How do I determine why SLM is not selecting jobs on one or more members of a JESplex (MAS)?
What sort of documentation should I collect before opening a case if the underlying issue(s) cannot be determined?

When SLM is not selecting batch jobs as expected, check the SDSF or JES2 STATUS field for a customized ThruPut Manager Rejection Reason Message (RRM).

One of the most common RRM codes reported by customers is:

• -SLM- Srv Cls Perf - Job cannot be selected on this system due to the performance of the Service Class.

If the job(s) in the above scenario has affinity to only the system experiencing poor Service Class performance, then no other system in the JESplex (MAS) is eligible to select the job and the job will wait in the JES2 input queue until the performance of the Service Class improves. SLM checks the queue and system conditions every 10 seconds. System affinity can be affected by JOBPARM SYSAFF, the SYSTEM keyword on the JOB statement, or JLS limiting or JBS binding agents that have been added to the job.

Note: An SLM Performance Focus Group can be defined to be more tolerant of Service Class performance.

• Performance focus definitions for ThruPut Manager AE

SLM commands for checking system status:

To determine why SLM is not selecting jobs, start by issuing the "SLM DISPLAY" or "SLM SELECT" commands to confirm if any operator constraints have been imposed via the "SLM SELECT" command:

/SLM D                                             
DTM8213I SLM SUMMARY                               
POLICY                          : TST715XY
SERVICE MODE                    : STANDARD         
ACTIVE PERFORMANCE FOCUS        : BALANCED         
SLM SERVER ON MEMBER            : P1               
GENERAL SERVICES CONSTRAINTS    : NONE
BATCH SERVICE                   : INSUFFICIENT DATA
PRODUCTION SERVICES CONSTRAINTS : NONE          
PRODUCTION SERVICES OPTION      : PREEMPT          
AUTOMATED CAPACITY MANAGEMENT   : NOT CONFIGURED   
MINIMUM ACM CAPACITY LEVEL      : NOT SET          
ONLINE PERFORMANCE PROTECTION   : NOT CONFIGURED  

/SLM SELECT                             
DTM8141I NO OPERATOR CONSTRAINTS APPLIED

Check the "Constraints" field of the "SLM DISPLAY JESPLEX" command to confirm there are no system related constraints:

/SLM D JPX                                                            
DTM8219I SLM JESplex Status                                           
Member: P1     State: SLM Server Active                               
  CPC       Type: 1090  Serial: 02-0DACB                              
  Constraints:                                                        
    Color(Green)
Member: P2     State: SLM Active                                      
  CPC       Type: 1090  Serial: 02-0DACB                              
  Constraints:                                                        
    Color(Green)

The "SLM DISPLAY SC ALL" command will provide the local system status for each WLM Service Class used by SLM:
                                                     
/SLM D SC ALL                                                         
DTM8233I SLM Service Class Display                                    
SLMIMP    G PI:N/A  Dly:N/A   ADly:N/A  U:N/A  ACPU:91%  AS:0.0    P:1
SLMPRIO   G PI:N/A  Dly:N/A   ADly:N/A  U:N/A  ACPU:91%  AS:0.0    P:1
SLMSTD    G PI:N/A  Dly:N/A   ADly:N/A  U:N/A  ACPU:91%  AS:0.0    P:1

If there are no SLM, System, or WLM Service Class constraints, it could be that the job has affinity to a particular system and the job cannot be selected there for some other reason like a JLS limiting or JBS binding agent.

The $DJnnnnnn,LONG command provides standard JES2 $HASP890 message output with an extended STATUS field containing the ThruPut Manager Rejection Reason Message (RRM) and also further ThruPut Manager display output showing RRM codes for each member of the JESplex (MAS).

In the example below:

• The JESplex (MAS) has 2 systems: P1 and P2;
• The ThruPut Manager General Services (GS) class is set to JOBCLASS(2);
• The queue affinity (QAFF) for JOBCLASS(2) is restricted to P1;
• J877631 has SYSAFF=(P2);
• So the job cannot be selected on either system for the following reasons:
  • P1   System affinity                                      
  • P2   Job class affinity

See the displays below for more details:

/TM CLASS                                                            
DTM3233I TM CLASS LIST                                               
Analysis............ 9                                               
Deferred............ D                                               
Selectable.......... A,B,C,D,E,G,H,I,J,K,L,M,N,O,P,R,S,T,TMANALYZ    
                     TMDEFR,TMDFLT,TMDMND,TMGS,TMJAL,TMPS,U,V,W,X,Y,Z
                     0,1,2,3,4,5,6,7,8,9                             
Exempt.............. F,Q                                             
On_Demand........... TMDMND                                          
Production_Services. 1                                               
General_Services.... 2                                               
Default............. TMDFLT      

$DJOBCLASS(2),LONG
$HASP837 JOBCLASS(2)         ACTIVE=YES,ACCT=NO,AUTH=(ALL),     
$HASP837                     BLP=YES,COMMAND=VERIFY,COPY=NO,    
$HASP837                     DSENQSHR=ALLOW,DUPL_JOB=DELAY,     
$HASP837                     GDGBIAS=JOB,GROUP=,HOLD=NO,        
$HASP837                     IEFUJP=YES,IEFUSO=YES,             
$HASP837                     JESLOG=(NOSPIN),JOBRC=MAXRC,       
$HASP837                     JOURNAL=YES,LOG=YES,MODE=JES,      
$HASP837                     MSGLEVEL=(1,1),OUTDISP=(HOLD,HOLD),
$HASP837                     OUTPUT=YES,PERFORM=000,            
$HASP837                     PROMO_RATE=0,PGMRNAME=NO,          
$HASP837                     PROCLIB=00,QAFF=(P1),QHELD=NO,  
$HASP837                     RESTART=NO,REGION=0001M,SCAN=NO,   
$HASP837                     SCHENV=,SWA=ABOVE,SYSSYM=ALLOW,    
$HASP837                     TIME=(001440,00),TYPE26=YES,       
$HASP837                     TYPE6=YES,XBM=,XEQCOUNT=(MAXIMUM=*,
$HASP837                     CURRENT=0),                        
$HASP837                     XEQMEMBER(P1)=(MAXIMUM=*,          
$HASP837                     CURRENT=0),                        
$HASP837                     XEQMEMBER(P2)=(MAXIMUM=*,          
$HASP837                     CURRENT=0),DESC=

$DJ877631,LONG                                                       
$HASP890 JOB(WTOR$P2)                                                
$HASP890 JOB(WTOR$P2)   STATUS=(DELAY-Class Affin.),CLASS=2,         
$HASP890                PRIORITY=2,SYSAFF=(P2),HOLD=(NONE),          
$HASP890                CMDAUTH=(LOCAL),OFFS=(),SECLABEL=,           
$HASP890                USERID=RSI01CD,SPOOL=(VOLUMES=(ZPSPL1),      
$HASP890                TGS=1,PERCENT=0.0010),ARM_ELEMENT=NO,        
$HASP890                CARDS=3,DUBIOUS=NO,PRIVILEGE=NO,DJC=NO,      
$HASP890                REBUILD=NO,SRVCLASS=SLMSTD,SCHENV=,          
$HASP890                SCHENV_AFF=(ANY),CC=(),DELAY=(JOBCLASS),     
$HASP890                CRTIME=(2022.122,15:25:41),JOBGROUP=,        
$HASP890                BERTNUM=2,JOENUM=0,JOEBERTS=0                

DTM0220I Job Display                                                 
J0877631 WTOR$P2  Awaiting Execution  For 0 Mins                     
                  Class   : 2         (General_Services)             
                  Priority: 2                                        
                  Sysaff  : P2                                       
Managed:  SLM                                                        
          JLS                                                        
Time     System Reason job was not selected                          
11:25:59 * P1   System affinity                                      
N/A        P2   Job class affinity   
                               
DTM6420I JLS DISPLAY                                                 
J0877631    WTOR$P2 AWAITING EXEC                                    
+LIMIT.ALLJOBS   - LIMIT=999/999 REF=11 ACT=0 WEIGHT=1  PANEL=JLSSAMP

DTM8226I SLM Job Display                                             
J0877631 WTOR$P2  Awaiting Execution                                 
Time     System Reason job was not selected                          
11:25:59 * P1   System affinity                                      
N/A        P2   Job class affinity                                   
Non SLM jobclass:      A                                             
Non SLM service class: BATCH3                                        
Workload Type : General Services                                     
Control Center: DEV      (Development Group)                         
Type          : STANDARD (Normal Workload)                           
Batch Service : Insufficient data                                    
Trend         : Insufficient data                                    
Service Mode  : Standard                                             
Importance    : 4                                                    
Job Service   : Before Target                                        
Estimated Time to Selection: Indeterminate                           
   Reason Estimated Time to Selection could not be calculated.       
   - Service level information is not available.         

Documentation to collect for diagnosing ThruPut Manager selection issues for SLM-managed jobs:

• TM D LEVEL – displays ThruPut Manager PTF level and a list of applied APARs.
• TM CLASS DISPLAY – displays ThruPut Manager class assignments for On_Demand, Production_Services (PS), and General_Services (GS).
• $DJnnnnnn,LONG – issued for one or more affected jobs while they are waiting for execution will confirm the affected WLM Service Class and generate an SLM display output showing Rejection Reason Messages (RRM) for each member of the JESplex.
• $DJOBDEF – display will provide details on duplicate job name checking.
• SLM D – displays server details and the active Performance Focus Definition.
• SLM D JESPLEX – displays SLM status for all members of the JESplex.
• SLM D SC ALL – this will display the performance of all service classes for the JES2 member where the command is issued.
• SLM SPIN TRACE TIME(ddd.hh:mm:ss,nn) SYSOUT – this will spin a trace of SLM activities to the TMSS started task.
• ADS F SLMTRACE TIME(ddd.hh:mm:ss,nn) SYSOUT – this will trace SLM Initiators (DTM8210I message output).
• ADS F GTRACE TIME(ddd.hh:mm:ss,nn) SYSOUT – this will produce a general trace of TM activities, including those by SLM.
• A list of the affected WLM Service Classes.
• WLM Policy Report.
• SLM Policy Report - can be generated to a pre-allocated dataset (RECFM=FBA,LRECL=121) using the TMISPF dialog (place a 'G' next to the active policy).
• SMF type 70,72 and TM SMF record subtype 17 for 1 hour before and after the event period:
  • Issue the TM SMF command to determine the record type set by the TM SMF statement in the TMSS initialization parmlib member (TM D PARMS TMSS).
• SLM START nnnnnn – can be issued to force SLM to start a job, regardless of performance concerns, but the job may still queue for CPU resources.