Difference between revisions of "Information Systems:IBM i System Jobs (WRKACTJOB)"
(Created page with "==Jobs== There are basically two types of user jobs on the iSeries – interactive and batch. An interactive job is one that runs on a workstation, and requires interaction...") |
m |
||
| (4 intermediate revisions by 2 users not shown) | |||
| Line 114: | Line 114: | ||
SMTP QNOTES BCI .0 PGM-SMTP SELW |
SMTP QNOTES BCI .0 PGM-SMTP SELW |
||
UPDATE QNOTES BCI .0 PGM-UPDATE SELW |
UPDATE QNOTES BCI .0 PGM-UPDATE SELW |
||
| + | |||
| + | <div id="monitor jobs"></div> |
||
These are the jobs to do with electronic order processing. EMASUBPRD runs a job that ‘wakes up’ every hour and checks for low order volumes. The next four subsystems run the processing for inbound messages, inbound message parsing, inbound order processing, and outbound messages. These jobs also snooze for a while, then wake up to process any available records. |
These are the jobs to do with electronic order processing. EMASUBPRD runs a job that ‘wakes up’ every hour and checks for low order volumes. The next four subsystems run the processing for inbound messages, inbound message parsing, inbound order processing, and outbound messages. These jobs also snooze for a while, then wake up to process any available records. |
||
| Line 260: | Line 262: | ||
DPRQ3_VA ASW4OWN BCH .0 PGM-ASGC900 DEQW |
DPRQ3_VA ASW4OWN BCH .0 PGM-ASGC900 DEQW |
||
ORDFLOW_PL ASW4OWN BCH .0 PGM-ASGC900 DEQW |
ORDFLOW_PL ASW4OWN BCH .0 PGM-ASGC900 DEQW |
||
| − | ORDFLOW_VA ASW4OWN BCH .0 PGM-ASGC900 DEQW |
+ | ORDFLOW_VA ASW4OWN BCH .0 PGM-ASGC900 DEQW |
===Cancelling a Job=== |
===Cancelling a Job=== |
||
| Line 280: | Line 282: | ||
In the ‘How to end’ field, change *CNTRLD (controlled) to *IMMED (immediate). If you make the mistake of pressing Enter instead of F4, the job will go right into a controlled end. You can still go back and repeat this procedure, remembering this time to press F4. |
In the ‘How to end’ field, change *CNTRLD (controlled) to *IMMED (immediate). If you make the mistake of pressing Enter instead of F4, the job will go right into a controlled end. You can still go back and repeat this procedure, remembering this time to press F4. |
||
| + | |||
| + | ====End Job Abnormal==== |
||
In extreme cases, this might not work - the job will just continue running. In that case, an end job abnormal (ENDJOBABN) can be run. Note that it can only be run ten minutes after an endjob *immed has been run. |
In extreme cases, this might not work - the job will just continue running. In that case, an end job abnormal (ENDJOBABN) can be run. Note that it can only be run ten minutes after an endjob *immed has been run. |
||
| Line 298: | Line 302: | ||
F3=Exit F4=Prompt F5=Refresh F12=Cancel F13=How to use this display |
F3=Exit F4=Prompt F5=Refresh F12=Cancel F13=How to use this display |
||
F24=More keys |
F24=More keys |
||
| + | |||
| + | [[Category: IBM i]] |
||
Latest revision as of 12:02, 15 July 2016
Jobs
There are basically two types of user jobs on the iSeries – interactive and batch. An interactive job is one that runs on a workstation, and requires interaction with a user. A batch job involves no user interaction, but requires more disk access and processor time. There is no reason for it to tie up a workstation.
The two types are handled differently by the system. Interactive jobs have a higher priority, but a smaller time slice. When the user presses a key, the system responses immediately and performs the request; which typically can be completed in the short time allowed. Then while it waits for the user to press another key, other jobs are serviced. If a job that should be batch (ie user interaction is not required, but there is a lot of disk access) is run interactively, the higher priority could degrade overall system performance, while the short time slice will not allow the job to run efficiently. (I once had a query that took 20 minutes to run interactively, and 2 minutes to run in batch.)
Note that while ASW jobs typically submit to batch when you press enter, you can press F2 to run them interactively. The advantages are that they will run immediately instead of lining up to run in batch, and that if the job is having problems, there will be a job log to review.
Job Queues
Key in the command WRKJOBQ to see the job queues, or the ‘lineup’ of jobs waiting for batch processing. These are jobs that have not yet started running.
Work with All Job Queues
Type options, press Enter.
2=Change 3=Hold 4=Delete 5=Work with 6=Release
8=Work with job schedule entries 9=Description 14=Clear
Opt Queue Library Jobs Subsystem Status
_ EOXBATCH QGPL 0 QBATCH RLS
_ IMPJOBQ QGPL 0 RLS
_ QBASE QGPL 0 RLS
_ QBATCH QGPL 0 QBATCH RLS
_ QBATCH1 QGPL 0 QBATCH RLS
_ QBATCH2 QGPL 0 QBATCH RLS
_ QBATCH3 QGPL 0 QBATCH RLS
_ QFNC QGPL 0 RLS
_ QINTER QGPL 0 QINTER RLS
_ QPGMR QGPL 0 QPGMR RLS
_ QSNADS QGPL 0 QSNADS RLS
_ QSPL QGPL 0 QSPL RLS
_ QSTATEMENT QGPL 0 QBATCH RLS
_ QS36EVOKE QGPL 0 QBATCH RLS
_ QS36MRT QGPL 0 QINTER RLS
_ QTXTSRCH QGPL 0 QBATCH RLS
_ QTCP QTCP 0 RLS
- DOMINO00 QUSRNOTES 0 DOMINO00 RLS
_ EMAIMPPRD UNIPHARMIT 54 EMASUBPRD RLS
The queues to worry about are QBATCH and QBATCH3. These are the line-ups of jobs submitted by users. If jobs start accumulating here, it could indicate a problem in the QBATCH subsystem – or that there are a couple of long running jobs running (each of these queues will submit two jobs at a time). This can happen just after the end of the month.
Job queue EMAIMPPRD will get a line-up of submitted jobs because of the way it is started every morning. These can be cancelled.
Active Jobs
To see all jobs that are currently running, key in the command WRKACTJOB.
Keep an eye on overall CPU % (5 to 20% is usual during the day). Press F5 to recalculate the usage over the length of time elapsed. Be aware of allowing the elapsed time to get too long, as the CPU % will start becoming more of an average than an actual figure. Press F10 to restart the elapsed time.
Each job also has a field indicating CPU %. For interactive jobs it is normally between 0 and 9% - it can peak as high as 60% but only for very short periods of time (press F10 to recalculate). If this percentage stays high, it could be a ‘runaway job’ (either a program that is in a loop, or a job that is no longer connected to a workstation, but is desperately trying to reconnect to it), or it could mean that the user is running a job interactively that should be run in batch. (In ASW pressing F2 will run a job online instead of submitting it to batch.) A batch job running online should be left to complete normally; a runaway job must be cancelled.
Individual batch jobs can use as much as 50 to 70% of the CPU without degrading performance. The QBATCH subsystem is tuned in such a way as to use all unused processor power. However, it has a lower run priority than QINTER, and so will defer to interactive jobs requiring processor time.
If the overall CPU % is high, the easiest way to find the culprit is to sort on the jobs’ percentages. Place the cursor in the CPU % column and press F16 (upper shift F4) to sort on this column from high to low. This will work for any column on the display – it does not work like windows; pressing F16 a second time will not reverse the sort sequence. Positioning the cursor under the ‘S’ or ‘u’ in Subsystem/Job will sort by subsystem then job; positioning it under any other letter will sort by job only.
Subsystems and Jobs
The DCRF subsystem shows all the RF guns that have a user signed on. (Interesting note – RFOSIGNON is for users who are only allowed to sign on to an RF gun. RFSIGNON will detect if the user is on an RF gun or a green screen, and go to either DCRF or QINTER, and run the required program.)
The advantage in having a separate subsystem, is that we can shut down picking in the warehouse, without affecting users in the office.
Work with Active Jobs BART
06/23/05 16:01:31
CPU %: 11.2 Elapsed time: 00:00:11 Active jobs: 446
Type options, press Enter.
2=Change 3=Hold 4=End 5=Work with 6=Release 7=Display message
8=Work with spooled files 13=Disconnect ...
Opt Subsystem/Job User Type CPU % Function Status
DCRF QSYS SBS .0 DEQW
RFTER0002A LEWISJ INT .0 PGM-RFSIGNON DSPW
RFTER0003A NATALIYAP INT .0 PGM-RFOSIGNON DSPW
RFTER0006A ALEXG INT .0 PGM-RFOSIGNON DSPW
RFTER0007A GORDONM INT .0 PGM-RFSIGNON DSPW
RFTER0008A MICHAELL INT .0 PGM-RFOSIGNON DSPW
RFTER0010A BIDOS INT .0 PGM-RFOSIGNON DSPW
RFTER0011A ERICC INT .0 PGM-RFOSIGNON DSPW
RFTER0012A VICTORIAM INT .0 PGM-RFOSIGNON DSPW
RFTER0013A JAMESS INT .0 PGM-RFOSIGNON DSPW
RFTER0014A MITAP INT .0 PGM-RFOSIGNON DSPW
RFTER0015A ELENAP INT .0 PGM-RFOSIGNON DSPW
RFTER0018A ANDYT INT .0 PGM-RFOSIGNON DSPW
RFTER0019A MELQUIADEM INT 3.6 PGM-RFOSIGNON DSPW
RFTER0020A JESUSAS INT .0 PGM-RFOSIGNON DSPW
RFTER0022A PARMJITG INT .0 PGM-RFOSIGNON DSPW
RFTER0024A MOHAMMEDR INT .0 PGM-RFOSIGNON DSPW
RFTER0026A INDERJITD INT .0 PGM-RFOSIGNON DSPW
RFTER0027A RENATON INT .0 PGM-RFOSIGNON DSPW
RFTER0029A PAULOV INT .0 PGM-RFOSIGNON DSPW
RFTER0030A CHOOSY INT .2 PGM-RFOSIGNON DSPW
The DOMINO subsystem shows the jobs to do with Lotus Notes.
DOMINO00 QSYS SBS .0 DEQW
ADMINP QNOTES BCI .6 PGM-ADMINP SELW
AMGR QNOTES BCI .0 PGM-AMGR SELW
AMGR QNOTES BCI .0 PGM-AMGR SELW
AMGR QNOTES BCI .0 PGM-AMGR CNDW
AMGR QNOTES BCI .0 PGM-AMGR CNDW
CALCONN QNOTES BCI .0 PGM-CALCONN CNDW
COLSRV400 QNOTES BCI .0 PGM-COLSRV400 SELW
CONTROLLER QNOTES BCI .0 PGM-CONTROLLER JVAW
DECS QNOTES BCI .0 PGM-DECS SELW
DIRCAT QNOTES BCI .0 PGM-DIRCAT SELW
EVENT QNOTES BCI .4 PGM-EVENT SELW
HTTP QNOTES BCI 1.1 PGM-HTTP SELW
MTC QNOTES BCI .0 PGM-MTC SELW
QJVAEXEC QNOTES BCI 1.0 PGM-SERVER SELW
REPLICA QNOTES BCI .0 PGM-REPLICA SELW
ROUTER QNOTES BCI .0 PGM-ROUTER SELW
SCHED QNOTES BCI .0 PGM-SCHED SELW
SMTP QNOTES BCI .0 PGM-SMTP SELW
UPDATE QNOTES BCI .0 PGM-UPDATE SELW
These are the jobs to do with electronic order processing. EMASUBPRD runs a job that ‘wakes up’ every hour and checks for low order volumes. The next four subsystems run the processing for inbound messages, inbound message parsing, inbound order processing, and outbound messages. These jobs also snooze for a while, then wake up to process any available records.
‘Delay’ can function in two different ways – ‘16:59:00’ means to wake up at one minute before 5PM, and ‘120’ means to wake up in two minutes or 120 seconds.
EMASUBPRD QSYS SBS .0 DEQW
EMAIOPPRD EODJOB BCH .0 DLY-16:59:00 DLYW
IMPSUBPRD QSYS SBS .0 DEQW
IMPPOSTPRD EODJOB BCH .0 DLY-120 DLYW
IMSSUBPRD QSYS SBS .0 DEQW
IMSPOSTPRD EODJOB BCH .0 DLY-120 DLYW
IOPSUBPRD QSYS SBS .0 DEQW
IOPPOSTPRD EODJOB BCH .0 DLY-120 DLYW
OMSSUBPRD QSYS SBS .0 DEQW
OMSPOSTPRD EODJOB BCH .0 DLY-300 DLYW
‘Exodus’ communicates with PC’s running BCD software. ‘Catapult’ and ‘Websmart’ are both using different versions of Exodus. This will be fixed when we upgrade Catapult. (Note – WebSmart is the tool we use to generate free form RPG programs to run in a browser, and Catapult takes i output and generates a PDF file.)
EXODUS QSYS SBS .0 DEQW
EXSERVER EXODUS ASJ .0 TIMW
EXODUS51 QSYS SBS .0 DEQW
EXSERV51 EXODUS51 ASJ .0 SELW
QCMN QSYS SBS .0 DEQW
EXODUS CATPULTADM EVK .0 DEQW
EXODUS CATPULTADM EVK .0 SELW
The jobs in the monitor subsystems are started by the end of day, and run all day.
MSGMONITOR keeps an eye on the QSYSOPR message queue, and forwards any unexpected messages to the addresses in data areas MSGREC to MSGREC5 (in library UNIPHARMIT).
ATLAS_INT checks the FTP server for scan batches from Atlas Courier. When it finds them, it updates the shipment archive.
EDIMON_PRD will send any EDI files that are ready, and will look to see if SPS Commerce has any files ready for us to pick up and process.
FTP_ORDERS checks the FTP server to see if any stores have sent us orders.
ORDALLOCFL checks to see if stock has come in for any future orders, or special orders. If a regularly stocked item has gone out of stock, and there is a future order, this program will set the item so that when the stock does come in, the future orders will be filled first. It used to be possible for a regular order to come in and scoop the stock.
MONITOR QSYS SBS .0 DEQW
MSGMONITOR EODJOB BCH .0 PGM-OPMESSAGES MSGW
MONITOR2 QSYS SBS .0 DEQW
ATLAS_INT ATLAS BCH .0 DLY-20 DLYW
EDIMON_PRD EDIUSER BCH .0 PGM-EDCUP040 MSGW
FTP_ORDERS EODJOB BCH .0 PGM-AICUP010 MSGW
ORDALLOCFL EODJOB BCH .0 DLY-13:30:00 DLYW
QBATCH runs normal batch jobs that users have submitted. Up to two jobs will run at the same time from each job queue – after that they will line up (in QBATCH or QBATCH3 job queues) and wait for their turn. (The subsystem and the job queue in this case have the same name, but it doesn’t have to be that way.)
PUR_ORDER jobs may cause problems. Only one can run at a time, so when it starts it checks to see if another version is already running. If so, the second version will delay and wait for the first to end. If the first does not end because it is in error, the second one will stay in delay. If a third PO print jobs starts it will also go into delay and the QBATCH subsystem will be blocked until someone notices the problem and corrects it. Meanwhile all submitted jobs will line up in the QBATCH and QBATCH3 job queues.
If no one corrects this, at end of day the three active jobs will be cancelled when the subsystem is ended (the only harm will be the PO will not be faxed or emailed). Any jobs lined up (but not started) will run when the subsystem is restarted the next morning.
QBATCH QSYS SBS .0 DEQW
RUNQRY SHEILAV BCH .0 PGM-ASGC900 RUN
PUR_ORDER SCOTTZ BCH .0 PGM-ASGC900 RUN
UPD_UD_SST SHEILAV BCH 25.1 PGM-ASGC900 RUN
We have three web servers –
WEBINTER for InfoNet, which is our browser based system for internal use
WEBSMART which is the Web Order system for our shareholder and customer use.
WEBTEST which is our test system.
There is another server for administration. We don’t use it much, but since version 6.1, when Domino starts, all TCP/IP servers are started at the same time.
These are Apache servers. See the section ‘WebSmart / Apache HTTP (Web) Servers’ for more information.
The jobs running QZSRCGI (Common Gateway Interface is a standard method used to generate dynamic content for web applications) call the programs that have been written with WebSmart. Multiple iterations can be running: as many as are needed to fill requests.
QHTTPSVR QSYS SBS .0 DEQW
ADMIN QLWISVR BCI .0 JVM-com.ibm.lw THDW
ADMIN QTMHHTTP BCH .0 PGM-QZHBMAIN SIGW
ADMIN QTMHHTTP BCI .0 PGM-QZSRLOG SIGW
ADMIN QTMHHTTP BCI .0 PGM-QZSRHTTP SIGW
ADMIN2 QLWISVR BCI .0 JVM-com.ibm.lw THDW
ADMIN3 QLWISVR BCI .0 JVM-com.ibm.lw THDW
ADMIN4 QWEBADMIN BCI .4 JVM-/qibm/prod THDW
WEBINTER QTMHHTTP BCH .0 PGM-QZHBMAIN SIGW
WEBINTER QTMHHTTP BCI .0 PGM-QZSRLOG SIGW
WEBINTER QTMHHTTP BCI .0 PGM-QZSRLOG SIGW
WEBINTER QTMHHTTP BCI .0 PGM-QZSRHTTP SIGW
WEBINTER QTMHHTTP BCI .0 PGM-QZSRHTTP DEQW
WEBINTER QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBINTER QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBINTER QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBINTER QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBINTER QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBINTER QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBINTER QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBINTER QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBSMART QTMHHTTP BCH .0 PGM-QZHBMAIN SIGW
WEBSMART QTMHHTTP BCI .0 PGM-QZSRLOG SIGW
WEBSMART QTMHHTTP BCI .0 PGM-QZSRLOG SIGW
WEBSMART QTMHHTTP BCI .3 PGM-QZSRHTTP SIGW
WEBSMART QTMHHTTP BCI .0 PGM-QZSRHTTP DEQW
WEBSMART QTMHHTP1 BCI 5.3 PGM-QZSRCGI TIMA
WEBSMART QTMHHTP1 BCI 6.6 PGM-QZSRCGI TIMA
WEBSMART QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBSMART QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBSMART QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBSMART QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBSMART QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBSMART QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBSMART QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBSMART QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBSMART QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBSMART QTMHHTP1 BCI .0 PGM-QZSRCGI TIMW
WEBTEST QTMHHTTP BCH .0 PGM-QZHBMAIN SIGW
WEBTEST QTMHHTTP BCI .0 PGM-QZSRLOG SIGW
WEBTEST QTMHHTTP BCI .0 PGM-QZSRHTTP SIGW
WEBTEST QTMHHTTP BCI .0 PGM-QZSRHTTP DEQW
All non-RF interactive sessions run in the QINTER subsystem. About the only thing to watch out for here is ‘runaway jobs’. This is either a program that is in a loop, or a job that is no longer connected to a workstation, but is desperately trying to reconnect to it.
QINTER QSYS SBS .0 DEQW
HABATERM GORDONM INT .0 PGM-RFSIGNON DSPW
OTCTERM GORDONM INT .0 PGM-RFSIGNON DSPW
QPADEV000B STEVENL INT .0 PGM-RFSIGNON DSPW
QPADEV000C NANCYN INT .0 PGM-UNITY DSPW
QPADEV000L MARYT INT .0 PGM-UNITY DSPW
+QPADEV000R MARYT INT .0 PGM-ASGC800 DSPW
+QPADEV000S RUBYS INT .0 PGM-ASGC800 DSPW
QPADEV001N SHANNONM INT .0 PGM-UNITY DSPW
QPADEV001P RUBYS INT .0 PGM-UNITY DSPW
QPADEV001Q RUBYS INT .0 PGM-UNITY DSPW
QPADEV001Z MARYC INT .0 PGM-UNITY DSPW
QPADEV0013 GORDONM INT .0 PGM-RFSIGNON DSPW
This subsystem runs the ASW direct jobs. The two characters after the underscore indicate the company (note VA for the main production company, and PL for test). Basically, these jobs control the flow of sales orders through the system.
UP1480BS QSYS SBS .0 DEQW
AMTRDR_PLA ASW4OWN BCH .0 PGM-ASGC900 DEQW
AMTRDR_PLT ASW4OWN BCH .0 PGM-ASGC900 DEQW
AMTRDR_PLT ASW4OWN BCH .0 PGM-ASGC900 DEQW
AMTRDR_VAA ASW4OWN BCH .0 PGM-ASGC900 DEQW
AMTRDR_VAT ASW4OWN BCH .0 PGM-ASGC900 DEQW
AMTRDR_VAT ASW4OWN BCH .0 PGM-ASGC900 DEQW
DPRQ1_PL ASW4OWN BCH .0 PGM-ASGC900 DEQW
DPRQ1_VA ASW4OWN BCH .0 PGM-ASGC900 DEQW
DPRQ2_PL ASW4OWN BCH .0 PGM-ASGC900 DEQW
DPRQ2_VA ASW4OWN BCH .0 PGM-ASGC900 DEQW
DPRQ3_PL ASW4OWN BCH .0 PGM-ASGC900 DEQW
DPRQ3_VA ASW4OWN BCH .0 PGM-ASGC900 DEQW
ORDFLOW_PL ASW4OWN BCH .0 PGM-ASGC900 DEQW
ORDFLOW_VA ASW4OWN BCH .0 PGM-ASGC900 DEQW
Cancelling a Job
In WRKACTJOB, find the job that is ‘runaway’, or perhaps the RF gun that is locked. Key ‘4’ in front of it, in the option field, and press F4.
End Job (ENDJOB)
Type choices, press Enter.
Job name . . . . . . . . . . . . > SHEILAV3 Name
User . . . . . . . . . . . . . > SHEILAV Name
Number . . . . . . . . . . . . > 101245 000000-999999
How to end . . . . . . . . . . . *CNTRLD *CNTRLD, *IMMED
Delay time, if *CNTRLD . . . . . 30 Seconds
Delete spooled files . . . . . . *NO *NO, *YES
Maximum log entries . . . . . . *SAME Number, *SAME, *NOMAX
Additional interactive jobs . . *NONE *NONE, *GRPJOB, *ALL
In the ‘How to end’ field, change *CNTRLD (controlled) to *IMMED (immediate). If you make the mistake of pressing Enter instead of F4, the job will go right into a controlled end. You can still go back and repeat this procedure, remembering this time to press F4.
End Job Abnormal
In extreme cases, this might not work - the job will just continue running. In that case, an end job abnormal (ENDJOBABN) can be run. Note that it can only be run ten minutes after an endjob *immed has been run.
End Job Abnormal (ENDJOBABN)
Type choices, press Enter.
Job name . . . . . . . . . . . . __________ Name
User . . . . . . . . . . . . . __________ Name
Number . . . . . . . . . . . . ______ 000000-999999
Additional Parameters
Duplicate job option . . . . . . *SELECT *SELECT, *MSG
F3=Exit F4=Prompt F5=Refresh F12=Cancel F13=How to use this display
F24=More keys