The CEMT PERFORM commands are not concerned with resource status. They are used to:
Descriptions of the CEMT PERFORM commands follow in alphabetic order. The operands for each of the commands are also presented in alphabetic order.
You can use the PERFORM CLASSCACHE command to initialize (start or reload) and terminate (phase out, purge or forcepurge) the shared class cache. While you are performing one of these operations, you can use other options on the command to set attributes of the shared class cache, as follows:
The PERFORM CLASSCACHE command is intended to operate on the active shared class cache in the CICS region. The Start and Reload operations only affect this shared class cache. The Phaseout, Purge and Forcepurge operations, however, also act on any old shared class caches that are still present in the region because they are waiting for worker JVMs that are dependent on them to be phased out. For example, if you issue a CEMT PERFORM CLASSCACHE Forcepurge command, all tasks using worker JVMs dependent on a shared class cache are forcepurged, both those dependent on the current shared class cache and those dependent on old shared class caches. Note, though, that JVMs running independently of the shared class cache are not affected by the CEMT PERFORM CLASSCACHE Phaseout, Purge and Forcepurge commands.
When you have issued one command to terminate the shared class cache, you can repeat the CEMT PERFORM CLASSCACHE command to make a stronger attempt to purge the tasks using the JVMs (for example, by using Purge instead of Phaseout). However, you should only repeat the command if autostart for the shared class cache is disabled. If autostart is enabled, and you repeat the command to terminate the shared class cache, the command could operate on the new shared class cache that has been started by the autostart facility, and terminate it.
When you use the Cachesize, Profile and Autostartst options, the new settings override the size, JVM profile and autostart status that were originally set by the system initialization parameters JVMCCSIZE, JVMCCPROFILE and JVMCCSTART. Subsequent CICS restarts use the most recent settings that you made using the PERFORM CLASSCACHE command (or the SET CLASSCACHE command), unless the system is INITIAL or COLD started, or the system initialization parameters are specified as overrides at startup. In these cases, the settings from the system initialization parameters are used.
CEMT PERFORM CLASSCACHE >>-CEMT Perform CLasscache--------------------------------------> >--+-+-Start--+-+-------------------+-+---------------+-+------>< | '-Reload-' '-Cachesize(number)-' '-Profile(name)-' | '-+-PHaseout---+-+----------+------------------------' +-PUrge------+ +-Enabled--+ '-Forcepurge-' '-Disabled-'
The operations you can perform with the PERFORM CLASSCACHE command are:
If autostart is enabled for the shared class cache, a new shared class cache is started as soon as a new JVM requests it. If autostart is disabled, the CEMT PERFORM CLASSCACHE Start command (or the equivalent EXEC CICS command) must be entered if you want to start a new shared class cache. If you do not restart the shared class cache, requests to run a Java application in a JVM whose profile requires the use of the shared class cache will fail. Specify the Autostartst option to set the status of autostart.
If autostart is enabled for the shared class cache, a new shared class cache is started as soon as a new JVM requests it. If autostart is disabled, the CEMT PERFORM CLASSCACHE Start command (or the equivalent EXEC CICS command) must be entered if you want to start a new shared class cache. If you do not restart the shared class cache, requests to run a Java application in a JVM whose profile requires the use of the shared class cache will fail. Specify the Autostartst option to set the status of autostart.
If autostart is enabled for the shared class cache, a new shared class cache is started as soon as a new JVM requests it. If autostart is disabled, the CEMT PERFORM CLASSCACHE Start command (or the equivalent EXEC CICS command) must be entered if you want to start a new shared class cache. If you do not restart the shared class cache, requests to run a Java application in a JVM whose profile requires the use of the shared class cache will fail. Specify the Autostartst option to set the status of autostart.
The additional options that you can use on the PERFORM CLASSCACHE command to set attributes of the shared class cache are:
When you use the name of a JVM profile anywhere in CICS, you must enter it using the same combination of upper and lower case characters that is present in the HFS file name. If you need to enter the name of a JVM profile in mixed case when you are using the CEMT transaction, ensure that the terminal you use is correctly configured, with upper case translation suppressed.
If the Start or Reload command is successfully executed but the master JVM fails to start, CICS retains the JVM profile name that you specify, and displays it as the JVM profile for the master JVM. Next time you issue the command to start or reload the shared class cache, the new JVM profile will be used, unless you change it again using the Profile option.
The PERFORM CORBASERVER command allows you to:
If no action is specified, the command has no effect.
You can issue a PERFORM CORBASERVER command only when the CorbaServer is in a steady state--that is, when it is in ENABLED or DISABLED state, but not when it is in ENABLING, DISABLING, or DISCARDING state.
In addition to the traditional DFSMS-managed data sets, such as partitioned data sets (PDS), CICS requires access to z/OS UNIX System Services and the hierarchical file store (HFS) in order to launch a JVM. This requires that the CICS region userid is authorized to z/OS UNIX System Services to enable CICS to access the required HFS directories, such as those specified on the JVMPROPS, LIBPATH, and CLASSPATH parameters.
A failure to perform the requested action will be reported by returning a condition and RESP2 value from the command and, in certain cases, by writing messages to CSMT. The requested action may have beens partially successful since a failure does not cause the effects of the command to be backed out. Also, failure does not necessarily stop the command from continuing in order to increase the extent to which it was partially successful.
publishes:
Publishing a bean means binding a reference to the home of the bean in the namespace. The naming context in which the bean is bound is named, relative to the initial context defined for the CICS region, using a concatenation of the JNDI PREFIX attribute of the CorbaServer and the name of the bean.
The Generic factory is bound with the name GenericFactory concatenated to the PREFIX attribute of the Corbaserver.
If CICS finds any new deployed JAR files in the pickup directory, it copies them to its shelf directory and dynamically creates and installs DJAR definitions for them.
If CICS finds any updated deployed JAR files in the pickup directory, it updates both the LASTMODTIME, DATESTAMP, and TIMESTAMP attributes of the installed DJAR definitions and the shelf copies of the deployed JAR files.
If you receive a "DJAR CONFLICT" message, it means that one or more deployed JAR files have failed to install due to a naming conflict. This error can occur for two reasons:
To avoid this error:
For more information about DJAR naming conflicts and how to avoid them, see the CICS® Resource Definition Guide.
Cause the CICS timeout delete mechanism to be invoked immediately.
The CICS timeout delete mechanism removes any shipped terminal definitions that have not been used for longer than the time specified on the DSHIPIDL system initialization parameter, or on a subsequent SET DELETSHIPPED IDLE command. The interval between scheduled invocations of the mechanism is specified on the DSHIPINT system initialization parameter, or on a subsequent SET DELETSHIPPED INTERVAL command.
PERFORM DELETSHIPPED invokes timeout delete immediately.
The PERFORM DJAR command performs certain actions against a specified deployed JAR file. If no action is specified, the command has no effect.
In addition to the traditional DFSMS-managed data sets, such as partitioned data sets (PDS), CICS requires access to z/OS UNIX System Services and the hierarchical file store (HFS) in order to launch a JVM. This requires that the CICS region userid is authorized to z/OS UNIX System Services to enable CICS to access the required HFS directories, such as those specified on the JVMPROPS, LIBPATH, and CLASSPATH parameters.
A failure to perform the requested action will be reported by returning a condition and RESP2 value from the command and, in certain cases, by writing messages to CSMT. The requested action may have beens partially successful since a failure does not cause the effects of the command to be backed out. Also, failure does not necessarily stop the command from continuing in order to increase the extent to which it was partially successful.
Obtain a CICS system dump.
To get a CICS system to dump and continue processing, you can use either the CEMT PERFORM DUMP command or the CEMT PERFORM SNAP command. Both commands produce a "snapshot" of the CICS system.
The use of CEMT PERFORM DUMP or CEMT PERFORM SNAP prevents all other tasks from running. Consequently, terminal operators may experience delays in response.
End an affinity owned by CICS.
Where CICS is a member of a VTAM generic resource group, PERFORM ENDAFFINITY instructs VTAM to end an affinity owned by CICS, whether or not the connection has been deleted. If the connection has not been deleted, it must be out of service and have no recovery information outstanding--that is, its RECOVSTATUS (as displayed by CEMT INQUIRE/SET CONNECTION) must be NORECOVDATA.
Generic resources and affinities are described in the CICS Intercommunication Guide.
If you do not specify a NETID, CICS takes the value from the installed connection, if it exists. If you do not specify a NETID and the connection does not exist, the command fails.
If a request to end an affinity is rejected by VTAM because no such affinity exists, CICS issues message DFHZC0181. This may mean either that you entered the NETNAME or NETID incorrectly, or that you (or CICS) was wrong in supposing that an affinity existed.
Use the PERFORM PIPELINE command to initiate a scan of the Web service binding directory that is specified in the WSBIND attribute of the PIPELINE definition. If the WSBIND attribute is not specified, there is nothing to scan, and control returns to your program.
If the directory location specified is valid, CICS examines the Web service binding files in the directory to determine if they should be installed into the system:
If, for any reason, CICS fails to install an individual Web service binding file, processing continues with the remaining files in the directory. When the scan completes, the PIPELINE is available for use with whichever of the binding files were installed successfully.
A failure to perform the requested action will be reported by returning a condition and RESP2 value from the command and, in certain cases, by writing messages to CSMT. The requested action may have beens partially successful since a failure does not cause the effects of the command to be backed out. Also, failure does not necessarily stop the command from continuing in order to increase the extent to which it was partially successful.
Synchronize the CICS date and time-of-day with the system date and time-of-day.
To ensure that the correct local time is used by all CICS functions, use the CEMT PERFORM RESET command immediately whenever you alter the system date or the time-of-day in the MVS TOD clock while a CICS region is running.
For example, when setting clocks forward or back an hour to adjust for Summer and Winter time, use this command to ensure that CICS immediately updates the correct local time.
Rebuild the in-store RACF resource profiles. The RACF resource profiles are rebuilt by reissuing the RACLIST macro for the active class names.
If the external security manager is RACF 2.1 or an equivalent product, the PERFORM SECURITY command is not required to refresh the classes used for resource checks. A message of NOT REQUIRED is displayed and no action is taken.
Shut down the CICS region. The shutdown can be either controlled or immediate. If a shutdown (either normal or
immediate) hangs, a subsequent SHUTDOWN IMMEDIATE resumes the shutdown process.
CEMT PERFORM SHUTDOWN >>-CEMT Perform SHUTdown--+------------+--+------------+--------> '-Xlt(value)-' '-Plt(value)-' >--+---------------+--+--------------------------+--------------> +-Sdtran(value)-+ '-Immediate--+-----------+-' '-NOSdtran------' '-NORestart-' >--+----------+--+------+-------------------------------------->< '-Takeover-' '-Dump-'
The default shutdown transaction, if neither SDTRAN nor NOSDTRAN are specified, is CESD.
Transactions listed in the specified XLT can be initiated after the SHUTDOWN request, and before the system quiesces. The default XLT should contain an entry for the master terminal transaction. This ensures that you can issue an immediate shutdown if, for example, a task is found to be suspended after an ordinary (nonimmediate) shutdown has been requested.
Table 5 shows the effect of the TAKEOVER and IMMEDIATE options on the active CICS system.
CEMT PERFORM SHUTDOWN | |||
(normal) | TAKEOVER | IMMEDIATE | |
Tasks | finish | finish | terminate |
SNA session (XRF) | unbind | leave | leave |
SNA session (non-XRF) | unbind | unbind | unbind |
Table 6 shows the effect of the TAKEOVER and IMMEDIATE options on the alternate CICS system.
CEMT PERFORM SHUTDOWN | |||
(normal) | TAKEOVER | IMMEDIATE | |
CICS alternate | shuts down | takes over | takes over |
For further information on XRF, see the CICS/ESA 3.3 XRF Guide.
Write the statistics for a named resource type immediately to the SMF data set, rather than wait for the current statistics-gathering interval to expire. The recording is made irrespective of the ON or OFF setting of statistics. See also topic CEMT INQUIRE STATISTICS and topic CEMT SET STATISTICS for a description of the CEMT INQUIRE and SET STATISTICS commands.
You can request either statistics for all resources (ALL), or statistics for individual resources in various combinations by selecting one or more keywords from the list.
The message ‘STATS MISSING’ appears when the resource type is either not functioning or not available. As much data is recorded as possible.
See the CICS Operations and Utilities Guide for information about printing statistics using the CICS-supplied utility program DFHSTUP, and for information about the reports produced by DFHSTUP.
CEMT PERFORM STATISTICS >>-CEMT Perform STatistics--+--------+--------------------------> '-RECORd-' .-----------------------. V | >----+-ALl--+----------+-+-+----------------------------------->< | '-RESetnow-' | +-AUtoinstall-------+ +-Bean--------------+ +-Connection--------+ +-CORbaserver-------+ +-DB2---------------+ +-DIspatcher--------+ +-Enqueue-----------+ +-FEpi--------------+ +-FIle--------------+ +-Journalname-------+ +-JVMPOol-----------+ +-JVMPROFile--------+ +-JVMPROGram--------+ +-Lsrpool-----------+ +-Monitor-----------+ +-PIpeline----------+ +-PROGAuto----------+ +-PROGRam-----------+ +-RECOVery----------+ +-REQuestmodel------+ +-STAts-------------+ +-STOrage-----------+ +-STReamname--------+ +-SYsdump-----------+ +-TAblemgr----------+ +-TCLass------------+ +-TCPIP-------------+ +-TCPipservice------+ +-TDqueue-----------+ +-TErminal----------+ +-TRAnsaction-------+ +-TRDump------------+ +-TSqueue-----------+ +-Urimap------------+ +-Vtam--------------+ '-Webservice--------'
ALL RESETNOW resets the counters for the collection of statistics. RESETNOW only works if the ALL operand is set, that is, ALL RESETNOW.