| View previous topic :: View next topic |
| Author |
Message |
warp5
Intermediate
Joined: 02 Dec 2002
Posts: 429
Topics: 18
Location: Germany
|
 Posted: Fri Jul 07, 2006 2:05 am Post subject: |
 |
|
| twb's or Mervyn's situation leave uncataloged datasets lying around. They should check their disks how many of those uncataloged datasets there are - they might get a lot of disk space freed by deleting those "ghost" datasets. |
|
| Back to top |
|
 |
Mervyn
Moderator
Joined: 02 Dec 2002
Posts: 415
Topics: 6
Location: Hove, England
|
| Posted: Fri Jul 07, 2006 6:06 am Post subject: |
|
|
| warp5 wrote: | | twb's or Mervyn's situation leave uncataloged datasets lying around. They should check their disks how many of those uncataloged datasets there are - they might get a lot of disk space freed by deleting those "ghost" datasets. |
That's absolutely right. I run a job most days looking for datasets unused for four weeks, and I pick up a few of these in a week.
Our PARMLIB has three ALLOCxx members, and they all contain | Code: | CATLG_ERR FAILJOB(YES)
ERRORMSG(YES) |
_________________
The day you stop learning the dinosaur becomes extinct |
|
| Back to top |
|
|
Jaya
Beginner
Joined: 02 Sep 2005
Posts: 77
Topics: 10
Location: Cincinnati
|
| Posted: Fri Jul 07, 2006 6:59 am Post subject: |
|
|
TWD,
I have seen CA11 doing it in our production system.
A utility CA11RMS is used in batch jobs to delete all already catalogued files that are to be created as output files i.e files with Disp=new
If the job contains a CA11RMS step in the beginning, one does not get a JCL error
_________________
"Great spirits have always encountered violent opposition from mediocre minds."
-Albert Einstein |
|
| Back to top |
|
|
twb
Beginner
Joined: 15 Jun 2006
Posts: 5
Topics: 2
Location: San Francisco
|
| Posted: Tue Jul 11, 2006 10:47 am Post subject: DISP=(NEW for Existing |
|
|
"Dedpending on what is coded in the ALLOCxx member of PARMLIB depends on what course of action will taken"
| Code: |
/********************************************************************/
/* */
/* CATLG_ERR defines policies for when */
/* datasets fail to get (re)cataloged at */
/* step unallocation. */
/* */
/********************************************************************/
CATLG_ERR ERRORMSG(YES) /* Issue an error message to the */
/* console if the dataset cannot be */
/* cataloged. */
FAILJOB(YES) /* Fail the job if the dataset */
/* cannot be cataloged. */
|
While I agree this information is correct, I doubt it is "on point." Our sysprog pointed me at the IBM doc (very nice RTFM note) and as I read it, my issue is that CATLG_ERR is not triggered. Instead of raising an error for DISP=(NEW with an existing dataset, something is quietly removing the old dataset.
I might mention, I have observed this on other systems but thought it was a SAS trick in the LIBNAME statement. I think DISP=NEW has destroyed existing datasets for several years.
I DO appreciate your thoughts--you give me credibility with the locals.
Thanks in advance for more ideas,
Tim Berryhill |
|
| Back to top |
|
|
taltyman
JCL Forum Moderator
Joined: 02 Dec 2002
Posts: 310
Topics: 8
Location: Texas
|
| Posted: Tue Jul 11, 2006 3:04 pm Post subject: |
|
|
Still sounds flaky to me. Can you recreate this with a test dataset name?
Create a test dataset, post a catalog listing for it. Run your job with disp=(NEW,CATLG,DELETE). Post pertinent job log messages. Post a catalog listing for the newly created dataset. |
|
| Back to top |
|
|
nevilh
Beginner
Joined: 11 Aug 2004
Posts: 115
Topics: 0
|
| Posted: Wed Jul 12, 2006 2:14 am Post subject: |
|
|
| If you specify DISP=(NEW,CATLG) and an existing dataset is overwritten then this is an integrity exposure and should be reported IBM |
|
| Back to top |
|
|
warp5
Intermediate
Joined: 02 Dec 2002
Posts: 429
Topics: 18
Location: Germany
|
| Posted: Thu Jul 13, 2006 1:31 am Post subject: |
|
|
| twb, you need to let us know more about your environment (Jes2? Scheduler? etc. ) and how you are running your job, is CA11 involved (see comments about that above) or some other tool? If you have DISP=(NEW,CATLG) then your job should abort if the dataset already exists, as a matter of fact you should get a JCL failure (in JES2). |
|
| Back to top |
|
|
nevilh
Beginner
Joined: 11 Aug 2004
Posts: 115
Topics: 0
|
| Posted: Thu Jul 13, 2006 2:21 am Post subject: |
|
|
twb stated in initial post that the old dataset was uncataloged. In this case it is quite valid to try and allocate a dataset with the same name on a different device.
We do this regularly . I agree with taltyman this sounds a bit flaky, I find it hard to believe such a basic function would have such an obvious error. |
|
| Back to top |
|
|
|
|
|
FAQ
Search
Quick Manuals
Register
Profile
Log in to check your private messages
Log in
