|
|
< Day Day Up > |
|
SMF Type 88 records contain information about usage of a log stream data for a system in a sysplex. Using the data from the Type 88 records can help you tune and perform capacity planning on your log streams.
Type 88 records summarize all a log stream's activity on a single system, as long as at least one address space is connected to the log stream on that system. If no System Logger write activity is performed against the log stream during a particular SMF interval, a record is produced showing zero for the various System Logger activity total fields.
The Type 88 record is produced for all log streams connected at the expiration of the SMF global recording interval. The creation of a Type 88 record is also triggered by the disconnection of the last connector to the log stream on that system.
There are some things you must remember about the Type 88 records however:
The Type 88 records only show activity from one system. If a log stream is used by more than one system, you must merge the information from all systems to get a comprehensive view of the use of that log stream
There are a number of subtype records produced by System Logger, but only one of them, the one dealing at the log stream level, has an IBM-provided formatter. Therefore, to determine what is happening at the structure level for structures containing more than one log stream, you need to merge the information from all log streams in that structure, from all systems in the sysplex.
In "SMF88" on page 260 you can find a discussion on how to set up for collecting Type 88 records. After the Type 88 records are collected, they can be sorted and
IXGRPT1 provides one of the few ways to effectively tune your log streams for better performance. However, interpreting the report that is produced by the IXGPRT1 program is not a trivial task. In "SMF Type 88 records and IXGRPT1 program" on page 281 you can find a discussion on using the Type 88 data to tune System Logger performance as well as highlighting particular fields of interest in some of the application chapters.
|
|
< Day Day Up > |
|
|
|
< Day Day Up > |
|
In this section we discuss maintaining the System Logger component and what documentation should be collected when a problem occurs to speed up problem diagnosis.
Depending on the problem, the following seven
Obtain a dump of the System Logger address space and any address spaces connected to the log stream you are experiencing a problem with. Refer to "DUMP command parmlib member" on page 260 for sample on the dump keywords.
Use the D LOGGER command to display the following information:
IXGLOGR address space status
D LOGGER,ST
Log stream, CF structure, and connection information.
D LOGGER,L
D LOGGER,STR
D LOGGER,C
Sysplex status for log streams.
D LOGGER,C,SYSPLEX
Specifics for DASD-only log streams.
D LOGGER,C,DASDONLY
D LOGGER,L,DASDONLY
Use IDCAMS to print the offload data sets for a log stream using the JCL shown in Example 2-11.
Example 2-11: IDCAMS job to print offload data sets.
|
|
//IDCAMS1 JOB (0,0),CLASS=A,MSGCLASS=X,NOTIFY=&SYSUID //*----------------------------------------------------------*/ //* PRINT contents of System Logger offload data set */ //* ---------------------------------------------------------*/ //* //PRINTIT EXEC PGM=IDCAMS //SYSPRINT DD SYSOUT=* //OFFLOAD DD DSN=hlq.xxx.A00000yyy.DATA,DUSP=SHR //SYSIN DD * PRINT INFILE(OFFLOAD) /*
| Notes |
|
|
|
Obtain offload data set characteristics. You can use the sample JCL in Example 2-12 to display the characteristics of the data set you are dumping.
Example 2-12: IDCAMS sample JCL to show DS characteristics
|
|
//IDCAMS2 JOB (0,0),CLASS=A,MSGCLASS=X,NOTIFY=&SYSUID //PRINTIT EXEC PGM=IDCAMS //SYSPRINT DD SYSOUT=* //SYSIN DD * LISTCAT ALIAS ALL LISTCAT ALIAS ALL CAT('SROCAT.CATALOG') LISTCAT ENT('USER.CATALOG.NAME') ALL CAT('
USER
.CATALOG.NAME') LISTCAT LVL('HLQ_NAME') ALL /*
This job will:
Display all alias
Display all alias names defined in a specified catalog.
Display the contents of a user catalog and the volume on which it resides.
Display all information
|
|
Obtain a System Logger inventory detailed listing by running an IXCMIAPU LIST report as shown in Example 2-13.
Example 2-13: Sample IXCMIAPU report job
|
|
//LOGRRPT JOB (999,POK),'LOGR POLICY',CLASS=A,REGION=4M, // MSGCLASS=X,TIME=10,MSGLEVEL=(1,1),NOTIFY=&SYSUID //STEP1 EXEC PGM=IXCMIAPU //SYSPRINT DD SYSOUT=* //SYSABEND DD SYSOUT=* //SYSIN DD * DATA TYPE(LOGR) REPORT(YES) LIST LOGSTREAM NAME(logstreamname) DETAIL(YES) /*
|
|
If you suspect that the LOGR CDS is corrupted, print its contents using the job shown in Example 2-14.
Example 2-14: Sample job to dump the LOGR CDS
|
|
//DUMPCDS JOB (0,0),CLASS=A,MSGCLASS=X,NOTIFY=SYSUID //******************************************************/ //* RUN ADRDSSU to dump off the LOGR Couple Data set */ //******************************************************/ //STEP1 EXEC PGM=ADRDSSU,REGION=4M //SYSPRINT DD SYSOUT=* //DD1 DD DISP=SHR,VOL=SER=xxxxxx,UNIT=SYSDA //SYSIN DD * PRINT DATASET(logr.couple.dataset) INDDNAME(DD1) TOL(ENQF) /*
|
|
| Note |
As we discussed in 2.5, "Log streams" on page 22, specifying DIAG(YES) on the log stream definition will provide additional documentation in certain error cases that could not be
|
Set a SLIP trap. Your IBM support representative will provide you with the information you need to set the appropriate SLIP for the problem you are experiencing. See z/OS MVS Diagnosis: Tools and Service Aids , GA22-7589 for more information on this.
To avoid rediscovering problems that already have a fix available, it is important that you keep your System Logger component up to date. Therefore, we recommend you apply all PTFs for:
HIPER APARs
APARs with the keyword LOGRSERVICE in the APAR description
APARs with the keyword CFDUPLEX in the APAR description
|
|
< Day Day Up > |
|