why have we lost all the volume



JES, JES2, JCL utilities, IDCAMS, Compile & Run JCLs, PROCs etc...

why have we lost all the volume

Postby samb01 » Tue Jul 09, 2013 8:23 pm

Heelo,

we have a job which add lines in a tape which is on DISP=MOD

 //S20S     EXEC PGM=Q0096                                       
 //SYSPRINT DD SYSOUT=*                                           
 //SYSUT1   DD DSN=XDPR.DCFRLUS.CACFD,DISP=(OLD,KEEP)             
 //SYSUT2   DD DSN=B.SOVSYSV
,DISP=(MOD,CATLG),UNIT=VTS2,         
 //         LABEL=RETPD=2000,VOL=(,,,100),                       
 //         DCB=(RECFM=VBA,LRECL=255,BLKSIZE=27998)               
 //IN       DD DSN=FRSD.HISTO.JOBDUP,DISP=SHR               
 //SYSUDUMP DD SYSOUT=W,HOLD=YES                                 



the tape : B.SOVSYSV
had the volume :

B.SOVSYSV                                    NOT RECTLGD 2 
VOL SER NOS= 240291,936181,245467,923538,229867,           
VOL SER NOS= 951478,213217,962483,201187,251861,           
VOL SER NOS= 219786,230750,234760,241058,939214,           
VOL SER NOS= 938109,966295,219468,968611,939282,           
VOL SER NOS= 246036,220636,932085,933460,966461,           
VOL SER NOS= 944209,935783,202485,219325,240991,           
VOL SER NOS= 953949,223453,208299,959474,969191,           
VOL SER NOS= 206165,267746,272637,959535,234631,           
VOL SER NOS= 273250,230873,973462,291705,254852,           
VOL SER NOS= 921167,224044,218236,290641,947143,           
VOL SER NOS= 965410,212990,913128,296399,983442,           
VOL SER NOS= 990971,284550,266308,908156,985729,           
VOL SER NOS= 308922,908218,310644,343646,221246,           
VOL SER NOS= 281867,352765,994506,353681,908058,           
VOL SER NOS= 223746,232769,982045,208481,910108,           
VOL SER NOS= 282050,947228,405430,983539,207371,           
VOL SER NOS= 303232,414486,367234,990609,427873,           
VOL SER NOS= 204630,263298,336499.                         



And the next day, all the volume was lost... And we can't anderstand why..

The tape B.SOVSYSV add a new volume :

IEF233A M 0DBB,PRIVAT,SL,EORSCAN2,S20S,B.SOVSYSV                   
IECTMS9 0DBB,233841,EORSCAN2,SYSUT2  ,18/12/2018,00001,B.SOVSYSV   
IEC705I TAPE ON 0DBB,233841,SL,COMP,EORSCAN2,S20S.B,B.SOVSYSV,MEDIA2



and so, we have lost all the data in those volumes :

VOL SER NOS= 240291,936181,245467,923538,229867,           
VOL SER NOS= 951478,213217,962483,201187,251861,           
VOL SER NOS= 219786,230750,234760,241058,939214,           
VOL SER NOS= 938109,966295,219468,968611,939282,           
VOL SER NOS= 246036,220636,932085,933460,966461,           
VOL SER NOS= 944209,935783,202485,219325,240991,           
VOL SER NOS= 953949,223453,208299,959474,969191,           
VOL SER NOS= 206165,267746,272637,959535,234631,           
VOL SER NOS= 273250,230873,973462,291705,254852,           
VOL SER NOS= 921167,224044,218236,290641,947143,           
VOL SER NOS= 965410,212990,913128,296399,983442,           
VOL SER NOS= 990971,284550,266308,908156,985729,           
VOL SER NOS= 308922,908218,310644,343646,221246,           
VOL SER NOS= 281867,352765,994506,353681,908058,           
VOL SER NOS= 223746,232769,982045,208481,910108,           
VOL SER NOS= 282050,947228,405430,983539,207371,           
VOL SER NOS= 303232,414486,367234,990609,427873,           
VOL SER NOS= 204630,263298,336499.                         


Nobody succes to find why it'has happened..

We suspect that all of those volume are scratch (because actually, other data use those tape) but we don't know why.....

Thank's for your help.
samb01
 
Posts: 236
Joined: Mon Nov 16, 2009 7:24 pm
Has thanked: 0 time
Been thanked: 0 time

Re: why have we lost all the volume

 

Re: why have we lost all the volume

Postby NicC » Tue Jul 09, 2013 8:36 pm

We do not know as we do not know what entries your tape library management system has for those tapes. But your tape librarian will know and will know if they have been scratched or not. Why don't you go and ask them. We cannot because we do not work where you work.
The problem I have is that people can explain things quickly but I can only comprehend slowly.
Regards
Nic
NicC
Global moderator
 
Posts: 2690
Joined: Sun Jul 04, 2010 12:13 am
Location: Pushing up the daisys (almost)
Has thanked: 4 times
Been thanked: 105 times

Re: why have we lost all the volume

Postby dick scherrer » Tue Jul 09, 2013 8:44 pm

Hello,

What is in the catalog for this dataset name?

Why does the process MOD onto an existing multi-volume dataset?

Why not create a GDG and each run create a +1 generation and when it is time to read these, specify the entire gdg as the input file . . . ?
Hope this helps,
d.sch.
User avatar
dick scherrer
Global moderator
 
Posts: 6304
Joined: Sat Jun 09, 2007 8:58 am
Has thanked: 3 times
Been thanked: 91 times

Re: why have we lost all the volume

Postby steve-myers » Tue Jul 09, 2013 10:32 pm

DISP=(MOD,...) for a miltivolume tape data set is just a bit chancy. I have to agree with Mr. Scherrer. Another problem is some tape management systems make DISP=MOD to any tape data set difficult. Something like Mr. Scherrer's proposal might work better, but I can see possible problems with it, too.

At one site where I worked we used an SMF data set manager that would write several data sets on a tape volume. From time to time it seemed we would mess up a volume and the tape manager would get some sort of x13 (I don't remember the x any more) ABEND opening a new data set on the volume. Us on call people had a standard response to this problem; we told the SMF data set manager to start a new tape. After a few years of this we persuaded the SMF tape manager program's vendor to make this an automatic response to x13 ABENDs. We never really uncovered the root cause of the problem. We thought, at the time, it was an issue with shared tape drive usage; another system would grab the drive and write something to the tape, but we never could prove the theory. Perhaps something like that is happening here.
steve-myers
Global moderator
 
Posts: 1886
Joined: Thu Jun 03, 2010 6:21 pm
Has thanked: 4 times
Been thanked: 197 times

Re: why have we lost all the volume

Postby BillyBoyo » Tue Jul 09, 2013 11:21 pm

NOT RECTLGD 2


Did you notice that?

I believe that there is an "option" to cause an abend for such messages.

At a minimum, new data from that run would be lost. Would it loose most of the stuff? I don't know. Was all the lost data "new" for that run?

I also think Dick's suggestion is a better way to do it. I've never liked DISP=MOD for sticking data on the end of other data, there are pitfalls (like depending what you read it with you may get an "early" end-of-file).

If the number of volumes (lots with small amounts of data) is seen as a problem, several generations can be copied at one time to create a consolidated dataset.
BillyBoyo
Global moderator
 
Posts: 3804
Joined: Tue Jan 25, 2011 12:02 am
Has thanked: 22 times
Been thanked: 264 times

Re: why have we lost all the volume

Postby steve-myers » Wed Jul 10, 2013 10:47 am

  1. I'm not sure I counted it correctly, but I think there are 105 volume serials in your first post.
  2. From z/OS 1.13 JCL Reference,
    If you do not code a volume count and volume serial number, the system can extend an existing cataloged data set that resides on a removable volume up to 20 volumes.
    I'm not sure how to interpret the statement.
    • Max 20 volumes? This does not seem to apply.
    • Add a max of 20 volumes? But your output - what you posted, anyway - just shows 1 new volume.
    You will have to review earlier runs, but is it adding a new volume for each run? If so, it seems to me to be pretty stupid - not you, z/OS.
I do not think your statement that there are other data sets in the volume group is correct.
steve-myers
Global moderator
 
Posts: 1886
Joined: Thu Jun 03, 2010 6:21 pm
Has thanked: 4 times
Been thanked: 197 times

Re: why have we lost all the volume

Postby samb01 » Wed Jul 10, 2013 1:50 pm

Hello,

i 'm going to try to answer questions :

1) We already ask for team system a dn many other people in the socity but nobody can't anderstand what has happened...

2) The dataset is in DISP=MOD beacause we want to keep the data five years

LABEL=RETPD=2000,VOL=(,,,100),


3) We can't use gdg because gdg is limited to 255 générations... And we need more because the batch work every days, and one genetaion par day give 255 days.....

The number of volume before the "black out" was : 88

B.SOVSYSV                                    NOT
VOL SER NOS= 240291,936181,245467,923538,229867,
VOL SER NOS= 951478,213217,962483,201187,251861,
VOL SER NOS= 219786,230750,234760,241058,939214,
VOL SER NOS= 938109,966295,219468,968611,939282,
VOL SER NOS= 246036,220636,932085,933460,966461,
VOL SER NOS= 944209,935783,202485,219325,240991,
VOL SER NOS= 953949,223453,208299,959474,969191,
VOL SER NOS= 206165,267746,272637,959535,234631,
VOL SER NOS= 273250,230873,973462,291705,254852,
VOL SER NOS= 921167,224044,218236,290641,947143,
VOL SER NOS= 965410,212990,913128,296399,983442,
VOL SER NOS= 990971,284550,266308,908156,985729,
VOL SER NOS= 308922,908218,310644,343646,221246,
VOL SER NOS= 281867,352765,994506,353681,908058,
VOL SER NOS= 223746,232769,982045,208481,910108,
VOL SER NOS= 282050,947228,405430,983539,207371,
VOL SER NOS= 303232,414486,367234,990609,427873,
VOL SER NOS= 204630,263298,336499.             


May be we reach the limit.... I don't know...

The only thing i know, is the day after, the dataset : B.SOVSYSV start with a new volume and the 87 volumes became scratch....

5) We use CA1

6) There is not a new volume each days
samb01
 
Posts: 236
Joined: Mon Nov 16, 2009 7:24 pm
Has thanked: 0 time
Been thanked: 0 time

Re: why have we lost all the volume

Postby NicC » Wed Jul 10, 2013 2:16 pm

Create a new GDG every day. every week copy them into a weekly GDG. Every month/quarter/year roll the weekly GDGs into a monthly/qur=arterly/yearly GDG. Max (assuming 5 working days ): 5 daily GDGs, 5 weekly GDGs, 3 monthly GDGs, 4 quarterly GDGs and 5 yearly GDGs. And if all those taqpes get mouinted every day - vast reduction in tape mounts and thus faster through-put time.
The problem I have is that people can explain things quickly but I can only comprehend slowly.
Regards
Nic
NicC
Global moderator
 
Posts: 2690
Joined: Sun Jul 04, 2010 12:13 am
Location: Pushing up the daisys (almost)
Has thanked: 4 times
Been thanked: 105 times

Re: why have we lost all the volume

Postby halfteck » Mon Jul 15, 2013 12:22 pm

Something else to consider regarding daily / weekly / monthly copying / consolidating multiple GDG's. if you attempt concatenation by leaving off the relative generation number, ie : DSN=MY.GDG all generations will be read sequentially STARTING WITH THE LATEST GENERATION (0), therefore the data will NOT be stored in the order you originally wanted, so some form of sort (huge if you could really fill up multiple tape volumes) will be required to place data in the corect order
halfteck
 
Posts: 42
Joined: Tue Nov 08, 2011 8:47 pm
Has thanked: 0 time
Been thanked: 0 time


Return to JCL

 


  • Related topics
    Replies
    Views
    Last post