Operational integration and reporting considerations
This chapter reviews operational integration considerations.
The Storage cloud setup is the first stage in of moving your data into the cloud. Use of a strong operational framework including set of instructions, housekeeping jobs, and security considerations are encouraged to ensure that you can take the best from your cloud implementation.
We encourage you to consider the suggestions that are described in this chapter and plan and implement your own operations and automation procedures that are based on your system requirements.
This chapter includes the following topics:
11.1 Pre-implementation reporting
Before implementing automatic migration to the cloud, a storage administrator might want to report on possible central processing unit (CPU) savings that are related to offloading the migration task to the DS8000.
IBM released a package to assist storage administrators to collect and report data regarding CPU consumption that is related to migration, recall, and recycle tasks. You can download the package from the following link:
You can use this package to extract and parse your SMF data. Then, you can generate reports and graphics to analyze the CPU savings that can be achieved with cloud migration.
11.2 Operational monitoring
After you finish configuring and activating your cloud settings, the cloud is ready for use. You then perform several tests, and they all complete with success, meaning that the hardware and software configurations are correct, and the network access to the cloud is functional.
Now, it is time to ensure that this access remains functional, and that any errors are tracked by automation systems.
11.2.1 Monitoring cloud setting changes
The first automation process that can be set up is to identify any cloud setting changes within DFSMShsm. Whenever a change to DFSMShsm storage cloud settings fails, a new ARC1581I message is issued with a description of the error that was encountered.
You might add message handling to our automation system to be notified whenever this message is issued and provide a timely reaction to the error. The error message code that is a result of an internal server error is shown in Example 11-1.
Example 11-1 ARC1581I message
Display Filter View Print Options Search Help
-------------------------------------------------------------------------------
SDSF SYSLOG 33.101 3090 3090 10/10/2016 0W 4,414 COLUMNS 52- 131
COMMAND INPUT ===> SCROLL ===> CSR
0010 ARC1581I UNEXPECTED HTTP STATUS 500 DURING A POST FOR 423
0010 ARC1581I (CONT.) URI
0010 ARC1581I (CONT.) https://IBMREDBOOKS.tuc.stglabs.ibm.com/v2.
0010 ARC1581I (CONT.) 0/tokens/ ERRTEXT HTTP/1.1 500 Internal Server Error
11.2.2 Monitoring migration activities
You can track data set migration failures by using the ARC0279I message, which is issued when a data set migration fails. Set up automation to report on these failures based on this message.
An alternative approach is to create a REXX program to read the DFSMShsm log or system log to periodically search for ARC0279I messages. A sample migration error that is the result of a non-existing cloud definition is shown in Example 11-2.
Example 11-2 ARC0279I message
Display Filter View Print Options Search Help
-------------------------------------------------------------------------------
SDSF OUTPUT DISPLAY DFHSM STC00131 DSID 2 LINE 1,361 COLUMNS 02- 81
COMMAND INPUT ===> SCROLL ===> CSR
17.00.38 STC00131 ARC0279I MIGRATION REJECTED - CLOUD NAME VOIDCLOUD NOT 770
770 ARC0279I (CONT.) FOUND
Use the information from the REXX-generated reports to track users that might be trying to migrate invalid data sets, or specifying wrong cloud information. The information helps to identify what information can be included in training for those users that are new to the cloud.
11.2.3 Monitoring reconnections
If you implemented periodic checks of the Hierarchical Storage Manager (HSM) SETSYS configurations, include the cloud configuration information. The new ARC0444I message identifies if cloud-recalled data sets can be reconnected to cloud objects. A sample output from the HSEND QUERY SETSYS command to display th cloud-reconnect setting is shown in Example 11-3.
Example 11-3 ARC0444I reconnection message
Display Filter View Print Options Search Help
-------------------------------------------------------------------------------
SDSF OUTPUT DISPLAY DFHSM STC00131 DSID 2 LINE 120 COLUMNS 02- 81
COMMAND INPUT ===> SCROLL ===> CSR
273 ARC0410I (CONT.) PERCENTAGE=020%, TAPEMAXRECALLTASKS=01, ML2
273 ARC0410I (CONT.) NOT ASSOCIATED GOAL=010, RECONNECT(NONE)
06.17.32 STC00131 ARC0444I CLOUDMIGRATION RECONNECT(ALL)
06.17.32 STC00131 ARC0411I TAPESECURITY=PASSWORD, DEFERMOUNT
11.2.4 Other messages to consider
Other situations that can be monitored include the DUMP and RESTORE processes that are performed by DFSMSdss. Several new messages identify and describe errors during DUMP or RESTORE processing. Consider investigating the following messages to determine whether they can help build a robust operational cloud:
•ADR600E: DFSMSdss did not process the data set because of the condition code detected.
•ADR601E: DFSMSdss invokes the ANTRQST macro for an MCLIST, STORE, or RETRIEVE request and ANTRQST fails with the listed hex return code, reason code, and return information.
•ADR602E: DFSMSdss found that a backup exists with the same object prefix in the specified container.
•ADR604E: A failure occurred while trying to store an object that is related to the dump process or a data set. All related objects stored that use the object-pre fix-name are not usable because of a previous error that was encountered.
•ADR606E: A failure occurred while performing the identified z/OS Client Web Enablement Toolkit service.
•ADR607E: A failure occurred while performing the identified request.
•ADR609E: I/O errors were encountered while the indicated type of dump meta-record was being read during logical data set RESTORE processing.
•ADR610E: DFSMSdss detected an unexpected internal error during processing of an HTTP/HTTPS request.
•ADR612E: DFSMSdss encountered an error obtaining a SYSZADRO enqueue; the resource might be in use.
•ADR705E: A nonexistent storage class, management class, or cloud was specified in the STORCLAS, MGMTCLAS, or CLOUD keyword.
Other monitor options can also be implemented in your systems to monitor and control how clouds are used.
11.3 Operational reporting
There are different options for reporting on DFSMShsm cloud usage. Whether by using the HSEND REPORT command or SMF records, plan to have a reporting and archiving job to analyze and retain storage cloud usage.
11.3.1 Building reports
The JCL and REXX that are included in this topic are intended to show you how to extract cloud migration and recall activity from a daily report, and append the data in a CSV format to output data sets. This file might then be downloaded and imported into a spreadsheet for further analysis.
The JCL that is used to run the report by running the REXX RXMEMBER procedure is shown in Example 11-4.
Example 11-4 JCL to run the REXX procedure
//JOBLIST1 JOB (XXXX),'RUN RPT',NOTIFY=&SYSUID,MSGLEVEL=(1,1),
// MSGCLASS=W
//STEP1 EXEC PGM=IKJEFT01,REGION=8M
//SYSTSPRT DD SYSOUT=A
//HSMREPT DD DSN=YOUR.INPUT.DATA,DISP=SHR
//CLOUDRPT DD DSN=YOUR.OUTPUT.REPORT,DISP=(NEW,CATLG),
// LRECL=80,RECFM=FB,SPACE=(TRK,(1,1)),DSORG=PS
//SYSTSIN DD *
EX 'YOUREXX.DATASET(RXMEMBER)'
The REXX source code is shown in Example 11-5 on page 113. You can use this code as a base to develop your own specific reports.
Example 11-5 REXX source code
/* REXX */
"EXECIO * DISKR HSMREPT (STEM HSMREPT. FINIS)"
/* NUMBER OF MIGRATION TO CLOUD LINES */
MIG=0
/* NUMBER OF RECALLS TO CLOUD LINES */
REC=0
DO Z=1 TO HSMREPT.0
/* GET REPORT DATE */
IF LASTPOS('DAILY STATISTICS REPORT FOR',HSMREPT.Z) > 0 THEN
PARSE VAR HSMREPT.Z 'DAILY STATISTICS REPORT FOR' REPDATE .
IF LASTPOS('PRIMARY - CLOUD',HSMREPT.Z) > 0 THEN DO
PARSE VAR HSMREPT.Z . . . NDS RTRK RBYT WTRK WBYT SYSR USRR FAIL ,
AGE QTIME WTIME PTIME TTIME
MIG = MIG + 1
OUTMIG.MIG = REPDATE','NDS','RTRK','RBYT','WTRK','WBYT','SYSR,
||','USRR','FAIL','AGE','QTIME','WTIME','PTIME','TTIME
END
IF LASTPOS('CLOUD - PRIMARY',HSMREPT.Z) > 0 THEN DO
PARSE VAR HSMREPT.Z . . . NDS RTRK RBYT WTRK WBYT SYSR USRR FAIL ,
AGE QTIME WTIME PTIME TTIME
REC = REC + 1
OUTREC.REC = REPDATE','NDS','RTRK','RBYT','WTRK','WBYT','SYSR,
||','USRR','FAIL','AGE','QTIME','WTIME','PTIME','TTIME
END
END
"EXECIO * DISKW CLOUDRPT (STEM HSMREPT. FINIS)"
Other reports can be created by using SMF records. Some suggestions of reports that can be generated include filtering migration and recall by data set high-level qualifiers, users, management class, data set size, and others. We suggest that you create at least one report that consolidates data sets by high-level qualifiers so that you can identify the applications that are making most use of cloud resources.
11.3.2 DCOLLECT reports
Along with the changes in DFSMShsm Control Data Sets (CDSs) and SMS constructs to enable the use of cloud storage, the DCOLLECT was also updated to reflect the extra information available.
In the DCOLLECT record type ‘MC’, cloud-related fields are also displayed, including the cloud names the management class relates to, and actions to take based on data set size during migration.
The ‘M’ records are updated to include the cloud name length, cloud name, container name, and number of objects stored.
A sample usage for this extra information includes using the DCOLLECT to gather information about the containers that are created and owned by DFSMShsm in the cloud, and the number of objects stored. This information might be specially valuable for large cloud environments, where list commands can take an extended amount of time to complete.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.