APAR status
Closed as documentation error.
Error description
1. TWSz behaves exactly the opposite as documented in the manual: Customization and Tuning in regard to the keyword DYNAMICDEL and its values: YES or NO. In Customization and Tuning you will find in: section 1.1.6 BATCHOPT in the explanation text of DYNAMICDEL(YES NO) the following """" DYNAMICDEL: determines if a special resource that has been dynamically added to the current plan can be deleted if the current plan is changed, without checking the normal conditions listed in the section "setting the global values" of the Managing the Workload manual. Specify **NO** if dynamically added changes can be deleted when the current plan is changed without further checking. Specify **YES** if dynamically added changes can be deleted when the current plan is changed. """"" """"" But this is not true and tests show that the TWSz behavior is the opposite. Therefor the sentence: Specify **NO** if dynamically added changes can be deleted when the current plan is changed without further checking. Specify **YES** if dynamically added changes can be deleted when the current plan is changed. in the manual Customization and Tuning should be changed to: Specify **NO** if dynamically added changes can be deleted when the current plan is changed. Specify **YES** if dynamically added changes can be deleted when the current plan is changed without further checking. . 2. The TWSz MANAGING THE WORKLOAD manual, SC32-1263, Chapter 17 - OVERVIEW OF JOB TRACKING ON z/OS, Topic - HOW TO MAKE SURE THAT EVENTS ARE NOT LOST, lists three techniques for ensuring that events are not lost between the time TWSz is taken down and JES is taken down. . The technique listed first is the LEAST DESIRABLE, and is incompletely documented. This section of the manual should be revised as follows. A) The method currently listed as #2 -- Removing the system being stopped from the JES2 MAS (placing it into INDEPENDENT MODE) should be listed as #1. This is the best option when a single system is being restarted. B) The method currently listed as #3 -- Leaving the TRACKER run until JES itself is stopped should be listed as #2. C) The method currently listed as #1 -- Restarting the TRACKER SUB=MSTR should be listed as #3. This is the LEAST desirable option, and has severe limitations which need to be more thoroughly documented. To that end, step #2 in this method should be changed to read as follows: 2. Bring it up again under the master scheduler (S OPCx,SUB=MSTR). Remember that a tracker cannot use JES services (SYSOUT DATA SETS, INTERNAL READER, JCC) if it runs under the master scheduler. Thus, this technique is not suitable if the TRACKER is a destination for any workstation, if the JCC is active, or if EQQMLOG is allocated to SYSOUT. Change 3 The 'Managing the Workload' Guide needs to be updated in order to clarify that if a dataset, present in a job undergoing Restart and Cleanup, is deleted or renamed, this can impact the results of the restarted job. . Note that a batch job run outside of TWSz or a TSO USERID can update, delete, or rename a GDG instance which is being used by a TWSz restart job. This can result in abend S737 ( ABENDS737 ABEND737 737 ) for the batch job case. . Change 4 In the 'Managing the Workload' Guide the description for EXT records needs to be updated . Change 5 In the 'Messages and Codes' manual the EQQE210I message needs to be modified Change 6 In the 'Managing the Workload' manual the behaviour of the period ETTRCY1 needs to be better explained.
Local fix
Problem summary
**************************************************************** * USERS AFFECTED: 1. TWS for z/OS users. * * FUNCTION=DP * * 2. TWS for z/OS users. * * FUNCTION=EM. * * 3. All TWS for zOS Restart & Cleanup users. * * FUNCTION=R&C * * 4. TWS for z/OS users. * * FUNCTION=EXIT. * * 5. TWS for z/OS users. * * FUNCTION=EXIT. * * 6. TWS for z/OS users. * * FUNCTION=ETT. * **************************************************************** * PROBLEM DESCRIPTION: 1. TWSz behaves the opposite of what * * documented in regard to the keyword * * DYNAMICDEL and its values:Yes or No. * * 2. Tracker running SUB=MSTR cannot use * * internal reader, JCC, or sysout. * * 3. A GDG data set deletion may impact * * Restart and Cleanup results. * * 4. EXT records needs to be correctly * * documented * * 5. Description of EQQE210I message * * needs to be changed * * 6. Description of the behaviour of * * ETTRCY1 period * **************************************************************** * RECOMMENDATION: APPLY THE PTF FIXING THIS APAR * **************************************************************** 1. According to the section "setting the global values" of the Managing the Workload manual the DYNAMICDEL description in the Customization and Tuning manual must be inverted, the description related to No must be Yes and vice versa. 2. The order of the methods to how to make sure that events are not lost must be changed and in the method that is currently listed as 1 must be partially rewritten. 3. The 'Managing the Workload' Guide needs to be updated in order to clarify that if a dataset, present in a job undergoing Restart and Cleanup, is deleted or renamed this can impact the results of the restarted job. 4. The 'Managing the Workload' Guide needs to be updated in order to change the description for the EXT records. 5. The 'Messages and Codes' manual needs to be updated in order to change the description for EQQE210I message. 6. The 'Managing the Workload' manual needs to be updated in order to change the description for ETTRCY1 behaviour.
Problem conclusion
THE FOLLOWING TWS for z/OS PUBLICATIONs WILL BE UPDATED: - - - Change 1: +--------------------------------------------------------------+ | | |TITLE: IBM Tivoli Workload Scheduler for z/OS, | | Customization and Tuning SC32-1265-02 | | SC32-1265-03 | | SC32-1265-04 | | SC32-1265-05 | | | |Chapter 1. Initialization Statements-Reference | | | | Paragraph : Specifying the Statements | | | | Section : Selecting Statements | | | | TOPIC : BATCHOPT | | | | | | Error in the DYNAMICDEL statemet, change the value | | description from: | | | | Specify NO if dynamically added changes can be deleted when | | the curren plan is changed without further checking. | | | | Specify YES if dynamically added changes can be deleted when | | the current plan is changed. | | | | to: | | | | Specify NO if dynamically added changes can be deleted when | | the current plan is changed. | | | | Specify YES if dynamically added changes can be deleted when | | the curren plan is changed without further checking. | | | +--------------------------------------------------------------+ - - - Change 2: +--------------------------------------------------------------+ | | |TITLE: IBM Tivoli Workload Scheduler for z/OS, | | Managing the Workload SC32-1263-02 | | SC32-1263-03 | | SC32-1263-04 | | | |CHAPTER : 17. Overview of job tracking on z/OS | | | | SC32-1263-05 | | | |CHAPTER : 14. Overview of job tracking on z/OS | | | | | | TOPIC : How to make sure that events are not lost | | | | | | Change the order of the bullets and ant semtences from: | | | | Method 1 | | 1. Take the tracker down (P OPCx). | | 2. Bring it up again under the master scheduler | | (S OPCx,SUB=MSTR). Remember that a tracker cannot use | | JES services (SYSOUT data sets, JCC) if it runs under | | the master scheduler. | | 3. Bring down JES. | | 4. Take the tracker down (P OPCx). | | | | Method 2 | | 1. Remove the system being stopped from the JES2 MAS by | | placing it into independent mode (issue $T MEM,IND=Y). | | 2. Allow the jobs currently running on this system to | | complete. | | 3. Stop the tracker (POPCx). | | 4. Stop JES. | | 5. Re-IPL. | | 6. Restart JES. | | 7. Restart the tracker | | 8. Resume normal work (issue $T MEM,IND=N). | | | | Method 3 | | 1. Abend JES2 ($PJES2,ABEND). | | 2. Stop the tracker (P OPCx). | | 3. Re-IPL | | 4. Restart JES. This will be a hot start. | | 5. Restart the tracker. | | | | to: | | | | Method 1 | | 1. Remove the system being stopped from the JES2 MAS by | | placing it into independent mode (issue $T MEM,IND=Y). | | 2. Allow the jobs currently running on this system to | | complete. | | 3. Stop the tracker (POPCx). | | 4. Stop JES. | | 5. Re-IPL. | | 6. Restart JES. | | 7. Restart the tracker | | 8. Resume normal work (issue $T MEM,IND=N). | | | | Method 2 | | 1. Abend JES2 ($PJES2,ABEND). | | 2. Stop the tracker (P OPCx). | | 3. Re-IPL | | 4. Restart JES. This will be a hot start. | | 5. Restart the tracker. | | | | Method 3 | | 1. Take the tracker down (P OPCx). | | 2. Bring it up again under the master scheduler | | (S OPCx,SUB=MSTR). | | Remember that a tracker cannot use JES services (SYSOUT | | DATA SETS, INTERNAL READER, JCC) if it runs under the | | master scheduler. Thus, this technique is not suitable | | if the TRACKER is a destination for any workstation, if | | the JCC is active, or if EQQMLOG is allocated to SYSOUT. | | 3. Bring down JES. | | 4. Take the tracker down (P OPCx). | | | | | +--------------------------------------------------------------+ - - - Change 3: +--------------------------------------------------------------+ | | |TITLE: TWS for z/OS Managing the Workload SC32-1263-02 | | Guide SC32-1263-03 | | SC32-1263-04 | | SC32-1263-05 | | | |CHAPTER: 17 TOPIC: "Restart and cleanup" | | | |PARAGRAPH: "JCL changes considerations" | |Change FROM: | | "As a consequence, if the JCL structure is changed (via | | exit01, ISPF, directives like FETCH, BEGIN, or END) in | | a way that makes the list passed....." | |TO: | | "As a consequence, if the JCL structure is changed (via | | exit01, ISPF, directives like FETCH, BEGIN, or END) or | | any data sets obtained from the previous run is renamed | | or deleted in a way that makes the list passed..... | | | +--------------------------------------------------------------+ +--------------------------------------------------------------+ | | |TITLE: TWS for z/OS Managing the Workload SC32-1263-02 | | Guide SC32-1263-03 | | SC32-1263-04 | | SC32-1263-05 | | | |CHAPTER: 17 TOPIC: "Restart and cleanup" | | | |PARAGRAPH: "JCL changes considerations" | |FROM: | | "In general you should not: | | ... | | - Delete a DD that is not a DUMMY DD (unless done via | | EQQUX001 at each call, so that the submitted JCL will | | never contain that DD). | | ... " | |TO: | | "In general you should not: | | ... | | - Delete a DD that is not a DUMMY DD (unless done via | | EQQUX001 at each call, so that the submitted JCL will | | never contain that DD). | | - Delete or rename a dataset obtained from a previous run | | that is needed to perform step restart actions | | ... " | | | +--------------------------------------------------------------+ Change 4: +--------------------------------------------------------------+ | | |TITLE: IBM Tivoli Workload Scheduler for z/OS, | | Managing the Workload SC32-1263-02 | | SC32-1263-03 | |CHAPTER : 5. Data areas | | | | All description for EXT record is wrong. | | Change it in: | | | | EXT - Set Time Event | | | | Name: DCLEXT | | | | Function: | | This segment declares a SET TIME (GMTOFFSET) event record. | | These event records are built by the EQQEVPGM and EQQEVCMD | | programs and then are passed to the Event Writer by using | | CSA buffers. The Event Writer writes them in the event | | datasets managing them as a part of an event record. | | | | Offsets | | Dec Hex Type Len Name (Dim) Description | | 0 (0) STRUCTURE 80 EXT READY TO RECEIVE EVENT | | 0 (0) CHARACTER 1 EXTETYPE EVENT TYPE = T | | 1 (1) CHARACTER 3 * RESERVED | | 4 (4) CHARACTER 4 EXTSID SYSTEM IDENTIFICATION | | 8 (8) SIGNED 4 EXTGMTOF GMT OFFSET IN MINUTES | | 12 (C) SIGNED 4 EXTCDAT CREATION DATE FORMAT | | (00YYDDDF) | | 16 (10) SIGNED 4 EXTCTIM CREATION TIME FORMAT (SEC*100) | | 20 (14) SIGNED 4 EXTEDAT EVENT DATE FORMAT (00YYDDDF) | | 24 (18) SIGNED 4 EXTETIM EVENT TIME FORMAT (SEC*100) | | 28 (1C) SIGNED 4 EXTGMTSEC GMT OFFSET | | 32 (20) CHARACTER 4 EXTENF T type ENF | | 36 (24) UNSIGNED 1 EXTENFB1 ENF 53 Byte1 | | 37 (25) UNSIGNED 1 EXTENFB2 ENF 53 Byte2 | | 38 (26) UNSIGNED 1 EXTENFB3 ENF 53 Byte3 | | 39 (27) CHARACTER 1 EXTRCHA REAL CHANGE Y/M | | 40 (28) CHARACTER 48 * RESERVED | | 76 (4C) CHARACTER 4 EXTOPCID EVENT ID 'EQQX' | | | | | +--------------------------------------------------------------+ Change 5: +--------------------------------------------------------------+ | | |TITLE: TIVOLI WORKLOAD SCHEDULER FOR z/OS, | | Messages and Codes SC32-1267-02 | | SC32-1267-03 | |CHAPTER: 10 EQQEnnn Messages | | | | Modify the System action of EQQE210I message FROM: | | | | System action: The tracker generates the event and | | modifies the GMT offset, on the controller site, if the | | event was created in the same system the controller | | will manage it and signal the change in its MLOG. If | | the event was created in another system the controller | | signals the message on its MLOG. | | | | IN: | | | | System Action: The tracker generates the event and, if the | | difference of GMT offset value is at least one minute long, | | modifies its GMT offset. Processing that event, the | | controller will: | | | | - Issue the EQQE210I message in its MLOG and, if the event | | was created on its system, update its GMT offset (but only | | if the difference is at least one minute long). | | - Issue the EQQE210I message but it will not revise its GMT | | offset value, if the event was created on a different | | system. | | | +--------------------------------------------------------------+ Change 6: +--------------------------------------------------------------+ | | |TITLE: TWS for z/OS Managing the Workload | | Version 8.3 SC32-1263-03 | | Version 8.5 SC32-1263-04 | | Version 8.5.1 SC32-1263-05 | | | |Section Automatically adding an occurrence to | | the current plan | | | |The sentence: | |"If an ETT-added occurrence contains a run cycle that uses | | the period ETTRCY1, Tivoli Workload Scheduler for z/OS uses | | the input arrival time defined on that run cycle to | | establish successor dependencies instead of using the | | occurrence input arrival time, which by default is the | | triggering-event time." | | | |has to be changed with: | |"If an ETT-added occurrence contains a run cycle that uses | | the period ETTRCY1, Tivoli Workload Scheduler for z/OS adds | | the occurrence into the CP with an input arrival time | | matching with the current time, but it uses the time | | supplied by the not cyclic runcycle ETTRCY1 to establish | | successor dependencies instead of using the occurrence input | | arrival time" | | | +--------------------------------------------------------------+
Temporary fix
Comments
APAR Information
APAR number
PM19724
Reported component name
TIV WRKLD SCHD
Reported component ID
5697WSZ01
Reported release
302
Status
CLOSED DOC
PE
NoPE
HIPER
NoHIPER
Special Attention
NoSpecatt / Xsystem
Submitted date
2010-08-03
Closed date
2010-10-22
Last modified date
2016-04-22
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
| SC32126302 | SC32126303 | SC32126304 | SC32126305 | SC32126502 |
| SC32126503 | SC32126504 | SC32126505 | SC32126702 | SC32126703 |
Fix information
Applicable component levels
[{"Business Unit":{"code":"BU058","label":"IBM Infrastructure w\/TPS"},"Product":{"code":"SSRULV","label":"IBM Workload Scheduler for z\/OS"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"302","Edition":"","Line of Business":{"code":"LOB35","label":"Mainframe SW"}},{"Business Unit":{"code":"BU054","label":"Systems w\/TPS"},"Product":{"code":"SG19M","label":"APARs - z\/OS environment"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"302","Edition":"","Line of Business":{"code":"","label":""}}]
Document Information
Modified date:
22 April 2016