in megabytes) that file system data and metadata buffers
should use.
- If you do not set a value for max, the system
assigns a default value that is equal to half the amount of real
storage available to the system when z/OS is
initialized. See z/OS UNIX System Services Planning for
more information about the VIRTUAL(max) parameter.
- FIXED(min)
- min specifies the amount of virtual storage
(in megabytes) that is fixed when z/OS is
initialized and remains fixed even if file system activity drops to
zero. min must be less than or equal to VIRTUAL(max).
- min cannot exceed 50% of real storage available
to the system. If the allowed amount of storage is exceeded, an informational
message is issued and min is set to 50% of real
storage. The minimum limit can be changed dynamically by invoking
the confighfs shell command. See z/OS UNIX System Services Command Reference for
more information about the confighfs shell
command.
- Default: 0
- FSFULL(threshold,increment)
- threshold specifies the percentage of the file
system (HFS, TFS, or zFS) capacity at which
an operator message is generated. The default is 100%.
- increment specifies the percentage of change
above the file system capacity at which an operator message is generated.
Messages are generated by either an increase or decrease greater than increment.
The default is 5%.
You can specify threshold and increment values
for the file system. The values can also be set on the MOUNT command
for a specific file system. Parameters on the MOUNT command override
parmlib values. If no values are specified in either place, no threshold
checking is done. If a threshold value is specified but no increment
is given, the increment defaults to 5%. The increment value applies
both to upgrading the message when the file system continues to fill
and to removing the message when more space becomes available due
to either deleting files, or to extending the file system. The values
are in terms of percent full. The values that are applied to a file
system can be changed only when the file system is mounted.
- TYPE(type_name)
- Specifies the
name of the file system type that is to control the file system. TYPE
is a required parameter. The name is 1 to 8 characters; the system
converts the name to uppercase. In the FILESYSTYPE statement, specify
one of the following types of the file system:
- HFS for a hierarchical file system (HFS).
- zFS for a z/OS File
System (zFS).
- NFS for accessing remote files.
- TFS for a temporary file system (TFS).
For planning information, see FILESYSTYPE in z/OS UNIX System Services Planning.
- FORKCOPY(COW|COPY)
- Specifies
how user storage is to be copied from the parent process to the child
process during a fork() system call.
FORKCOPY(COW)specifies that
all fork() calls are processed with the copy-on-write mode if the
suppression-on-protection (SOP) hardware feature is available. Before
the storage is modified, both the parent and child process refer to
the same view of the data. The parent storage is copied to the child
only if either the parent or the child modifies the storage. FORKCOPY(COW)
causes the system to use the ESQA to manage page sharing.
FORKCOPY(COPY)
specifies that fork() immediately copies the parent storage to the
child, whether the SOP is available or not. Use this option to avoid
any additional ESQA use in support of fork.
Follow these guidelines:
- If the runtime library is in the link pack area, specify FORKCOPY(COPY).
- If the runtime library is not in the link pack area, specify FORKCOPY(COW).
Default: COW
Use the SETOMVS or SET OMVS
command to change the value of FORKCOPY dynamically. To make a permanent
change, edit the BPXPRMxx member used for IPLs.
- IPCMSGNIDS(nnnnn)
- Specifies
the maximum number of unique system-wide message queues.
Value
Range: nnnnn is a decimal value from 1 to 20000.
Default:
500
Use the SETOMVS or SET OMVS command to change the value
of IPCMSGNIDS dynamically. The new minimum is the current value. The
new maximum is calculated as follows:
MIN(initial maximum,MAX(4096,3*initial value))
You
can increase but not decrease the value, as described in z/OS UNIX System Services Planning.
- IPCMSGQBYTES(nnnnn)
- Specifies
the maximum number of bytes in a single message queue.
Value
Range: nnnnn is a decimal value from 0 to 2147483647.
Note: The
high end of this range is not obtainable due to storage constraints.
The actual maximum range varies due to storage allocation and system
usage.
Default: 2147483647 (2G)
Use
the SETOMVS or SET OMVS command to change the value of IPCMSGQBYTES
dynamically.
- IPCMSGQMNUM(nnnnn)
- Specifies
the maximum number of system-wide messages for each queue.
Value
Range: nnnnn is a decimal value from 0 to 2147483647.
Note: The
high end of this range is not obtainable due to storage constraints.
The actual maximum range varies due to storage allocation and system
usage.
Default: 10000
Use the SETOMVS or
SET OMVS command to change the value of IPCMSGQMNUM dynamically.
- IPCSEMNIDS(nnnnn)
- Specifies
the maximum number of unique system-wide semaphore sets.
Value
Range: nnnnn is a decimal value from 1 to 20000.
Default:
500
You can change the value of IPCSEMNIDS dynamically using
the SETOMVS or SET OMVS command, as described in z/OS UNIX System Services Planning.
- IPCSEMNOPS(nnnnn)
- Specifies
the maximum number of operations for each semop call.
Value
Range: nnnnn is a decimal value from 0 to 32767.
Default:
25
You can change the value of IPCSEMNOPS dynamically using
the SETOMVS or SET OMVS command.
- IPCSEMNSEMS(nnnnn)
- Specifies
the maximum number of semaphores for each semaphore set.
Value
Range: nnnnn is a decimal value from 0 to 32767.
Default:
1000
You can change the value of IPCSEMNSEMS dynamically using
the SETOMVS or SET OMVS command.
- IPCSHMMPAGES(nnnnn)
- Specifies
the maximum number of pages for shared memory segments.
Value
Range: nnnnn is a decimal value from 1 to 4 petabytes (that is, 4 * 1 125 899 906 842 624).
If
you obtain memory segments below the 2-gigabyte address range, then
a realistic maximum is about 1.5 gigabytes; the actual maximum range
varies due to storage allocation and system usage. If you obtain
memory segments above the 2-gigabyte address range, the maximum depends
on the IEASYS HVSHARE parameter, which specifies the size of the high
virtual shared area.
Default: 25600
Use the
SETOMVS or SET OMVS command to change the value of IPCSHMMPAGES dynamically.
- IPCSHMNIDS(nnnnn)
- Specifies
the maximum number of unique system-wide shared memory segments.
Value
Range: nnnnn is a decimal value from 1 to 20000.
Default:
500
Use the SETOMVS or SET OMVS command to change the value
of IPCSHMNIDS dynamically. The new minimum is the same as the current
value. The new maximum is calculated as follows:
MIN(initial maximum,MAX(4096,3*initial value))
You
can increase but not decrease the value, as described in z/OS UNIX System Services Planning.
- IPCSHMNSEGS(nnnnn)
- Specifies
the maximum number of attached shared memory segments for each address
space.
Value Range: nnnnn is a decimal
value from 0 to 1000.
Default: 10
You can change
the value of IPCSHMNSEGS dynamically using the SETOMVS or SET OMVS
command.
- IPCSHMSPAGES(nnnnn)
- Specifies
the maximum number of system-wide shared pages that are created by
calls to the fork and 31–bit shmat functions.
You can increase,
but not decrease, the value, as described in z/OS UNIX System Services Planning.
Shared memory segments obtained above the 2-gigabyte range in 64-bit
programs do not affect this limit.
Value Range: nnnnn is
a decimal value from 0 to 2621440.
You can set
a denomination (or multiplier) value when defining this value. The
C suffix can have a one-character value as presented in Table 1, but must not exceed
the parameter-specific upper limit. The denomination value is retained
and is used again within a subsequent D OMVS command.
Default:
262144
You can change the value of IPCSHMSPAGES dynamically
using the SETOMVS or SET OMVS command. The new minimum is the same
as the current value. The new maximum is calculated as follows:
MIN(initial maximum,MAX(4096,3*initial value))
You
can increase but not decrease the value, as described in z/OS UNIX System Services Planning.
Because
each page of shared storage requires the associated consumption of
extended system queue area (ESQA) storage, limiting the shared storage
usage provides a way to limit the ESQA usage by z/OS UNIX users.
If you use the __IPC_MEGA or __MAP_MEGA options, then the shared pages
limits are not affected because MEGA does not affect the system ESQA
overhead.
- LIMMSG(NONE|SYSTEM|ALL)
- Specifies how
console messages that indicate when parmlib limits are reaching critical
levels are to be displayed.
- NONE
- Do not display console messages when any of the parmlib limits
are reached.
- SYSTEM
- Display console messages for all processes that reach system limits.
In addition, messages are displayed for each process limit of a process
if any of the following conditions are met:
- The process limit or limits are defined in the OMVS segment of
the owning user ID
- The process limit or limits were changed with a SETOMVS PID=pid,process_limit
- ALL
- Display console messages for both the system limits and the process
limits, regardless of which process reaches a process limit.
Default: NONE
- LOSTMSG(ON|OFF)
- LOSTMSG(ON)
detects lost and duplicate XCF messages in a shared file system configuration.
It is ignored if the file system does not have a shared file system
configuration; for example, when SYSPLEX(NO) is specified. While LOSTMSG
can be specified differently for each member in the sysplex, the same
LOSTMSG setting should be specified throughout the sysplex. To disable
the detecting of lost and duplicate messages, specify LOSTMSG(OFF).
Tip: Do not use LOSTMSG(ON) when z/OS UNIX sysplex
traffic is high, such as when many file systems that are not sysplex-aware
are being accessed remotely because performance might be affected.
LOSTMSG(ON) is the default.
- MAXASSIZE(nnnnn)
- Specifies
the RLIMIT_AS resource values that will be established as the initial
values for new processes. RLIMIT_AS indicates the address space region
size. For more information about RLIMIT_AS, refer to the description
of setrlimit in z/OS UNIX System Services Programming: Assembler Callable Services Reference.
The
soft limit is obtained from MVS™;
if it is greater than the MAXASSIZE value, the soft limit is set to
the hard limit. This value is also used when processes are initiated
by a daemon process using an exec() after setuid(). In this case,
both the RLIMIT_ AS hard and soft limit values are set to the MAXASSIZE
specified value.
When processes are initiated by a daemon process
using an exec()after setuid(), this value is used. Therefore, MAXASSIZE
will be the region size for all processes created through rlogin or
telnet. In this case, both the RLIMIT_AS hard and soft limit values
are set to the MAXASSIZE value.
A superuser can override this
value by specifying a new region size in the spawn inheritance structure
on __spawn(). Or you can change the value of MAXASSIZE dynamically
by using the SETOMVS or SET OMVS command. This change only affects
the new processes that are created after the change was made.
Note: The
IEFUSI user exit can modify the region size of an address space.
Users are strongly discouraged from altering the region size of address
spaces in the OMVS subsystem category.
Value
Range: nnnnn is a decimal value from 10485760
(10 megabytes) to 2147483647 (2 gigabytes).
You
can set a denomination (or multiplier) value when defining this value.
The C suffix can have a one-character value as presented in Table 1, but must not exceed
the parameter-specific upper limit. The denomination value is retained
and is used again within a subsequent D OMVS command.
Default:
209715200
Use the SETOMVS or SET OMVS command to dynamically
increase or decrease the value of MAXASSIZE. To make a permanent change,
edit the BPXPRMxx member that is used for IPLs.
For planning
information, see MAXASSIZE in z/OS UNIX System Services Planning.
- MAXCORESIZE(nnnnn)
- Specifies the RLIMIT_CORE soft and hard resource values
that will be established as the initial values for new processes. RLIMIT_CORE
indicates the maximum core dump file size (in bytes) that a process
can create. It also specifies the limit when they are initiated by
a daemon process using an exec() after a setuid(). For more information
about RLIMIT_CORE, see the description of setrlimit() in z/OS UNIX System Services Programming: Assembler Callable Services Reference.
Value
Range: nnnnn is a decimal value from 0 to 2147483647
(2 gigabytes).
You can set a denomination (or
multiplier) value when defining this value. The C suffix can have
a one-character value as presented in Table 1, but must not exceed
the parameter-specific upper limit. The denomination value is retained
and is used again within a subsequent D OMVS command.
Default:
4194304 (4 megabytes) Specifying a value of 2147483647 (2 gigabytes)
indicates an unlimited core file size.
Use the SETOMVS or SET
OMVS command to dynamically increase or decrease the value of MAXCORESIZE.
To make a permanent change, edit the BPXPRMxx member that is used
for IPLs.
- MAXCPUTIME(nnnnn)
- Specifies
the RLIMIT_CPU resource values that will be established as the initial
values for new processes. RLIMIT_CPU indicates the CPU time, in seconds,
that a process can use. For more information about RLIMIT_CPU, refer
to the description of setrlimit() in z/OS UNIX System Services Programming: Assembler Callable Services Reference.
If
the soft limit value from MVS is
greater than the MAXCPUTIME value, the soft limit is set to the hard
limit. This value is also used when processes are initiated by a
daemon process using an exec() after setuid(). In this case, both
the RLIMIT_CPU hard and soft limit values are set to the MAXCPUTIME
value.
A superuser can override this value by specifying a
new time limit in the spawn inheritance structure on __spawn().
For
processes that are running in or forked from TSO or BATCH, the MAXCPUTIME
value has no effect. The TIME limit is inherited from the parent.
If a TIME parameter is specified on the JCL for the started task,
then that value is used. If not, then the TIME value is taken from
the JES default TIME value.
For processes created by the rlogind
command or other daemons, MAXCPUTIME is the time limit for the address
space.
Specifying a MAXCPUTIME or CPUTIMEMAX of 86400 seconds
disables the JWT, SWT, or TWT timeout the same way that JCL TIME=1440
does. See Parameters for SMFPRMxx for details on
the JWT, SWT, and TWT parameters.
Value
Range: nnnnn is a decimal value from 7 to 2147483647
seconds.
Default: 1000
Use the SETOMVS or SET
OMVS command to dynamically increase the value of MAXCPUTIME. To make
a permanent change, edit the BPXPRMxx member that will be used for
IPLs.
For planning information, see MAXCPUTIME in z/OS UNIX System Services Planning.
- MAXFILEPROC(nnnnnn)
- Specifies
the maximum number of descriptors for files, sockets, directories,
and any other file system objects that a single process can have concurrently
active or allocated. MAXFILEPROC is the same as the OPEN_MAX variable
in the POSIX standard.
Value Range: nnnnnn is
a decimal value from 3 to 524287.
Default: 64000
Use
the SETOMVS or SET OMVS command to dynamically increase or decrease
the value of MAXFILEPROC. To make a permanent change, edit the BPXPRMxx
member that is used for IPLs.
On the ADDUSER command, you can
increase the maximum number of open descriptors on a per process basis.
The FILEPROCMAX parameter on the ADDUSER command overrides the MAXFILEPROC
parameter in the BPXPRMxx profile.
For planning information,
see MAXFILEPROC in z/OS UNIX System Services Planning.
- MAXFILESIZE(nnnnn|NOLIMIT)
- Specifies
the RLIMIT_FSIZE soft and hard resource values that will be established
as the initial values for new processes. RLIMIT_FSIZE indicates the
maximum file size (in 4 KB increments) that a process can create.
It also specifies the limit when they are initiated by a daemon process
using an exec() after a setuid(). For more information about RLIMIT_FSIZE,
see the description of setrlimit() in z/OS UNIX System Services Programming: Assembler Callable Services Reference.
Value
Range: nnnnn is a decimal value from 0 to 2147483647,
which indicates an unlimited file size. If MAXFILESIZE is not specified
or MAXFILESIZE(NOLIMIT) is specified, there will be no limit to the
size of files created, except for the architectural limit of the system.
You can set a denomination (or multiplier) value when
defining this value. The C suffix can have a one-character value as
presented in Table 1,
but must not exceed the parameter-specific upper limit. The denomination
value is retained and is used again within a subsequent D OMVS command.
If
you specify 0, the process does not create any files. Omitting this
statement indicates an unlimited file size.
Default: NOLIMIT
Use
the SETOMVS or SET OMVS command to dynamically increase or decrease
the value of MAXFILESIZE. To make a permanent change, edit the BPXPRMxx
member that is used in IPLs.
- MAXIOBUFUSER(nnnnn)
- Specifies
that maximum amount of persistent I/O virtual storage that z/OS UNIX obtains
on behalf of a user when a process performs I/O in a Unicode Services
conversion environment (that is, the AUTOCVT setting is ALL). This
storage remains allocated for the life of an open file. The amount
that is allocated for each open depends on the CCSID of the file and
the size of a read or write requests that are used by the process.
The limit does not apply to UID 0 processes.
Value Range: 0
to 2 G, representing the number of megabytes of storage. Thus, 2G
represents 2G x 1M = 2P (petabytes) of storage.
Default:
2048 (equivalent to 2G of storage).
- MAXMMAPAREA(nnnnn)
- Specifies
the maximum amount of data space storage space (in pages) that can
be allocated for memory mappings of z/OS UNIX files. Storage is not allocated
until the memory mapping is active.
Using memory map services
causes a large amount of system memory to be consumed. For each page
(4 KB) that is memory-mapped, 96 bytes of ESQA are consumed when a
file is not shared with any other users. When a file is shared by
multiple users, each user after the first causes an additional 32
bytes of ESQA to be consumed for each shared page. Assuming that the
default of 40960 pages is taken, and assuming that no sharing is done
by mmap() users, a maximum of 3840 KB of ESQA could be consumed. The
ESQA storage is consumed when the mmap() function is invoked rather
than when the page is accessed by the memory mapping application program.
If
you have applications using the __MAP_MEGA option, you can map very
large files without the system overhead in ESQA. For more information,
see Extended System Queue Area (ESQA) in z/OS UNIX System Services Planning.
Value
Range: nnnnn is a decimal value from 1 to 16777216.
You can set a denomination (or multiplier) value when
defining this value. The C suffix can have a one-character value as
presented in Table 1,
but must not exceed the parameter-specific upper limit. The denomination
value is retained and is used again within a subsequent D OMVS command.
Default:
40960
You can change the value of MAXMMAPAREA dynamically using
the SETOMVS or SET OMVS command. To make a permanent change, edit
the BPXPRMxx member that will be used for IPLs.
For planning
information, see MAXMMAPARE in z/OS UNIX System Services Planning.
- MAXPIPEUSER(nnnnn)
- Specifies
the maximum number of z/OS UNIX named or unnamed pipes that
a real UID can open concurrently. The value only applies to non-UID=0
users. For UID=0 users, the MAXPIPERUSER limit of 8730 is always enforced.
Value
Range: nnnnn is a decimal value from 256 to
8730.
Default: 8730.
Use the SETOMVS or SET OMVS
command to dynamically increase or decrease the value of MAXUIDS.
To make a permanent change, edit the BPXPRMxx member that will be
used for IPLs.
For planning information, see z/OS UNIX System Services Planning.
Syntax
example in BPXPRMxx: MAXPIPEUSER(8730)
- MAXPROCSYS(nnnnn)
- Specifies
the maximum number of processes that the system allows.
Value
Range: nnnnn is a decimal value from 5 to 32767.
Default:
900
You can use the SETOMVS or SET OMVS command to dynamically
increase or decrease the value of MAXPROCSYS. To make a permanent
change, edit the BPXPRMxx member that is used for IPLs.
If you
are using SETOMVS or SET OMVS to change the value, the new value must
be within a certain range, or you will get an error message. The range
that you can use has a minimum value of 5; the maximum value is based
on the following calculation:
MIN(32767,MAX(4096,3*initial value))
The
initial value is the MAXPROCSYS value that was specified during BPXPRMxx
initialization. You cannot use a value less than 5. If you want to
use a value greater than the current maximum (as calculated by the
formula) but lower than the initial maximum (32767), you must change
the value in BPXPRMxx and re-IPL. For an example of how to calculate
the maximum value in the range, see the section on dynamically changing
certain BPXPRMxx parameter values in z/OS UNIX System Services Planning.
For planning information, also see MAXPROCSYS in z/OS UNIX System Services Planning.
- MAXPROCUSER(nnnnn)
- Specifies
the maximum number of processes that a single z/OS UNIX user
ID can have concurrently active, regardless of how the processes were
created. MAXPROCUSER is the same as the CHILD_MAX variable in the
POSIX standard.
A value of 25 is required for FIPS 151-2 compliance
and a value of 16 is required for POSIX.1 (ISO/IEC 9945-1:1990[E]
IEEE Std 1003.1-1990) standard compliance.
The number of processes
is tracked by user ID (UID). When a user attempts to create a new
process, the limit value for the user (defined by either the user
profile or the default OPTN value) is compared to the value maintained
for the user's UID. If the user maximum is larger than the current
process count for the UID, the user can create another process. If
not, the user is not allowed to create a new process. For example,
if user A, with a user-defined limit of 10, tries to create a process
and the UID limit is already 12, user A is not allowed to create the
new process. Since only 12 processes are currently created, user B,
with a user-defined limit of 20, is allowed to create a new process.
Use
the SETOMVS or SET OMVS command to dynamically increase or decrease
the MAXPROCUSER values. To make a permanent change, edit the BPXPRMxx
member that is used for IPLs.
There are certain
cases when a new process can be created with a UID that has already
reached the MAXPROCUSER number of processes. Super users (with UID=0)
and processes blind dubbed with the default OMVS segment are not limited
by MAXPROCUSER.
For planning information, see MAXPROCUSER in z/OS UNIX System Services Planning.
Value
Range: nnnnn is a decimal value from 3 to 32767.
Default:
25
- MAXPTYS(nnnnn)
- Specifies
the maximum number of pseudoterminals (pseudo-TTYs or PTYs) for the
system.
Value Range: nnnnn is a decimal
value from 1 to 10000.
Default: 800
You can
use the SETOMVS or SET OMVS command to dynamically increase the value
of MAXPTYS. To make a permanent change, edit the BPXPRMxx member that
is used for IPLs.
If you are using SETOMVS or SET OMVS to change
the value, the new value must be within a certain range. If it is
outside the range, you will get an error message. To use a value that
is outside this range, you must change the MAXPTYS specification in
BPXPRMxx and re-IPL. The range's minimum value is 1 and the maximum
is based on the following calculation:
MIN(10000,MAX(256,2*initial value)
The
initial value is the MAXPTYS value that was specified during BPXPRMxx
initialization. For an example of how to calculate the maximum value
in the range, see the section on dynamically changing BPXPRMxx values
in
z/OS UNIX System Services Planning.
For
planning information, see MAXPTYS in z/OS UNIX System Services Planning.
- MAXQUEUEDSIGS(nnnnnn)
- Specifies
the maximum number of signals that z/OS UNIX allows to be concurrently queued
within a single process.
Value Range: nnnnnn is
a decimal value from 1 to 100000.
Default: 1000
You
can change the value of MAXQUEUEDSIGS dynamically by using the SETOMVS
or SET OMVS command. To make a permanent change, edit the BPXPRMxx
member that will be used for future IPLs.
- MAXSHAREPAGES(nnnnn)
- Specifies
the maximum amount of shared system storage pages that can be used
by z/OS UNIX functions. The purpose of MAXSHAREPAGES
is to limit the amount of ESQA storage necessary to maintain the shared
pages. See Predicting and limiting ESQA usage in z/OS UNIX System Services Planning for
the formula to determine the maximum amount of ESQA
that is consumed during loading user-shared libraries.
The usage
of shared pages is helpful but not critical to the loading of user
shared library modules, ptrace, and fork; it serves to increase performance
but does not affect functionality. As the amount of shared pages being
used reaches certain limits, fewer functions are allowed to continue
using them. User shared library loads, ptrace, and fork stop using
shared pages when the limit reaches 60% (the only time shared storage
is used by the fork service is when FORKCOPY(COW) is specified),
mmap stops at 80%, and shmat, the most critical function, uses shared
pages until their total capacity has been reached. If
while running a 64–bit program, you allocate shared memory segments
above the bar by using the shmget() service, the shared page limit
is not affected.
Because each page of shared storage requires
the associated consumption of extended system queue area (ESQA) storage,
limiting the shared storage usage provides a way to limit the ESQA
usage by z/OS UNIX users. If you use the __IPC_MEGA
or __MAP_MEGA options, then the shared pages limits are not affected
because MEGA does not affect the system ESQA overhead.
Note: Evaluate
adjusting MAXSHAREPAGES on an active system. Dynamically decreasing
the number of pages available to ESQA for active work can cause errors.
This is due to the fact that for those jobs, the ESQA limit may now
be reached or exceeded. It is possible that shared programs will not
be able to be loaded and fork() may not succeed. This situation will
exist until the workload adjusts to the new lower limit.
Value
Range: nnnnn is a decimal value from 0 to 32768000
specifying a number of 4K pages.
You can set a denomination
(or multiplier) value when defining this value. The C suffix can have
a one-character value as presented in Table 1, but must not exceed
the parameter-specific upper limit. The denomination value is retained
and is used again within a subsequent D OMVS command.
Default:
131072
Use the SETOMVS or SET OMVS command to dynamically increase
or decrease the MAXSHAREPAGES value. To make a permanent change, edit
the BPXPRMxx member that will be used for IPLs.
- MAXTHREADS(nnnnnn)
- Specifies
the maximum number of pthread_created threads, including running,
queued, and exited but undetached, that a single process can have
concurrently active. Specifying a value of 0 prevents applications
from using pthread_create.
Value Range: nnnnnn is
a decimal value from 0 to 100000.
Default: 200
You
can change the value of MAXTHREADS dynamically using the SETOMVS or
SET OMVS command. To make a permanent change, edit the BPXPRMxx member
that will be used for IPLs.
For planning information, see MAXTHREADS
in z/OS UNIX System Services Planning.
- MAXTHREADTASKS(nnnnn)
- Specifies
the maximum number of MVS tasks
that a single process can have concurrently active for pthread_created
threads.
Value Range: nnnnn is a decimal
value from 0 to 32768.
Default: 1000
You can change
the value of MAXTHREADTASKS dynamically using the SETOMVS or SET OMVS
command. To make a permanent change, edit the BPXPRMxx member that
is used for IPLs.
For planning information, see MAXTHREADTASKS
in z/OS UNIX System Services Planning.
- MAXUIDS(nnnnn)
- Specifies
the maximum number of z/OS UNIX user IDs (UIDs) that can operate
concurrently.
Value Range: nnnnn is
a decimal value from 1 to 32767.
Default: 200
Use
the SETOMVS or SET OMVS command to dynamically increase or decrease
the value of MAXUIDS. To make a permanent change, edit the BPXPRMxx
member that will be used for IPLs.
For planning information,
see the section on MAXUIDS in z/OS UNIX System Services Planning.
- MAXUSERMOUNTSYS(nnnn)
- Specifies the maximum number of nonprivileged user mounts for
the system or for the shared file system configuration environment.
The MAXUSERMOUNTSYS limit applies only to nonprivileged users and
does not affect privileged mounts.
Value Range: nnnnn is
a decimal value from 0 to 35000 that indicates the maximum number
of nonprivileged user mounts allowed in the system. For those using
a shared file system configuration, this value is the maximum number
of nonprivileged user mounts allowed in the shared file system environment.
The most recent specification prevails for all of the systems that
are participating in a shared file configuration.
Default: 0.
Use
the SETOMVS or SET OMVS command to dynamically increase or decrease
the value of MAXUSERMOUNTSYS. To make a permanent change, edit the
BPXPRMxx member that is used for IPLs. If MAXUSERMOUNTSYS is not specified,
default values are used. The default value indicates that user mounts
are not allowed in the system.
For planning information, see
MAXUSERMOUNTSYS in z/OS UNIX System Services Planning.
- MOUNT FILESYSTEM('fsname') DDNAME(ddname) TYPE(type_name) MOUNTPOINT('pathname')
MODE(access) PARM('parameter') TAG(NOTEXT|TEXT,ccsid) SETUID|NOSETUID SECURITY|NOSECURITY AUTOMOVE[(INCLUDE,sysname1,sysname2,...,sysnamen|*)]
[(EXCLUDE,sysname1,sysname2,...,sysnamen)]
- |NOAUTOMOVE|UNMOUNT SYSNAME(sysname) MKDIR('pathname')
- Specifies a
file system that z/OS UNIX is to logically mount onto
the root file system or another file system.
Mount statements
are processed in the sequence in which they appear. If they are cascading,
the system mounts the first file system first. Make sure that a mount
point exists before the file system is mounted. If you mount a file
system over an existing directory that contains files, you will cover
up the existing files.
If a MOUNT statement uses a DDNAME parameter
to identify the HFS data set, allocate that HFS data set in the OMVS
cataloged procedure. See the section on customizing the OMVS cataloged
procedure to run the kernel initialization program in z/OS UNIX System Services Planning.
The
MOUNT statement is optional; the BPXPRMxx member can contain one or
more MOUNT statements. The MOUNT parameters are:
- AUTOMOVE[(INCLUDE|EXCLUDE,sysname1,sysname2,...,{sysnameN|
*})]|
- NOAUTOMOVE|UNMOUNT
- The AUTOMOVE, NOAUTOMOVE, and UNMOUNT parameters apply only in
a sysplex where systems are participating in shared file system. The
parameters indicate what happens if the system that owns a file system
goes down.
AUTOMOVE indicates that ownership of the file system
automatically changes to another system that is participating in shared
file system. You can specify AUTOMOVE on its own to allow the system
to randomly select a new owner for the file system. You can direct
the system how to choose a new owner for the file system by using
the indicators INCLUDE (I) or EXCLUDE (E).
Specify
INCLUDE with a system list to provide a prioritized list of systems
to which the file system can be moved if the owning system goes down.
For example, AUTOMOVE(INCLUDE,SYS1,SYS4,SYS9) tells the system that
the file system can be moved to SYS1, SYS4, or SYS9, in that order,
or AUTOMOVE(INCLUDE,SYS1,SYS4,*) tells the system that the file system
can be moved to SYS1 or SYS4, in that order, then to any other available
system. This selection is not totally random; rather, MVS attempts to move the file system to a new
server system in which the file system is actively in use.
Restriction: If specified, the asterisk must be listed last
or listed as the only system name in the INCLUDE system list.
Specify EXCLUDE with a system list to provide a list
of systems to which the file system cannot be moved. For example,
AUTOMOVE(EXCLUDE,SYS3,SYS5,SYS7) tells the system that the file system
can be moved to any system except SYS3, SYS5, and SYS7. If the file
system cannot be moved as you directed in the system list, the file
system is unmounted when the owning system goes down.
NOAUTOMOVE
indicates that ownership of the file system is not moved in some situations;
as a result, the file system becomes inaccessible.
Note: When specifying
NOAUTOMOVE, although the file system becomes inaccessible when the
owning system unexpectedly goes down, it still exists in the file
system hierarchy. The file system remains unowned until the original
owning system re-IPLs. Changing the AUTOMOVE value in BPXPRMxx of
an unowned file system before re-IPLing will not change the AUTOMOVE
value of this file system since it is already mounted. To change the
AUTOMOVE value, the file system must be unmounted before the IPL.
UNMOUNT indicates that the file system should be unmounted
in certain situations.
See z/OS UNIX System Services Planning for
more information about the behavior of the AUTOMOVE options.
Note: - Use AUTOMOVE for the version file system and the sysplex root
file system.
- For file systems that are associated with a single system, specify
UNMOUNT. The file systems include /etc,/tmp, /var, /dev,
and the system-specific file system. For descriptions of the sysplex
root, system-specific, and version file systems, see the section on
sharing file systems in a sysplex in z/OS UNIX System Services Planning.
- For sysplex-unaware file systems that are mostly exported by the DFS or SMB server
to their remote clients, consider specifying NOAUTOMOVE on the MOUNT
statement. By doing so, the file systems will not change ownership
if the system is suddenly recycled and they are available for automatic
re-export by DFS or
SMB.
Consider specifying NOAUTOMOVE because a file system can only
be exported by the DFS or
SMB server at the system that owns the file system. After a file system
is exported by DFS,
it cannot be moved until it has been unexported by DFS. The same holds true of file
systems that are exported by SMB. When recovering from system outages,
you need to weigh sysplex availability against availability to the DFS or SMB clients.
When an owning system recycles and a file system that is exported
by DFS or SMB
is taken over by one of the other systems, DFS or SMB cannot automatically re-export
that file system. When an owning system is recycled and an exported
file system is taken over by one of the other systems, that file system
is not automatically reexported. The file system must be moved from
its current owner back to the original system, the one that has just
been recycled, and then exported again.
Default: AUTOMOVE
- DDNAME(ddname)
- The ddname
on the JCL DD statement that defines the file system. To use the DDNAME
parameter, a DD statement for the HFS data set containing the mountable
file system should be placed in the OMVS cataloged procedure. Either
FILESYSTEM or DDNAME is required; do not specify both. The name is
1 to 8 characters; the system converts the ddname to uppercase.
Restriction: zFS does not support DDNAME; the
FILESYSTEM keyword must be used.
- FILESYSTEM('fsname')
- The name
of the file system. The name must be unique in the system.
In the FILESYSTYPE statement, you can mount the following
file systems:
Either FILESYSTEM or DDNAME is required; do not specify
both. The name is 1 to 44 characters; the characters can be in uppercase,
lowercase, or both. The name must be enclosed in single quotation
marks. An HFS data set name must conform to the rules of MVS data set names.
- MKDIR('pathname')
- Specifies the
name of a directory that the system will create dynamically after
the file system has been successfully mounted. This allows the system
to create mount points that subsequent MOUNT statements can reference.
MKDIR is an optional keyword.
You can specify more than one MKDIR
keyword on each ROOT or MOUNT statement. The directories are created
in the order they are listed. Currently, the maximum number of MKDIR
statement that is allowed for each ROOT or MOUNT statement is 50.
You
must specify a relative path name; the path name cannot start with
a slash (/). Enclose the path name in single quotation marks.
The
path name is relative to the file system mount point specified on
the MOUNTPOINT keyword.
The path name can contain intermediate
directories, but these intermediate directories must exist in the
file system hierarchy. If these intermediate directories do not exist,
you can use additional MKDIR keywords to create the necessary intermediate
directories.
The directory to be created must be in a file
system that was mounted as RDWR. The permission bits for the created
directory will be 755. The UID and GID is inherited from this directory's
parent. These attributes are overlaid when this directory is used
as a mount point.
Do not use the MKDIR keyword for file systems
that mount asynchronously, such as the NFS file system. When a file
system is mounted asynchronously, message BPXF025I is issued to the
system log. Similarly, do not use MKDIR with the SYSNAME keyword,
when SYSNAME identifies a remote system to perform the mount. The
results are unpredictable.
A failure to create a directory
does not cause a failure of the mount. A message is written to the
system log when a problem occurs during the creation of the directory.
If the directory exists, no message is written.
- MODE(access)
- Specifies access to the mounted file system by all users:
- READ: Users can only read the file system that is being mounted.
- RDWR: Users can read and write in the file system that is being
mounted.
Default: RDWR
- MOUNTPOINT('pathname')
- Specifies
the absolute path name, or a symbolic link that resolves to the path
name of the directory onto which the file system is to be mounted.
Mount point restrictions are as follows:
- The mount point must be a directory.
- Any files in the directory are not accessible while the file system
is mounted.
- Only one mount can be active at any time for a mount point.
- A file system can be mounted at only one directory at any time.
MOUNTPOINT is required and must specify
an absolute path name. The path name is up to 1023 characters long;
the characters can be in uppercase, lowercase, or both.
- PARM('parameter')
- Provides a parameter
to be passed directly to the file system type. The parameter format
and content are specified by the file system type.
PARM is an
optional parameter. The parameter is up to 500 characters long; the
characters can be in uppercase, lowercase, or both. The parameter
must be enclosed in single quotation marks.
Note: For
information about specifying the TFS configuration parameters, see
the section on parameter key options for the mount statement and mount
command in
z/OS UNIX System Services Planning. For
example:
- -ea count
- Allows the TFS file system to automatically grow count times.
- -em count
- Allows the TFS file system to manually grow count times.
If the physical file system specified
does not expect a PARM operand, it ignores all PARM operands. Refer
to the documentation for the specific physical file system for valid
entry point names.
SYNC(
t), NOWRITEPROTECT,
NOSPARSE, and SYNCRESERVE are valid only when ENTRYPOINT is GFUAINIT.
FSFULL(
threshold,
increment)
is supported by HFS,
TFS, and z/OS File System (zFS).
Note: If a syntax error
is found in any of these parameters (SYNC(t), NOWRITEPROTECT, NOSPARSE,
FSFULL, and SYNCRESERVE), an error message is issued and all five
parameters are set to the default values.
- SYNC(t)
- t specifies the number of seconds used as a
default for the sync daemon interval. When the sync daemon is active,
the metadata for a file system is hardened. Setting t to 0
indicates that the file system should harden metadata synchronously
with syscall requests.
- Sync interval values are rounded up to the next 30-second value.
For example, specifying 31 seconds results in a sync interval of 60
seconds.
- The maximum value that can be specified for t is
65535. Values between 65535 and 99999 are rejected.
- A value of 99999 specifies that no sync daemon intervals are specified,
and thus, the metadata is not hardened.
- Default: 60 seconds
- NOWRITEPROTECT
- This keyword overrides the WRITEPROTECT function. When NOWRITEPROTECT
is specified, the file system is not protected from being read/write
mounted by multiple systems simultaneously. Read/write mounting by
multiple systems corrupts the file system.
Extreme care should
be taken when specifying this keyword. It should only be used when
there is no possibility of the file system that is being mounted by
multiple systems.
Use of the NOWRITEPROTECT keyword avoids
an additional file system read operation that is required at Sync
time to support the WRITEPROTECT function.
- Default: WRITEPROTECT
- NOSPARSE(DUMP|LOGREC)
- When NOSPARSE is specified on the MOUNT statement, HFS will not
allow any files in that file system to be sparse. A file becomes
sparse when all of the data cannot be written. For example, suppose
we are only able to write the first 10,000 bytes of a file, and then
the system has to lseek out to offset 50,000 and resume writing from
there. The file is considered sparse because bytes 10,000-50,000
were never written to the file. If the user attempts to read bytes
10,000 to 50,000, binary 0's will be returned as the value. NOSPARSE
handles this by causing a dump to be taken or LOGREC record to be
created when either of the following situations occur:
- The file system attempts to read metadata from disk for a file
and detects that the subject file is sparse, or
- An application attempts to write to a page beyond the end of the
file, causing the file to become sparse.
- DUMP causes a dump to be created. Only one dump is created for
each of the possible reason codes while a file system is mounted.
DUMP is the default if you specify NOSPARSE without the DUMP or LOGREC
keywords.
- LOGREC causes a LOGREC record to be written. A dump is not created.
- Default: DUMP
- FSFULL(threshold,increment)
- threshold specifies the percentage of the file
system (HFS, TFS, or zFS) capacity at which
an operator message is generated. The default is 100%.
- increment specifies the percentage of change
above the HFS or zFS capacity at which an operator message is generated.
Messages are generated by either an increase or decrease greater than increment .
The default is 5%.
You can specify threshold and increment values
for all file systems. The values can also be set on the MOUNT command
for a specific file system. Parameters on the MOUNT command override
parmlib values. If no values are specified in either place, no threshold
checking is done. If a threshold value is specified but no increment
is given, the increment defaults to 5%. The increment value applies
both to upgrading the message when the file system continues to fill
and to removing the message when more space becomes available due
to either deleting files, or to extending the file system. The values
are in terms of percent full. The values that are applied to a file
system can be changed only when the file system is mounted.
- SYNCRESERVE(nn)
- This keyword controls the number of pages to be reserved for sync
processing of the file system metadata.nn represents
the percentage of the file system space that is to be reserved for
the sync shadow write mechanism. nn is a decimal
number between 1 and 50. There is no reason to reserve more than 50%
of the file system space, because the reserved space must always be
less than the actual index size and the index size plus the reserved
space cannot be greater than the file system space.
- When this parameter is specified on the MOUNT statement, it overrides
the internal reserved page estimation algorithm. Only use this parameter
if the internal algorithm is not providing the desired results.
- SYSNAME(sysname)
- For systems that are participating in a shared file system, SYSNAME
specifies the particular system on which a mount should be performed.
This system will then become the owner of the file system mounted.
This system must be IPLed with SYSPLEX(YES).
Default: The
name of the system, if IPLed with SYSPLEX(YES), that the mount is
processed on.
Note: - Use the defaults for SYSNAME and AUTOMOVE to ensure that the root
is always available.
- Only specify a SYSNAME() value if you want only the specified
system to be the file system owner.
- During z/OS UNIX initialization processing the MOUNT statement
is ignored if SYSNAME() specifies another system.
- For SET OMVS and SETOMVS processing, the MOUNT statement is processed
and the MOUNT is function-shipped to the system specified by SYSNAME().
If SYSNAME() is used with a value that resolves to another system,
do not include any subsequent parmlib MOUNT statements that specify
a MOUNTPOINT() with a path name that includes a directory in this
file system.
- TAG (NOTEXT|TEXT,ccsid)
- Specifies whether
implicit file tags are assigned to untagged files in the mounted file
system. File tagging controls whether a file's data can be converted
during file reading and writing. "Implicit" in this case means that
the tag is not permanently stored with the file. Instead, the tag
is associated with the file during reading and writing, or when stat()
type functions are issued. Either TEXT, or NOTEXT, and ccsid must
be specified when TAG is specified.
NOTEXT specifies that none of
the files in the file system are automatically converted during file
reading and writing.
TEXT specifies that each untagged file
is implicitly marked as containing pure text data that can be converted.
ccsid names
the coded character set identifier to be implicitly set for the untagged
file. ccsid is specified as a decimal value from
0 to 65536. However, when TEXT is specified, the values of 0 and 65536
are illegal because those values imply no conversion. Other than this,
the value is not checked as being valid and the corresponding code
page is not checked as being installed.
For example:
- TAG(TEXT,819) identifies text files that contain ASCII (ISO-8859-1)
data.
- TAG(TEXT,1047) identifies text files that contain EBCDIC ((ISO-1047)
data.
- TAG(NOTEXT,65536) tags files as containing binary or unknown data.
- TAG(NOTEXT,0) is the equivalent of not specifying the TAG parameter.
- TAG(NOTEXT,273) tags file with the German code set
(ISO-273) but is not eligible for automatic conversion.
Default: NOTEXT
- SECURITY|NOSECURITY
- SECURITY
specifies that security checks should be performed. NOSECURITY specifies that security
checks should not be performed.
Default: SECURITY
- SETUID|NOSETUID
- SETUID specifies
that the setuid() and setgid() mode bit on an executable file will
be supported.
NOSETUID
specifies that the setuid() and setgid() mode bit on an executable
file will not be supported. The UID or GID will not be changed when
the program is executed and the APF and Program Control extended attributes
are not honored. The entire HFS is uncontrolled.
Default:
SETUID
- TYPE(type_name)
- Specifies the
name of a file system type that is identified in a FILESYSTYPE statement.
The TYPE(type_name) parameter must be the same as the TYPE(type_name)
parameter on a FILESYSTYPE statement. TYPE is a required parameter.
The name is 1 to 8 characters; the system converts the name to uppercase.
For additional information, see the MOUNT section
in z/OS UNIX System Services Planning.
- NETWORK DOMAINNAME(sockets_domain_name) DOMAINNUMBER(sockets_domain_number)
MAXSOCKETS(number) TYPE(type_name) INADDRANYPORT(starting_port_number)
INADDRANYCOUNT(number_of_ports_to_reserve)
- Specifies
that a socket physical file system domain should be readied for use. The TYPE in this statement matches the TYPE on the previous
FILESYSTYPE statement.
Use the SETOMVS RESET command to dynamically
change the MAXSOCKET value or add a new NETWORK. To make a permanent
change, edit the BPXPRMxx member that is used for IPLs. For more
information, see the section on dynamically adding FILESYSTYPE statements
in BPXPRMxx in z/OS UNIX System Services Planning.
Provide
a NETWORK statement for each socket file system domain to be initialized.
- For AF_UNIX file systems, always include a FILESYSTYPE statement
that specifies ENTRYPOINT(BPXTUINT) and a NETWORK statement with
a matching TYPE, usually TYPE(UDS), on both.
- For TCP/IP sockets, always include a FILESYSTYPE statement that
specifies ENTRYPOINT(EZBPFINI) and a NETWORK statement with a matching
TYPE, usually TYPE(INET), on both.
- To activate an Internet Protocol Version 6 (IPv6) socket on a
system, you must configure both the AF_INET domain and the AF_INET6
domain. You cannot code a NETWORK statement for domain name AF_INET6
without coding a NETWORK statement for domain name AF_INET.
- For CINET sockets, include a FILESYSTYPE statement with ENTRYPOINT
(BPXTCINT) and a NETWORK statement with a matching TYPE, usually TYPE(CINET),
that specifies INADDRANYPORT and INADDRANYCOUNT. See the section on
specifying INADDRANYPORT and INADDRANYCOUNT in z/OS UNIX System Services Planning for
more information.
- DOMAINNAME(sockets_domain_name)
- The 1 to
16 character name by which this socket file system domain is to be
known.
- DOMAINNUMBER(sockets_domain_number)
- A number
that matches the value that is defined for this domain name. The
currently supported values for this field are:
- 1
- AF_UNIX
- 2
- AF_INET
- 19
- AF_INET6
Table 1 shows some supported
domain names, domain numbers, and their associated entry point names.
See the documentation for the physical file system you are using to
get the correct entry point name.
Table 1. Supported
domainsDomain name |
Domain number |
Entry point |
AF_UNIX |
1 |
BPXTUINT |
AF_INET |
2 |
EZBPFINI, BPXTCINT |
AF_INET6 |
19 |
EZBPFINI, BPXTCINT |
- INADDRANYCOUNT(number_of_ports_to_reserve)
- Specifies
the number of ports that the system reserves, starting with the port
number specified in the INADDRANYPORT parameter. This value is only
needed for CINET.
Value Range: number_of_ports_to_reserve is
a decimal value from 1 to 8000.
Default: If neither
INADDRANYPORT or INADDRANYCOUNT is specified, the default for INADDRANYCOUNT
is 1000. Otherwise, no ports are reserved (0).
- INADDRANYPORT(starting_port_number)
- Specifies
the starting port number for the range of port numbers that the system
reserves for use with PORT 0, INADDR_ANY binds. This value is
only needed for CINET.
Value Range: starting_port_number is
a decimal value from 1024 to 65534. Ports 1 — 1023 are well-known
ports that cannot be reserved for use with PORT 0, INADDR_ANY binds.
Default:
If neither INADDRANYPORT or INADDRANYCOUNT is specified, the default
for INADDRANYPORT is 63000. Otherwise, no ports are reserved (0).
Note: If
you do not want to support INADDRANY with CINET, you should specify
INADDRANYPORT(xx), where xx is
a valid value, without specifying INADDRANYCOUNT.
Note: When activating IPv6 on a system, the INADDRANYPORT
is shared across domains. The INADDRANYPORT value is taken from the
NETWORK statement for the AF_INET domain. Any INADDRANYPORT value
specified for the AF_INET6 domain is ignored.
- MAXSOCKETS(nnnnn)
- Specifies
the maximum number of sockets supported by this file system for this
address family. You can specify a value from 0 to 16777215. This is
an optional parameter. The maximum value that this field can have
is defined by each domain. If a value larger than the maximum is
specified, an informational message is issued and the value used is
the maximum.
Note: Ensure that this number is large enough for socket
connections for all applications using your
z/OS UNIX environment. This upper limit
is set when the NETWORK statement is processed during IPL. It can
only be changed if the NETWORK statement is changed using the SETOMVS
RESET command.
When the Common Event Adapter (CEA) is active, each
CEA client requires two socket connections for communication with
the CEA server.
For AF_UNIX, MAXSOCKETS will be ignored
if it is specified. The maximum number of AF_UNIX sockets is 10000.
When
activating IPv6 on a system, you can specify separate MAXSOCKETS values
for domains AF_INET and AF_INET6. If you do not specify a MAXSOCKETS
value for the AF_INET6 domain, the default will be the MAXSOCKETS
value specified or defaulted to for the AF_INET domain.
Table 2 shows the maximum and
default values for MAXSOCKETS by domain.
Table 2. Maximum and default values for MAXSOCKETS
by domainDomain name |
MAXSOCKETS maximum |
MAXSOCKETS default |
AF_UNIX |
10000 |
10000 |
AF_INET |
16777215 |
64000 |
AF_INET6 |
16777215 |
Same as the current value for AF_INET |
- TYPE(type_name)
- Specifies the
name of a file system type that is identified in a FILESYSTYPE statement.
The TYPE(type_name) must be the same as the TYPE(type_name) parameter
on a FILESYSTYPE statement. TYPE is a required parameter. The name
is 1 to 8 characters; the system converts the name to uppercase.
- NONEMPTYMOUNTPT(NOWARN|WARN|DENY)
- Specifies how the system is to mount any file system on a mount
point when it is a non-empty directory.
NOWARN causes the system
to mount any file system on mount point without any warning message
when the mount point is a non-empty directory. The contents of that
directory are hidden for the duration of the mount.
WARN causes
the system to mount any file system on mount point with a warning
message when the mount point is a non-empty directory. The contents
of that directory are hidden during the mount.
DENY causes
the system not to mount any file system when the mount point is a
non-empty directory.
Default: NOWARN
If NONEMPTYMOUNTPT
is not specified, default values are used.
For planning information,
see NONEMTPYMOUNTPT in z/OS UNIX System Services Planning.
- PRIORITYGOAL(service_class_name1,...service_class_name40)
- Specifies a list of 1 to 40 service class names of 8 characters
or less separated by commas, which are used in association with the
setpriority, nice and chpriority callable services when the system
is running in goal mode. These functions allow a program to alter
the priority of one or more processes.
Generally, it is recommended that you not set PRIORITYGOAL
unless the nice(), setpriority() or chpriority() values is enabled.
If
the list has less than 40 entries, the system propagates the last
service class specified into the remaining unspecified entries in
the table. For example:
PRIORITYGOAL(CICS4,CICS4,CICS4,CICS3,CICS2,CICS1,TSO2,TSO1,BAT3,BAT2)
If
you do not specify this statement, arrays are not created for it.
All service classes that are specified on the PRIORITYGOAL statement
must also be specified in your workload manager service policy.
PRIORITYGOAL(NONE)
means that there are no values. If you do not specify PRIORITYGOAL,
that means that there are no values.
If you do not want to
allow users to increase the priority but still want to enable the nice() and setpriority() functions,
define a range of service classes with priority increments on a base
that is normal for the users. Using these functions lets the user
order the priority of processes, but will not let a user improve performance
over that of other users.
Value Range: service_class_name is
a 1 to 8 character value.
Default: None
You can
dynamically change the values of PRIORITYGOAL by using the SETOMVS
or SET OMVS command. To make a permanent change, edit the BPXPRMxx
member that is used for IPLs.
- PWT(SMF|SMFENV|ENV)
- Allows installations
to time out z/OS UNIX processes that are waiting on terminal
activity. When specified, the timeout value applies to all z/OS UNIX processes
that are waiting on terminal activity. To override that value for
a specific process, the _BPXK_TIMEOUT environment value can be set
for an individual process. The PWT value can be changed with the SETOMVS
PWT operator command. Unless a BPXK_TIMEOUT value is set, timeouts
do not occur.
Specify SMF to honor the SMPFPRMxx JWT|TWT|SWT values
for the timeout value. The _BPXK_TIMEOUT environment value is ignored.
An individual process cannot override the system settings.
Specify
SMFENV to honor the SMFPRMxx JWT|TWT|SWT value and to allow the _BPXK_TIMEOUT
settings to override the setting.
Specify ENV to allow the _BPXK_TIMEOUT
variable be set. Only processes setting the _BPXK_TIMEOUT value are
timed out.
Default: ENV
- RESOLVER_PROC(procname|DEFAULT|NONE)
- Specifies how
the resolver address space is processed during z/OS UNIX initialization. The resolver
is used by TCP/IP applications for name-to-address or address-to-name
resolution. In order to create a resolver address space, a system
must be configured with an AF_INET or AF_INET6 domain.
procname is
the name of the address space for the resolver and the procedure member
name in the appropriate proclib. procname is one
to eight characters long. The procedure must reside in a data set
that is specified by the MSTJCLxx parmlib member's IEFPDSI DD card
specification.
DEFAULT causes an address space named RESOLVER
to start, using the system default procedure of IEESYSAS. The address
space is started with SUB=MSTR so that it runs under the MASTER address
space instead of the JES address space.
NONE specifies that
no address space is to be started. If you are using z/OS Communications Server IP, the resolver
must be started before TCP/IP can be started. TCP/IP does not initialize
until the resolver address space is started.
- RUNOPTS('string')
- Specifies
the _CEE_RUNOPTS environment variable that is used when z/OS UNIX initialization invokes /etc/init or /usr/sbin/init. z/OS UNIX passes the _CEE_RUNOPTS value
and all programs that are invoked from /etc/rc to
the shell.
If you want to change the value of RUNOPTS, you will
must edit the BPXPRMxx member and then re-IPL. You cannot use the
SET OMVS or SETOMVS command to change the value. After the value is
specified in BPXPRMxx, you can use one of the following methods to
change this string:
- The system is re-IPLed with a new BPXPRMxx RUNOPTS string.
- The user or installation sets _CEE_RUNOPTS in /etc/rc or /etc/init.config.
- A program or shell script sets _CEE_RUNOPTS.
If you do not specify a value for RUNOPTS, the RUNOPTS
string or _CEE_RUNOPTS environment variable is not provided.
The
TSO/E OMVS command uses the specified options as the Language Environment® runtime
options, by default.
The setting of RUNOPTS has no effect on
BPXBATCH jobs.
Specifying the RUNOPTS parameter causes the kernel
to set the _CEE_RUNOPTS environment variable when
starting /etc/init, or when the TSO/E OMVS command
is entered. This environment variable is normally propagated to subsequent
processes (such as /etc/init to /bin/sh to /etc/rc to /bin/inetd to /bin/rlogind to
/bin/sh for shell users).
To do this,
you must make sure that any other steps in the flow (such as export
statements in /etc/rc) do not overwrite the value
of _CEE_RUNOPTS. If more runtime options are needed, they should be
concatenated to the old value of _CEE_RUNOPTS.
Value Range: From
1 to 250 characters.
Default: No RUNOPTS string or _CEE_RUNOPTS
environment variable is provided.
Restrictions:
- The string must be enclosed in parentheses and quotation marks
('').
- An empty string (' ') is not valid.
- Although all characters are allowed, nulls, slashes (/), unbalanced
SO/SI, and unbalanced parentheses and quotation marks cause unpredictable
problems in areas such as the TSO/E OMVS command.
For more information about specifying RUNOPTS strings,
see Customizing the BPXPRMxx parmlib member in z/OS UNIX System Services Planning.
- ROOT FILESYSTEM('fsname') DDNAME(ddname) TYPE(type_name) MODE(access)
PARM('parameter') SETUID|NOSETUID AUTOMOVE|NOAUTOMOVE
SYSNAME(sysname) TAG(NOTEXT|TEXT,ccsid)
MKDIR('pathname')
- Specifies a
file system that z/OS UNIX is to logically mount as the
root file system. It is optional. If it is not specified, a TFS file
system is mounted as the root.
To change the value of the ROOT
statement without having to re-IPL, use the TSO/E MOUNT and UNMOUNT
commands.
The root file system can be unmounted using the TSO/E
UNMOUNT command or ISHELL. Ensure that you specify the IMMEDIATE option.
The
ROOT statement does not support the SYSNAME() keyword.
The parameters
are as follows:
- DDNAME(ddname)
- The ddname
on the JCL DD statement that defines the root file system. To use
the DDNAME parameter, a DD statement for the HFS data set containing
the root file system should be placed in the z/OS UNIX
cataloged procedure.
Either FILESYSTEM or DDNAME is required;
do not specify both. The ddname is 1 to 8 characters; the system converts
the ddname to uppercase.
Restriction: zFS does not support DDNAME; the
FILESYSTEM keyword must be used.
- FILESYSTEM('fsname')
- The name
of the root file system. The name must be unique in the system.
You can specify the following types of the file system:
- HFS for a hierarchical file system (HFS).
- zFS for a z/OS File System
(zFS).
- NFS for accessing remote files.
- TFS for a temporary file system (TFS).
Either FILESYSTEM or DDNAME is required; do not specify
both. The name is 1 to 44 characters; the characters can be in uppercase,
lowercase, or both. The name must be enclosed in single quotation
marks. An HFS data set name must conform to the rules of MVS data set names.
- MKDIR('pathname')
- Specifies the
name of a directory that the system will create dynamically after
the file system is successfully mounted. This allows the system to
create mount points that can be referenced by subsequent MOUNT statements.
MKDIR is an optional keyword.
You can specify more than one MKDIR
keyword on each ROOT statement. The directories are created in the
order they are listed.
You must specify a relative path name;
the path name cannot start with a slash (/). Enclose the path name
in single quotation marks.
The path name is relative to the
file system mount point specified on the MOUNTPOINT keyword.
The
path name can contain intermediate directories, but these intermediate
directories must exist in the file system hierarchy. If these intermediate
directories do not exist, you can use additional MKDIR keywords to
create the necessary intermediate directories.
The directory
to be created must reside in a file system that was mounted as RDWR.
The permission bits for the created directory is 755. The UID and
GID is inherited from this directory's parent. These attributes will
be overlaid when this directory is used as a mount point.
Do
not use the MKDIR keyword for file systems that mount asynchronously,
such as the NFS file system. When a file system is mounted asynchronously,
message BPXF025I is issued to the system log. Also, do not use MKDIR
with the SYSNAME keyword, when SYSNAME identifies a remote system
to perform the mount. The results are unpredictable.
A failure
to create a directory does not cause a failure of the mount. A message
is written to the system log when a problem occurs when the directory
is created. If the directory exists, no message is written.
Restriction: MKDIR works only on systems that are at z/OS Version 1 Release 5 level
or later. If you are using MKDIR in a sysplex that shares a common
BPXPRMxx member, make sure that all systems are at least at the z/OS V1R5 level.
- MODE(access)
- Specifies access
to the root file system by all users:
- READ: Users can only read the root file system.
- RDWR: Users can read and write in the root file system.
Default: RDWR
- PARM('parameter')
- Provides a parameter
to be passed directly to the file system type. The parameter format
and content are specified by the file system type.
PARM is an
optional parameter. The parameter is up to 500 characters long; the
characters can be in uppercase, lowercase, or both. The parameter
must be enclosed in single quotation marks.
If the physical
file system specified does not expect a PARM operand, it ignores all
PARM operands. Refer to the documentation for the specific physical
file system for valid entry point names.
SYNC(
t),
NOWRITEPROTECT, NOSPARSE, and SYNCRESERVE are valid only when ENTRYPOINT
is GFUAINIT. FSFULL(
threshold,
increment)
is supported by HFS,
TFS, and zFS.
Note: If
a syntax error is found in any of these parameters (SYNC(t,
NOWRITEPROTECT, NOSPARSE, FSFULL, and SYNCRESERVE), an error message
is issued and all five parameters are set to the default values.
- SYNC(t)
- t specifies the number of seconds used as a
default for the sync daemon interval. When the sync daemon is active,
the metadata for a file system is hardened. Setting t to
0 indicates that the file system should harden metadata synchronously
with syscall requests.
- Sync interval values are rounded up to the next 30-second value.
For example, specifying 31 seconds results in a sync interval of 60
seconds.
- The maximum value that can be specified for t is 65535.
Values between 65535 and 99999 are rejected.
- A value of 99999 specifies that no sync daemon intervals are specified,
and thus, the metadata is not hardened.
- Default: 60 seconds
- NOWRITEPROTECT
- This keyword overrides the WRITEPROTECT function. When NOWRITEPROTECT
is specified, the file system is not protected from being read/write
mounted by multiple systems simultaneously. Read/write mounting by
multiple systems corrupts the file system.
Extreme care should
be taken when specifying this keyword. Use it only when there is no
possibility of the file system being mounted by multiple systems.
Use
of the NOWRITEPROTECT keyword avoids an additional file system read
operation that is required at Sync time to support the WRITEPROTECT
function.
- Default: WRITEPROTECT
- NOSPARSE(DUMP|LOGREC)
- When NOSPARSE is specified on the MOUNT statement, files are not
allow to be sparse. A file becomes sparse when all of the data cannot
be written. For example, suppose we are only able to write the first
10,000 bytes of a file, and then the system has to lseek out to offset
50,000 and resume writing from there. The file is considered sparse
because bytes 10,000-50,000 were never written to the file. If the
user attempts to read bytes 10,000 to 50,000, binary 0’s are
returned as the value. NOSPARSE handles this by causing the file system
to create a dump or a LOGREC record when either of the following situations
occur:
- The file system attempts to read metadata from disk for a file
and detects that the subject file is sparse.
- An application attempts to write to a page beyond the end of the
file, causing the file to become sparse.
- DUMP causes the file system to create a dump. Only one dump is
created for each of the possible reason codes while a file system
is mounted. DUMP is the default if you specify NOSPARSE without the
DUMP or LOGREC keywords.
- LOGREC will cause the file system to write a LOGREC record instead
of creating a dump.
- Default: DUMP
- FSFULL(threshold,increment)
- threshold specifies the percentage of the file
system (HFS, TFS, or zFS) capacity at which
an operator message is generated. The default is 100%.
- increment specifies the percentage of change
above the file system capacity at which an operator message is generated.
Messages are generated by either an increase or decrease greater than increment.
The default is 5%.
You can specify threshold and increment values
for all file systems. The values can also be set on the MOUNT command
for a specific file system. Parameters on the MOUNT command override
parmlib values. If no values are specified in either place, no threshold
checking is done. If a threshold value is specified but no increment
is given, the increment defaults to 5%. The increment value applies
both to upgrading the message when the file system continues to fill
and to removing the message when more space becomes available due
to either deleting files, or to extending the file system. The values
are in terms of percent full. The values that are applied to a file
system can be changed only when the file system is mounted.
- SYNCRESERVE(nn)
- This keyword controls the number of pages to be reserved for sync
processing of the file system metadata.nn represents
the percentage of the file system space which is to be reserved for
the sync shadow write mechanism. nn is a decimal
number between 1 and 50. There is no reason to ever reserve more than
50% of the file system space, because the reserved space must always
be less than the actual index size and the index size plus the reserved
space cannot be greater than the file system space.
- When this parameter is specified on the MOUNT statement, it overrides
the internal reserved page estimation algorithm. Only use it if the
internal algorithm does not provide the expected results.
- SERV_LINKLIB('dsname', 'volser')
- Specifies the target service library where the UNIX System Services modules that are normally
loaded from SYS1.LINKLIB into the private area of the OMVS address
space are located.
Value Range: dsname is
a 1-to-44 character value that represents a valid MVS load library data set name. The alphabetic
characters in the load library name must be uppercase. volser is
a 1-to-6 character value that represents a valid volume serial number
for the volume that contains the specified MVS load library. The alphabetic characters in
the volume serial number must be uppercase.
You can change
the value of SERV_LINKLIB dynamically using the SETOMVS or SET OMVS
command. To make a permanent change, edit the BPXPRMxx member that
will be used for future IPLs.
- SERV_LPALIB('dsname', 'volser')
- Specifies the target service library where the UNIX System Services modules that are normally
built into LPA are located.
Value Range: dsname is
a 1-to-44 character value that represents a valid MVS load library data set name. The alphabetic
characters in the load library name must be uppercase. volser is
a 1-to-6 character value that represents a valid volume serial number
for the volume that contains the specified MVS load library. The alphabetic characters in
the volume serial number must be uppercase.
You can change
the value of SERV_LPALIB dynamically using the SETOMVS or SET OMVS
command. To make a permanent change, edit the BPXPRMxx member that
will be used for future IPLs.
- SETUID|NOSETUID
- SETUID specifies
that the setuid() and setgid() mode bit on an executable file will
be supported.
NOSETUID
specifies that the setuid() and setgid() mode bit on an executable
file will not be supported. The UID or GID will not be changed when
the program is executed and the APF and program control extended attributes
are not honored. The entire file system is uncontrolled.
Default:
SETUID
- SHRLIBMAXPAGES(nnnnn)
- If
you specify the SHRLIBMAXPAGES parameter, it is accepted but will
not have any impact on the system. The value that you specify will
never be reached, because user-shared library objects are no longer
supported. This parameter is intended to control the maximum number
of pages that can be allocated in the system to contain user shared
library modules. This value, used with MAXSHAREPAGES, can be used
to control the amount of ESQA consumed by user shared library modules.
Refer to z/OS UNIX System Services Planning for
more details.
Value Range: nnnnn is
a decimal value between 1 and 16777216 specifying
a number of 4K pages.
Default: 4096
- SHRLIBRGNSIZE(nnnnn)
- Specifies
the maximum size of the shared library region for address spaces that
load system shared library modules. For these address spaces, the
size that is specified is allocated from high private storage and
is used for the loading of system shared library modules. This storage
is not allocated in an address space until it loads a system shared
library module. This parameter applies to modules loaded from system
shared libraries, which allocate storage on megabyte boundaries.
Therefore, this storage does not count against the MAXSHAREPAGES limit,
and does not consume ESQA.
Value Range: nnnnn is
a decimal value between 16777216 (16 megabytes)
and 1610612736 (1.5 gigabytes).
Default: 67108864
Use the SETOMVS or SET
OMVS command to dynamically increase or decrease the SHRLIBRGNSIZE
value. To make a permanent change, edit the BPXPRMxx member that is
used for IPLs.
Because the shared library region is allocated in
megabytes, the value for SHRLIBRGNSIZE must be evenly divisible by
1048576. Otherwise, the D OMVS,L command might show values in the
Current Usage and Highwater Usage columns that are larger than the
value in the System Limit column.
- STARTUP_EXEC
- STARTUP_EXEC
names a REXX exec that does application environment initialization
for z/OS UNIX. This statement is optional;
if it is specified, the BPXOINIT process will not run /etc/init.
The startup exec is typically used by an installation that does not
have an HFS, but is using a TFS for a file system. It can be used
to populate the TFS with any directories and files that are needed.
It is specified as:
STARTUP_EXEC('Dsname(Memname)',SysoutClass)
where:
- Dsname is a 1-to-44-character valid data set name.
- Memname is a 1-to-8-character valid REXX exec member.
- SysoutClass is 1 character and is alphanumeric and specifies the
sysout class that the REXX exec will run under. Specifying SysoutClass
is optional.
If you want to change the value of STARTUP_EXEC, you will
have to edit the BPXPRMxx member and then reIPL. You cannot use the
SET OMVS or SETOMVS command to change the value.
Default:
There is no default value for STARTUP_EXEC.
- STARTUP_PROC
- This
statement specifies a 1-to-8-character name of a started JCL procedure
that initializes the kernel. The name specified in this statement
must exist on the system before IPL or errors will occur.
Using
a started procedure other than OMVS is strongly discouraged. If you
want to change the value of STARTUP_PROC, you must edit the BPXPRMxx
member and then re-IPL. You cannot use the SET OMVS or SETOMVS command
to change the value.
If you decide to use a started procedure
other than OMVS:
- The replacement started procedure must also be a single job step
procedure that invokes the BPXINIT program (EXEC PGM=BPXINIT). If
it invokes any other program, the OMVS initialization will fail.
- Change the procedure name in the RACF® started
procedures table or the definitions in the STARTED Class. See Preparing for RACF in z/OS UNIX System Services Planning.
Note: Renaming OMVS to another value might affect the setup
of other products such as TCP/IP.
Default: STARTUP_PROC(OMVS).
- SUBFILESYSTYPE NAME(transport_name) TYPE(type_name) ENTRYPOINT(entry_name)
PARM('parameter') DEFAULT
- Specifies
an AF_INET or AF_INET6 physical file system that is to run underneath
the CINET socket file system. The TYPE() value is usually CINET and
matches the TYPE operand on a previous FILESYSTYPE and NETWORK statement.
In the case of TCP/IP, the NAME() value is the procname. The system
attaches the EZBPFINI load module during initialization, and this
file system should be used as the default INET physical file system.
The SUBFILESYSTYPE
statement is associated with its corresponding FILESYSTYPE and NETWORK
statements by matching the value that is specified in the TYPE operand.
The value that is specified on all of the TYPE operands
must match, but can be any 1- to 8-character value. The value that
is specified on the NAME parameter on the SUBFILESYSTYPE statement
is the name to be used by the physical file system when it is initialized.
The first character of the NAME parameter must be non-numeric.
For SecureWay Communications
Server, the SUBFILESYSTYPE statement must match the TCPIPJOBNAME of
that stack. For more information about defining file systems, see
the section on customizing the BPXPRMxx member in z/OS UNIX System Services Planning.
New
SUBFILESYSTYPE statements can be added dynamically. However, you cannot
dynamically change (or delete) a value. For more information, see
the section on dynamically adding FILESYSTYPE statements in z/OS UNIX System Services Planning.
The
parameters are as follows:
- DEFAULT
- Identifies
this file system as the default CINET file system. DEFAULT is an optional
parameter. If it is not specified, the file system that is specified
in the first SUBFILESYSTYPE statement found in the parmlib member
is designated as the default. See the section on setting up for CINET
AF_INET sockets in z/OS UNIX System Services Planning for
more information about the use of the DEFAULT parameter.
- ENTRYPOINT(entry_name)
- Specifies
the name of the load module that contains the entry point into the
file system type. ENTRYPOINT is a required parameter. The name is
1 to 8 characters; the system converts the name to uppercase.
- NAME(transport_name)
- Specifies the
name that identifies this file system to the CINET physical file system.
NAME is a required parameter. The name is 1 to 8
characters with the first character non-numeric; the system converts
the name to uppercase. The value specified by the NAME parameter
on the SUBFILESYSTYPE statement is the name that the physical file
system uses to identify itself when it is initialized. For example,
for TCP/IP, this is the starting procedure name.
- PARM('parameter')
- Provides a parameter
to be passed to the transport driver. The parameter format and content
are specified by the file system that is receiving the data.
PARM
is an optional parameter. The parameter is up to 1024 characters
long; the characters can be in uppercase, lowercase, or both. If
the characters are not all in uppercase, the parameter must be enclosed
in single quotation marks.
If the physical file system specified
does not expect a PARM operand, it ignores all PARM operands. Refer
to the documentation for the specific physical file system for valid
entry point names.
- TYPE(type_name)
- Specifies the
name of the CINET file system type that is identified in a FILESYSTYPE
statement. The TYPE(type_name) parameter must be the same name that
was used for the TYPE(type_name) parameter on the FILESYSTYPE statement
for the CINET physical file system. TYPE is a required parameter.
The name is 1 to 8 characters; the system converts the name to uppercase.
For more information, see SUBFILESYSTYPE in z/OS UNIX System Services Planning.
- SUPERUSER(user_name)
- Superuser
name, which must conform to the restrictions for the z/OS user
ID. The user name must also be defined to RACF (or another security product) and must
have a z/OS UNIX user ID (UID) of 0. For example,
in RACF, specify OMVS(UID(0))
on the ADDUSER command.
When a daemon issues a setuid() to set
a UID to 0 and the user ID is not known, setuid() uses the user ID
from the SUPERUSER statement.
Never permit the user ID BPXROOT
to the BPX.DAEMON profile (described in the section on setting up
the BPX.* FACILITY class profiles in z/OS UNIX System Services Planning).
This warning applies even if you use a name other than BPXROOT.
Value
Range: user_name is a 1 to 8 character value.
Default:
BPXROOT
Use the SETOMVS or SET OMVS command to dynamically
change the value of SUPERUSER. To make a permanent change, edit the
BPXPRMxx member that is used for IPLs.
- SWA(ABOVE|BELOW)
- Specifies whether
SWA control blocks should be allocated above or below the 16-megabyte
line. The SWA parameter only affects the OMVS address space and its
SWA blocks. It has no impact on the colony address spaces.
- ABOVE
- All SWA control blocks are to be allocated above the 16-megabyte
line.
- BELOW
- All SWA control blocks are to be allocated below the 16-megabyte
line.
Default: BELOW
- SYSCALL_COUNTS(YES | NO)
- Specifies
that syscall counts are to be accumulated in internal kernel data
areas so that the RMF™ data gatherer
can record the information.
If you specify YES, the path length
for the most frequently used z/OS UNIX system calls is increased by
more than 150 instructions. This setting will also cause the reporting
of CPU time for z/OS UNIX to be more accurate. This
is reflected in the output from the BPX1TIM, BPX1GPS, BPX1GTH, and
BPX1RMG services and from BPXESMF.
If
you specify NO (or default to NO) the counts for syscall and CPU time
are not accumulated.
If you switch between YES and NO for
actively running processes, the accumulated data is inaccurate.
The
SMF30OST field of the SMF type 30 record can be seen as a negative
number when the CPU time used to compute the time value is less than
the previous CPU time and in some cases can even appear as zero (for
example, when SYSCALL_COUNTS is set to NO after having been YES for
an earlier part of the run.) In addition, if SYSCALL_COUNTS=YES,
the SMF30OST value is calculated from OCVTSYSTIME (accumulated syscall
time for all syscalls issued in the system) rather than from OASBSYSCALLCOUNT
(syscalls for the current process).
Default: NO
Use
the SETOMVS or SET OMVS command to dynamically change the value of
SYSCALL_COUNT. To make a permanent change, edit the BPXPRMxx member
that is used for IPLs.
- TAG (NOTEXT|TEXT,ccsid)
- Specifies whether
implicit file tags are assigned to untagged files in the mounted file
system. File tagging controls whether a file's data can be converted
during file reading and writing. “Implicit” in this case
means that the tag is not permanently stored with the file. Instead,
the tag is associated with the file during reading and writing, or
when stat() type functions are issued. Either TEXT,
or NOTEXT, and ccsid must be
specified when TAG is specified.
NOTEXT specifies
that none of the files in the file system are automatically converted
during file reading and writing.
TEXT specifies
that each untagged file is implicitly marked as containing pure text
data that can be converted.
ccsid names
the coded character set identifier to be implicitly set for the untagged
file. ccsid is specified as a decimal value from
0 to 65536. However, when TEXT is specified, the values of 0 and 65536
are illegal because those values imply no conversion. Other than this,
the value is not checked as being valid and the corresponding code
page is not checked as being installed.
For example:
- TAG(TEXT,819) identifies text files that contain ASCII (ISO-8859-1)
data.
- TAG(TEXT,1047) identifies text files that contain EBCDIC ((ISO-1047)
data.
- TAG(NOTEXT,65536) tags files as containing binary or unknown data.
- TAG(NOTEXT,0) is the equivalent of not specifying the TAG parameter.
- TAG(NOTEXT,273) tags file with the German code set
(ISO-273) but is not eligible for automatic conversion.
Default: NOTEXT
- SYSPLEX(YES|NO)
- For z/OS UNIX, the SYSPLEX statement specifies
whether a system is to join the SYSBPX XCF group to share resources
across the sysplex. If SYSPLEX(YES) is specified, the system participates
in shared file system. If SYSPLEX(NO) is specified, the system does
not participate in shared file system. If the SYSPLEX statement is
not provided, the default is SYSPLEX(NO). Also, to participate in
shared file system, the systems must be at the R9 level or later.
For more information about shared file systems, see Sharing
file systems in a sysplex in z/OS UNIX System Services Planning.
It contains information about parameters specific to shared file system:
SYSPLEX, VERSION, AUTOMOVE, NOAUTOMOVE, and SYSNAME.
Note: You
cannot adjust the SYSPLEX field dynamically. There is no SETOMVS,
SET OMVS, or SETOMVS RESET=(xx) capability. To change the value of
SYSPLEX, you must re-IPL the system.
Default: NO
- TYPE(type_name)
- Specifies the
name of a file system type that is identified in a FILESYSTYPE statement.
The TYPE(type_name) parameter must be the same as the TYPE(type_name)
parameter on a FILESYSTYPE statement.
TYPE is a required parameter.
The name is 1 to 8 characters; the system converts the name to uppercase.
- TTYGROUP(group_name)
- Specifies
the z/OS group
name given to slave pseudoterminals (PTYs) and OCS remote terminals
(RTYs). This group name must defined to the security product and
have a unique group ID (GID). No users should be connected to this
group.
The group_name is used by certain setgid() programs, such
as talk and write, when writing to another user's PTY or RTY.
Value
Range: group_name is a 1 to 8 character value.
Default:
TTY
You can change the value of TTYGROUP dynamically using
the SETOMVS or SET OMVS command. To make a permanent change, edit
the BPXPRMxx member that will be used for future IPLs.
- USERIDALIASTABLE('/etc/tablename')
- Specifies
the path name of a z/OS UNIX file. This file is intended
to contain a list of z/OS user
IDs and group names with their corresponding alias names. The alias
names can contain any characters in the portable file name character
set.
You can change USERIDALIASTABLE dynamically using the SETOMVS
or SET OMVS command. To make a permanent change, edit the BPXPRMxx
member that will be used for IPLs.
Once a user is logged into
the system, changing the user ID or group name alias table does not
change the alias name immediately. If a change needs to be activated
sooner, you can use the SETOMVS or SET OMVS command to change the
table more quickly.
For planning information, see USERIDALIASTABLE
in z/OS UNIX System Services Planning.
- VERSION('nnnn')
- The VERSION statement applies only to systems that are exploiting
shared file systems. VERSION allows multiple releases and service
levels of the binaries to coexist and participate in shared file system.
A directory with the value nnnn specified
on VERSION is dynamically created at system initialization under the
sysplex root that is used as a mount point for the version file system.
This directory, however, is only dynamically created if the sysplex
root HFS is mounted read/write.
Note: nnnn is a
case-sensitive character string no greater than 8 characters in length.
It indicates a specific instance of the version file system. The most
appropriate values for nnnn are the name
of the target zone, &SYSR1, or another qualifier meaningful to
the system programmer. For example, if the system is at V2R9, you
can specify REL9 for VERSION.
When SYSPLEX(YES) is specified,
you must also specify the VERSION parameter. The VERSION value
is substituted in the content of symbolic links that contain $VERSION.
For scenarios that describe the use of the version file system, see Sharing
file systems in a sysplex in z/OS UNIX System Services Planning.
When
testing or changing to a new maintenance level, dynamically change
the VERSION value by using the SETOMVS command.
SETOMVS VERSION='string'
You
can also change the settings of this parameter by using SET OMVS=(xx)
and SETOMVS RESET=(xx) parmlib specifications.
Note: Do not change VERSION dynamically
if you have any users that are logged on or running applications because
replacing the system files for these users might be disruptive.