| View previous topic :: View next topic | 
	
	
		| Author | Message | 
	
		| Ajay Jain Beginner
 
 
 Joined: 07 Jan 2003
 Posts: 1
 Topics: 1
 
 
 | 
			
				|  Posted: Tue Jan 07, 2003 3:12 am    Post subject: How to know CPU time of abend jobs |   |  
				| 
 |  
				| We have been getting huge data center bills. We have tried to analyse the data for the past one month and have found a lot of abends. Using ABEND-AID we found the reasons for the abend also as well as other data like what transaction, module etc caused the abend. Is there any way we can find out the CPU time that elapsed for a particular abend so that we can calculate how much each abend has cost us in terms of CPU time usage and then we can focus to resolve the same????. We also has TMONF2 (Omegamon Utilities) to monitor CICS regions.
 |  | 
	
		| Back to top |  | 
	
		|  | 
	
		| Bill Dennis Advanced
 
  
 Joined: 03 Dec 2002
 Posts: 579
 Topics: 1
 Location: Iowa, USA
 
 | 
			
				|  Posted: Tue Jan 07, 2003 9:00 am    Post subject: |   |  
				| 
 |  
				| Where does the data you are analyzing come from?  Does the data center provide it?  If they provide detail on each job (including abends), I would expect that they could provide CPU time to you as well. 
 Otherwise, you will need some type of history file, such as SMF records from the systems.
 
 Bill
 |  | 
	
		| Back to top |  | 
	
		|  | 
	
		| kolusu Site Admin
 
  
 
 Joined: 26 Nov 2002
 Posts: 12394
 Topics: 75
 Location: San Jose
 
 | 
			
				|  Posted: Tue Jan 07, 2003 9:14 am    Post subject: |   |  
				| 
 |  
				| Ajay Jain, 
 
 Can't you just look at the SYSMSG from the job output?? I believe that it has the stastics of each and every step execueted as well as a cummulative cpu and elapsed time of each and every job.It would be some this like this :
 
 
  	  | Code: |  	  | ********************************************************************************
 *                                                  STEP TERMINATION STATISTICS
 *
 * STEP NAME     STEP10      START TIME  04:00:40.41   VIRT SYS  USED      452K
 * PGM NAME      DFSRRC00    STOP  TIME  04:19:48.42   VIRT CORE USED    2,308K
 * DISP. PRTY      0         ELAP. TIME  00:19:08.01   NO. OF SWAPS           1
 * PERF. GROUP     0         SERV UNITS    2,339,852   TRANS ACT TIME  00:19:05.7
 *  MAIN STORAGE UNITS     300,789     CPU UNITS       2,037,266     SRB UNITS
 ********************************************************************************
 
 
 ******************************************************
 *
 *
 NO. OF PAGE-INS          0   STEP CPU  00:05:39.05 *
 NO. OF PAGE-OUTS         0   JOB  CPU  00:05:39.41 *
 PAGES SWAPPED-OUT       68   SRB TIME  00:00:00.25 *
 .71                        COMPLETION CODE     U3820 *
 1,117     I/O UNITS             680        *
 ******************************************************
 
 
 | 
 
 Look at the second flowered box.
 
 It has both step and job cpu times and completion code of the job.
 
 
 Hope this helps...
 
 cheers
 
 kolusu
 |  | 
	
		| Back to top |  | 
	
		|  | 
	
		| DaveyC Moderator
 
  
 Joined: 02 Dec 2002
 Posts: 151
 Topics: 3
 Location: Perth, Western Australia
 
 | 
			
				|  Posted: Fri Jan 10, 2003 9:09 am    Post subject: |   |  
				| 
 |  
				| Kolusu, 
 That is not standard MVS. What you have shown is the SMF exit used in your site. IBM have shipped a default as of OS/390 2.x but it's not well established yet. You have to check the JES output for CPU/SRB/EXCP metrics.
 _________________
 Dave Crayford
 |  | 
	
		| Back to top |  | 
	
		|  | 
	
		| Bithead Advanced
 
  
 Joined: 03 Jan 2003
 Posts: 550
 Topics: 23
 Location: Michigan, USA
 
 | 
			
				|  Posted: Fri Jan 10, 2003 12:07 pm    Post subject: |   |  
				| 
 |  
				| This looks like a problem in a CICS region so the MVS statistics suggested will probably not be available. |  | 
	
		| Back to top |  | 
	
		|  | 
	
		|  | 
	
		|  |