Has anybody experienced getting a "(FOC198) FATAL ERROR IN DATABASE I/O..." error when using the IBATCH JCL to run multiple WebFocus reports? It appears to happen right on the MATCH FILE command. Funny, when I run the report in IBATCH by itself, it runs fine. But when I run multiple reports, then it breaks. I was running 3 reports and the MATCH FILE command is in the last report. Here is the simple IBATCH jcl step:

//IPROCS JCLLIB ORDER=CUST.HVP.HVP.WFOCUS.V716.WFM.DATA
//* Call IRUNTEST proc
//IBATCH EXEC PROC=IRUNBATT
//TSCOM300.EDAPARM DD *
-t
//TSCOM300.EDAENV DD
// DD *
TRMIN=//DD:INPUT
TRMOUT=//DD:OUTPUT
*EDACURDIR=/u/iadmin/ibi/srv71/wfm/edatemp
//TSCOM300.INPUT DD *
EX SETOFF
EX SETOFF2
EX SETOFF3
//*
//TSCOM300.OUTPUT DD SYSOUT=*
/*
//


And here is the MATCH FILE statement that breaks it:

MATCH FILE APU1
SUM
APUCT AS 'APUCT'
BY WEEK AS 'WEEK'
RUN
FILE INT2
SUM
INTCT AS 'INTCT'
BY WEEK AS 'WEEK'
AFTER MATCH HOLD AS INTER1 OLD-OR-NEW
RUN
FILE CCD3
SUM
CCDCT AS 'CCDCT'
BY WEEK AS 'WEEK'
AFTER MATCH HOLD AS INTER OLD-OR-NEW
END
-RUN

This is a legacy report from FOCUS 7.0.9 that I am trying to convert to WebFocus.

Please advise if anybody had experienced this and found a way around it.

Thanks in advance for replying.

"(FOC198) FATAL ERROR IN DATABASE I/O..."

That's FOCSORT bombing out. I am guess that the DD statement that allocs this file in the new proc is allowing smaller initial and extents in comparison to the original JCL.

Look for the original FOCSORT DD card on the 7.0.9 proc and use it to override on the new proc. (Used to be good at this for o/n failures etc but I have forgotten it all!)

I see the FOCSORT DD stmt in our Focus 7.0.9 production sysout for that report and it's set to "UNIT=SYSDA,SPACE=(CYL,(4,4))". But I don't see the FOCSORT in my IRUNJCL job or even in the IBATCH job. Although when viewing the WebFocus SYSOUT for my report that bombed with an I/O error, I see where the FOCSORT appears, there's an MVS error saying "Inaccessible storage".

I tried hardcoding "FOCSORT DD UNIT=SYSDA,SPACE=(CYL,(4,4))" in my IRUNJCL job but that did not change anything.

I also tried harcoding "TSCOM300.FOCSORT DD UNIT=SYSDA,SPACE=(CYL,(4,4))" in my IBATCH job but that also did not change anything.

My problem right now is I can't find where that FOCUS or TSCOM300 proc is so I can code it directly in it. But coding it in the IBATCH or IRUNJCL should be good enough, right?

4 initial cylinders and 15 extensions of 4 cylinders may not be enough disk space for FOCSORT.

This is what I've had in my JCL:

//FOCSORT DD UNIT=PUBLIC,SPACE=(CYL,(400,50),RLSE),DCB=BLKSIZE=0
//S001WK01 DD UNIT=PUBLIC,SPACE=(CYL,(400,50),RLSE),DCB=BLKSIZE=0
//S001WK02 DD UNIT=PUBLIC,SPACE=(CYL,(400,50),RLSE),DCB=BLKSIZE=0
//S001WK03 DD UNIT=PUBLIC,SPACE=(CYL,(400,50),RLSE),DCB=BLKSIZE=0
//S001WK04 DD UNIT=PUBLIC,SPACE=(CYL,(400,50),RLSE),DCB=BLKSIZE=0
//S001WK05 DD UNIT=PUBLIC,SPACE=(CYL,(400,50),RLSE),DCB=BLKSIZE=0
//S001WK06 DD UNIT=PUBLIC,SPACE=(CYL,(400,50),RLSE),DCB=BLKSIZE=0

Lots more disk space allocated here.

You will have to add the PROC name to the DDNames to correctly override the PROC DD statements.

//IRUNBATT.FOCSORT DD UNIT=PUBLIC,SPACE=(CYL,(400,50),RLSE),DCB=BLKSIZE=0

This message has been edited. Last edited by: Francis Mariani,

quote:

//IBATCH EXEC PROC=IRUNBATT

This is in your production PROCLIB PDS. I'm sure an on-site resource can help you change FOCSORT DD to:

FOCSORT DD UNIT=SYSDA,SPACE=(CYL,(40,4),RLSE)

First, thanks for all your replies and suggestions.

I've tried both suggestions all the way to allocating 1000 cylinders even on all the S001WKnn DD's, but the error messages are all the same.

Any other suggestions I can try?

Ronn,

In your PROC, do you have DDNAMES for INTER and INTER1. If not, add them (,CATLG,DELETE) with plenty of space, then RLSE the extra space.

Try that...

Sorry, I made a mistake in the JCL override - instead of proc name, the override should be step name, eg.

//TSCOM300.FOCSORT DD UNIT=PUBLIC,SPACE=(CYL,(400,50),RLSE),DCB=BLKSIZE=0

(I haven't worked with mainframe JCL in quite a while).

I added the ff in my IRUNBATT proc:

INTER DD UNIT=SYSDA,SPACE=(CYL,(1000,100),RLSE),DCB=BLKSIZE=0,
DISP=(,CATLG,DELETE)
INTER1 DD UNIT=SYSDA,SPACE=(CYL,(1000,100),RLSE),DCB=BLKSIZE=0,
DISP=(,CATLG,DELETE)

But still got the same error message.

To solve this one I think you need to see how big the focsort grows on the original jcl by allocing FOCSORT to a permenant dataset on your 7.0.6 run and seeing what size you get.

Do the same on wf and if its much bigger when it busts - its probably an out of control table request that's internally looping for some reason.

If this is the case you will have to play with the code and reorder the order of your fields etc until you come up with something that works.

The other thing you can try quickly is setting the new XFOCUS features on SET XFOCUS=ON in your code and hope that the new version of FOCSORT cures your problem. Worth a try as it s quick.