z/OS JES2 Initialization and Tuning Reference
Previous topic | Next topic | Contents | Contact z/OS | Library | PDF


Parameter Description for MASDEF

z/OS JES2 Initialization and Tuning Reference
SA32-0992-00

AUTOEMEM=ON|OFF
Specifies whether (ON) or not (OFF) this JES2 MAS member wants to have its work restarted if this member fails. If a MAS member has specified AUTOEMEM=ON and the MVS™ system where the member resides fails, its jobs will be warm started if another active member has specified RESTART=YES on its MASDEF statement.
Note: If you specify AUTOEMEM=ON, but no other MAS member has RESTART=YES, and the MVS system on which this member resides fails, then the active work on this member will not be restarted.

However, if the job is registered with the automatic restart manager, the automatic restart manager restarts jobs regardless of the settings of the AUTOEMEM= and RESTART= parameters.

Modification: $T MASDEF operator command.

CKPTLOCK=INFORM|ACTION
CKPTLOCK=INFORM causes JES2 not to issue message $HASP479 during a cold or warm start of JES2 if the following conditions exist:
  • The checkpoint lock cannot be obtained
  • The member id of the member holding the lock is the same as the one being started.
Instead, JES2 issues message $HASP459 to indicate that the member has bypassed confirmation of the lock release.

If you specify CKPTLOCK=ACTION, JES2 issues message $HASP479 in all cases.

Modification: $T MASDEF command during initialization only.

DORMANCY=(mmmm,nnnn)|(100,500)
Use the first positional operand (mmmm) to specify, in hundredths of a second (0-3000), the minimum time a member of a multi-access spool configuration must wait after releasing control of the checkpoint data set (that is, shared queues) before again attempting to gain control of them. This parameter is used to prevent one member of a multi-access spool configuration from monopolizing control of the checkpoint data set.

Use the second positional operand (nnnn) to specify, in hundredths of a second (100-6000), the maximum time a member of a multi-access spool configuration may refrain from attempting to gain control of the checkpoint data set.

When processors are active in JES2, this parameter has little meaning because control of the checkpoint data set is frequently requested. However, when JES2 is idle, this parameter ensures that JES2 periodically looks at the shared queues for eligible work that another member of the configuration may have placed there.
Note:
  1. If the value specified for nnnn is too small, excessive system time could be expended in unnecessary attempts to reacquire the checkpoint data set. However, if the value specified is too large, the start of certain functions and the responses to certain display commands may be delayed.
  2. See z/OS JES2 Initialization and Tuning Guide for a discussion of the relationship between DORMANCY= and HOLD=, and recommended configuration-dependent initial values.
HOLD=nnnnnnnn|99999999
Specifies, in hundredths of a second (0-99999999), the minimum length of time a member of a multi-access spool configuration must maintain control of the checkpoint data set after gaining control of them.
This parameter is provided to minimize the contention that could occur with the checkpoint data set in an environment in which all members of the configuration specify HOLD=0.
Note:
  1. If you have more than one member in the MAS, you MUST specify a value for this parameter.
  2. Setting this parameter to a high value will tend to limit access of other members of the configuration to the shared queues.
  3. See z/OS JES2 Initialization and Tuning Guide for a discussion of the relationship between DORMANCY= and HOLD=, and recommended configuration-dependent initial values.

Performance Considerations: If you specify MODE=DUAL on CKPTDEF, also specifying HOLD=99999999 on the MASDEF statement causes performance degradation. If you must use HOLD=99999999 on MASDEF, be certain to also specify MODE=DUPLEX on CKPTDEF.

Modification: $T MASDEF operator command.

LOCKOUT=nnnnn|1000
Specifies, in hundredths of a second, the time interval (500-15000) from the first-denied request for access to the shared queues of a member of a multi-access spool configuration to the time that the configuration will assume the member controlling the queues is down. When this situation occurs, JES2 issues either the $HASP263 WAITING FOR ACCESS TO JES2 CHECKPOINT VOLUME or $HASP263 WAITING FOR ACCESS TO JES2 CHECKPOINT. LOCK HELD BY MEMBER member message, indicating a lock-out situation, and resets the timer interval to the LOCKOUT value.

Modification: $T MASDEF operator command.

OWNMEMB=xxxx
Specifies the member identifier of this member, where xxxx is the 1- to 4-character alphanumeric name to be used in place of that provided by System Management Facility (SMF). This parameter may be required to warm start JES2 on a system with a different SMF-defined system ID or to warm start JES2 on the same system following an IPL with different SMF specifications. The member identifier for OWNMEMB defaults to the generated system SMF ID.
Note: When naming a member, avoid 1- or 2-digit numbers such as ‘6’ and ‘23’. This can cause confusion when the name is used in commands or displayed in messages. Consider the following initialization statements and series of commands:
MEMBER(3) NAME=6
MEMBER(6) NAME=SYSA
If you enter:
$E MEMBER(3)
JES2 resets the third member (NAME=6).
If you enter:
$E MEMBER(SYSA)
JES2 resets the sixth member (NAME=SYSA).
However, if you enter:
$E MEMBER(6)
it might not be obvious that JES2 resets the third member (MEMBER(3) NAME=6), not MEMBER(6) NAME=SYSA.
Therefore, if you must use 1- or 2-digit names, to reduce any possibly of confusion, pad the number on the left with 0s. The above example becomes:
MEMBER(3) NAME=0006
MEMBER(6) NAME=SYSA

Now $E MEMBER(6) resets the sixth member (MEMBER(6) NAME=SYSA), and $E MEMBER(0006) resets the third member (MEMBER(3) NAME=0006).

Modification: Single-member warm start

RESTART=Yes|No
Specifies whether (Yes) or not (No) a JES2 MAS member can make the jobs on a failed member eligible for restart.
Note:
  1. If the job is registered with the automatic restart manager, the automatic restart manager restarts jobs regardless of the settings of the AUTOEMEM= and RESTART= parameters.
  2. If a JES2 member has been identified as the target of an automatic reconfiguration facility takeover, specify RESTART=NO. For more information about the Processor Resource/Systems Manager™ (PR/SM™) environment to which this applies, see z/OS MVS Setting Up a Sysplex.

Modification: $T MASDEF operator command.

SHARED=CHECK|NOCHECK
Specifies whether the DASD used in your configuration for the spool and checkpoint devices is shared (CHECK) or non-shared (NOCHECK).
Note: You must specify CHECK in a multi-access spool environment and you must specify NOCHECK in a single-member environment. The exceptions to this rule are the following:
  • Use this parameter specifically to test your multi-access spool complex when in a poly-JES environment; shared DASD is not required.
  • If the shared pack is damaged, the installation can use a non-shared pack to bring up one JES2 member.

Modification: $T MASDEF operator command, cold start or warm start.

SYNCTOL=nnn|120
Specifies, in seconds (0-300), the time interval that must elapse before one JES2 member in a multi-access spool configuration is assumed to be not operating.

Actions such as cold start, warm start, or $E MEMBER operator commands are rejected unless the time stamps of the affected members in the shared checkpoint record are greater than the current time minus this parameter value.

Note: JES2 processing uses the SYNCTOL= keyword value to identify when members of differing sysplexes are using the same checkpoint. JES2, in this way, can identify this exceptional condition and take appropriate action. Generally, all members in a sysplex use a unique checkpoint, and any given checkpoint is not typically shared among sysplexes.

Modification: $T MASDEF operator command.

XCFGRPNM=XCF_groupname | nodename
Each MAS in a sysplex is required to use a unique XCF groupname. The groupname must be a 1- to 8-alphanumeric or special character ( $ # @ ) name that follows all XCF naming conventions. To avoid using the names IBM uses for its XCF groups, do not begin group names with the letters A through I or the character string SYS. Also, do not use the name UNDESIG, which is reserved for use by the system programmer in your installation.

The default is the local node name as defined on the NAME= parameter of the local NODE(xxxxxxxx) initialization statement. IBM suggests that you use the default unless it conflicts with an existing XCF group name.

If you set the XCFGRPNM to the SYSPLEX value, ensure that no other functions are also registering to that same name or else you risk receiving $HASP792 at JES2 startup stating that some members were found active but are not part of the SYSPLEX.

Modification: All-member warm start.

Go to the previous page Go to the next page




Copyright IBM Corporation 1990, 2014