Nancy DiFilippo shares MVS Solutions' experiences with Automation, and how their customers are using ThruPut Manager AE+ to automate work load management, capacity management and production control. This presentation shows how AE+ extends all the functionality of ThruPut Manager AE to CA 7 workloads.
For more information on Mainframe solutions from CA Technologies, please visit: http://bit.ly/1wbiPkl
3. Automating Workload
3
How does AE+ automate workload ?
–Dynamically track and manage utilization
–Automatically balance workloads across available resources
Why Automate?
-Machine speed
-Predictable (policy driven)
-Dependable and repeatable
4. Batch Job Life Cycle
4
Reader Converter
QUEUE
Execution
JES/WLM Initiators
Submission
Initiation
Execution
CA Workload
Automation
CA 7® Edition
5. (Over) Utilization
5
Allowing more cars on this road only makes everyone arrive later
6. Understanding Utilization
6
0
5
10
15
20
25
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
55%
60%
65%
70%
75%
80%
85%
90%
95%
Percent Utilization
Time vs Utilization
Elapsed Time
Elapsed time grows exponentially with utilization. You will not complete very much work if everything takes twice as long to run!
7. Optimizing Utilization
TM AE+ adds batch load when and where it makes sense
–LPARs that have available capacity
–Service Classes that are performing well
–Results
•Jobs end earlier
•Other workloads continue to perform
7
8. Benchmark: WLM vs Automation
1000+ batch jobs over 10 hours using Workload Manager vs ThruPut Manager initiators
–CPU and I/O mix
Same environment
–Hardware
–z/OS level
–Service classes
Results:
–Automated Initiators complete much more work
–Automated Initiators uses fewer resources
8
9. 0
50
100
150
200
250
300
350
0
1
2
3
4
5
6
7
8
9
Jobs
Time (hours)
Benchmark: WLM vs TM AE concurrent Jobs
AE
WLM
AE jobs ahead
Complete More Work, use less Resources
9
10. Automated Workload Balancing
Workload balancing helps to avoid the effects of over utilization.
Automatic and Dynamic
–AE+ controls the number of available initiators on each system
–AE+ balances batch workload across each system in the JESplex
AE+ is well informed.
–Current utilization of all LPARS
–CEC, LPAR and Service Class performance reevaluated every 10 seconds
–The business specified priorities of all batch workloads
–System/resource affinities of all batch workloads
10
11. Automated Workload Balancing
AE+ rebalances batch workload as CPU demand and availability change
–Avoids overloading
–Considers capacity changes such as
•Capacity On Demand
•LPAR weight changes
•Soft capping.
11
13. Defined or Group Capacity Soft Cap
13
Soft caps don’t restrict demand while the R4HA is under the limit
14. Hitting the Wall
14
LPAR is capped while R4HA exceeds limit
Poor application performance due to instant overload
15. Managing Caps manually
15
Recovering; resetting back to normal; loop back to monitoring
Monitoring & reacting to workload demand & resource use
Potentially hours of fire fighting
16. 16
Managing with Automation
Caps can be lowered while protecting high priority workloads
19. The + for CA 7 Users
PCS (Production Control Services)
–Brings the scheduling goals of CA Workload Automation CA 7® Edition (CA 7) to z/OS execution
–Adds automation to critical path management and batch deadline scheduling to z/OS
–PCS dashboard simplifies production control operations
–Unique Closed Loop function
19
20. Why Automate CA 7?
Facilitates interaction between CA 7 and z/OS
–CA 7 is “hands off” after submission until completion
–z/OS selection/execution/prioritization logic is unaware of CA 7 goals and critical path
Production batch is complicated to manage manually
–Can be 100,000 + daily production jobs
–Operations personnel often has long tenure
–“PCS has become a repository of knowledge” Cliff Demoff, Humana
20
21. Batch Job Life Cycle With TM AE
21
Reader
Converter
TM Analysis
TM
XEQ
QUEUE
TM Job Selection
TM Initiators
TM Managed Execution
JES/WLM Initiators
Goals and Constraints
Submission
Initiation
Execution
CA Workload
Automation
CA 7® Edition
Execution
22. The Dashboard
22
------------------------------- TM/PCS Services ------------------------------ View/Edit Active Battle Plan Command ===> Production Cycle Name: WED Type: Day of Week (Wed) Description: Cycle Date/Time - From: Apr 17,2013 10:00 To: Apr 18,2013 09:59 Activated: Apr 17,2013 09:55 by: **Auto** ... from AMD, Volser: TMPCS.CA75.DEMO.AMD MDS002 Current Date/Time: Apr 17,2013 11:58:36 Time Remaining in Cycle: 22:01:24 -Job Instances- -Applications- -Job Trees- -Job Paths- Total Culprits Is Late: 2 3 3 7 3 Will be Late: 1 1 2 5 0 Was Late: 0 0 0 0 To-Do: 85 111 287 827 3 Done: 14 14 46 136 Total: 99 125 333 963 Adhoc: 0 For Detailed Battle Plan Information, Select one of the following: 1 List of All Applications 2 Job Instances... 3 List of Incomplete Culprit Job Instances X Exit
23. Integration with CA 7, z/OS
23
------------------------------- TM/PCS Services ------------------------------
View/Edit Active Battle Plan Job Instances
Command ===> Scroll ===> CSR
Cycle Date/Time - From: Apr 17,2013 10:00 To: Apr 18,2013 09:59
Current Date/Time: Apr 17,2013 12:23:46
Line Commands: S-Paths P-Detailed Paths X-Experience M-Monitor D-Delete
O-View/Edit Def./Options G-Progress I-Info(TM/UDF) A-Add to R-Remove from
List of All Job Instances in Battle Plan Line 1 of 963
Filter Status: Disabled Repeat: Next Sorted by: Rerun Factor/A
Job App Job Inst P CA-7 CA-7
- --Instance Id--- --Name-- Progress I -Status- --------Details--------- -Status- Job#
_ XMK0STR1_001 XMK 0 1 Done C0000
_ XMK1003_001 XMK C 2 1 Done C0000
_ XMK2003_001 XMK 0 1 Done C0000
_ XMK3END3_001 XMK 0 1 Active EXECUTING MVSD/XTRACT1 / ACT-Q 4333
_ XMK0STR1_001 XMK 0 1 Done C0000
_ XMK1003_001 XMK C 2 1 Ready Awaiting Execution RDY-S
_ XMK2003_001 XMK 1 1 Pending FCAST
_ XMK3END3_001 XMK 2 1 Pending FCAST
_ AAP0STR1_001 AAP 0 3 Done C0000
_ AAP1001_001 AAP 0 3 Done C0000
_ AAP2001_001 AAP 0 3 Active EXECUTING MVSD/UPDATE / ACT-Q 4452
_ AAP3END1_001 AAP 1 3 Pending FCAST
24. Job Instance Lists Colors Tell the Story
------------------------------- TM/PCS Services ------------------------------
View/Edit Active Battle Plan Job Instances
Command ===> Scroll ===> CSR
Cycle Date/Time - From: Apr 17,2013 10:00 To: Apr 18,2013 09:59
Current Date/Time: Apr 17,2013 12:02:38
Line Commands: S-Paths P-Detailed Paths X-Experience M-Monitor D-Delete
O-View/Edit Def./Options G-Progress I-Info(TM/UDF) A-Add to R-Remove from
List of All Job Instances in Battle Plan Line 1 of 963
Filter Status: Disabled Repeat: Next Sorted by: Rerun Factor/A
Job App CA-7 CA-7 # of M H Rp
- --Instance Id--- --Name-- --------Details--------- -Status- Job# Paths M T Nu
_ XMK2003_001 XMK FCAST 1 M
_ XMK3END3_001 XMK FCAST 1 T
_ XMK3END3_001 XMK FCAST 1 T
_ ZBK0STR1_001 ZBK C0000 6561 2 H
_ ZBK1001_001 ZBK C0000 6828 1 M
_ ZBK2001_001 ZBK C0000 6839 1 M
_ ZBK3END1_001 ZBK EXECUTING MVSD/XTRACT1 / ACT-Q 6854 1 T
_ ASI0STR1_001 ASI EXECUTING MVSD/STEPX / ACT-Q 6685 2 H
_ ASI1001_001 ASI FCAST 2 M
_ ASI2END2_001 ASI FCAST 1 T
_ ASI2END1_001 ASI FCAST 1 T
_ AAP0STR1_001 AAP EXECUTING MVSD/STEP3 / ACT-Q 6698 2 H
_ AAP1001_001 AAP FCAST 2 M
_ AAP2001_001 AAP FCAST 2 M
24
25. “Culprit” Jobs – Be Proactive
------------------------------ TM/PCS Services ------------------------------
View/Edit Active Battle Plan Job Instances
Command ===> Scroll ===> CSR
Cycle Date/Time - From: Apr 17,2013 10:00 To: Apr 18,2013 09:59
Current Date/Time: Apr 17,2013 11:59:32
Line Commands: S-Paths P-Detailed Paths X-Experience M-Monitor D-Delete
O-View/Edit Def./Options G-Progress I-Info(TM/UDF) A-Add to R-Remove from
List of Incomplete Culprit Job Instances Line 1 of 3
Filter Status: Ignored Repeat: All Sorted by: Rerun Factor/A
Job App C -Urgent Path- Job Inst P
- --Instance Id--- --Name-- P Rerun Slack Progress I -Status- --------Details
_ XMK1003_001 XMK C - 5.4 -01:11 C 2 1 Pending
_ XMK2003_001 XMK C - 5.4 -01:11 C 2 1 Pending
_ ZBK2001_001 ZBK - 0.4 -00:05 C 1 2 Active EXECUTING MVSD/
*******************************************************************************
-
25
26. ProGress – Analysis of Job’s Status
------------------------------- TM/PCS Services ------------------------------
Job Instance Progress Analysis
Command ===>
Time of G Line Cmd: Apr 17 12:20:44 (Culprit, IS Late)
Job Instance Name: XMK1003_001 (5812) Status: Pending
Application Name: XMK CA-7 Status: REQ-Q
-All- To-Do
Number of Preceding Jobs: 1 0 | Tab cursor to the
Number of Preceding Culprit Jobs: 0 0 | number and press ENTER
Number of Affected Jobs: 2 2 | to display list of jobs.
Why this Job Instance ...
1 has an Acceptable time of Apr 17,2013 10:51 ?
2 is a Culprit ?
3 is Late ?
4 has Not Completed ?
To Return to the previous panel, press END.
26
27. Direct CA 7 Communications
------------------------------- TM/PCS Services ------------------------------
Why This Job Instance has NOT Completed
Command ===>
Time of G Line Cmd: Apr 17 12:20:44 (Culprit, IS Late)
Job Instance Name: XMK1003_001 (5812) Status: Pending
Application Name: XMK CA-7 Status: REQ-Q
This Job Instance (CA 7 # 5812) appears to be in the REQ Queue.
According to PCS, all JOB requirements have been satisfied for
the Job. However, it is possible that the job has other requirements
(such as Manual Verification) or there are Job Requirements not
known to PCS.
For additional and up-to-date information, use the following commands:
1 LQ,LIST=RQMT,JOB=...
2 LJOB,LIST=ALL,JOB=...
3 LJOB,LIST=NODD,JOB=...
4 LRLOG,SPAN=48,SEQ=DATE,JOB=...
5 LDTM,LIST=ALL,JOB=...
* TM/UDF Info for the Job Instance
To return to the previous panel, press END.
27
28. Dynamically Updated Status CA 7 and z/OS Information Together
------------------------------- TM/PCS Services ------------------------------
View/Edit Active Battle Plan Job Instances
Command ===> Scroll ===> CSR
Cycle Date/Time - From: Apr 17,2013 10:00 To: Apr 18,2013 09:59
Current Date/Time: Apr 17,2013 12:23:46
Line Commands: S-Paths P-Detailed Paths X-Experience M-Monitor D-Delete
O-View/Edit Def./Options G-Progress I-Info(TM/UDF) A-Add to R-Remove from
List of All Job Instances in Battle Plan Line 1 of 963
Filter Status: Disabled Repeat: Next Sorted by: Rerun Factor/A
Job App Job Inst P CA-7
- --Instance Id--- --Name-- Progress I -Status- --------Details--------- -Statu
_ XMK0STR1_001 XMK 0 1 Done C0000
_ XMK1003_001 XMK C 2 1 Done C0000
_ XMK2003_001 XMK 0 1 Done C0000
_ XMK3END3_001 XMK 0 1 Active EXECUTING MVSD/XTRACT1 / ACT-Q
_ XMK0STR1_001 XMK 0 1 Done C0000
_ XMK1003_001 XMK C 2 1 Ready Awaiting Execution RDY-S
_ XMK2003_001 XMK 1 1 Pending FCAST
_ XMK3END3_001 XMK 2 1 Pending FCAST
_ AAP0STR1_001 AAP 0 3 Done C0000
_ AAP1001_001 AAP 0 3 Done C0000
_ AAP2001_001 AAP 0 3 Active EXECUTING MVSD/UPDATE / ACT-Q
_ AAP3END1_001 AAP 1 3 Pending FCAST
_ LAP0STR1_001 LAP 0 3 Pending FCAST
_ LAP1001_001 LAP 1 3 Pending FCAST
_ LAP2001_001 LAP 2 3 Pending FCAST
28
29. Production Control Can Influence Job Selection
+------------------------------ TM/PCS Services ------------------------------+
| View/Edit Defaults/Options for Selected Job Instance |
| Command ===> |
| |
| Job Instance: |
| Production Cycle Name: WED |
| Description: |
| Job Instance Id: HMK3END3_001 |
| |
| Job Instance Defaults: |
| Production Importance: D (1 derived from Application for All Cycles L2) |
| Generate Job Summary: D (Y derived from All Apps for All Cycles L1) |
| Early Analysis/Setup: D (N derived from All Apps for All Cycles L1) |
| hh:mm |
| Culprit Job Time: |
| Acceptable Time: 14:00 / Apr 17 Apr 18 . Apr 19 |
| |
| Job Instance Event Management: |
| - Alerts (0) | Tab cursor and press ENTER |
| |
| Press END to accept changes or CANCEL to discard. |
+-----------------------------------------------------------------------------+
Path: HMK0STR1_001_HMK4END2_001
_ HMK0STR1_001 HMK C 0.5 00:11 0 5 1 Pending
_ HMK1002_001 HMK 0.5 00:11 1 1 1 Pending
_ HMK2002_001 HMK 0.5 00:11 2 1 1 Pending
_ HMK4END2_001 HMK 0.5 00:11 4 1 1 Pending
Path: HMK0STR1_001_HMK3END3_001
29
32. Summary
Automating Workload Balancing:
–Managing batch workload placement with its dynamic initiators
–Managing job selection priority based on importance and time constraints information ultimately CA 7
Automating Capacity Management:
–Managing R4HR to reduce software costs with sub-capacity pricing & lower soft cap limits
Automating Production Control
–Dynamic adjustment of job queue placement based on job urgency
–Operations can monitor and effect change directly from dialog
32
33. For More Information
To learn more about Mainframe, please visit: http://bit.ly/1wbiPkl
Insert appropriate screenshot and text overlay from following “More Info Graphics” slide here; ensure it links to correct page
Mainframe