View previous topic :: View next topic
|
Author |
Message |
vasanthz
Global Moderator
Joined: 28 Aug 2007 Posts: 1742 Location: Tirupur, India
|
|
|
|
Hi,
We have a SAS sort step and SAS calls host sort(DFSORT) to perform sorting, during this step the job is failing with the below abend, the unusal thing is, this job runs fine once restarted without any changes made to the job.
Code: |
SASLOG:
.ERROR: Sort did not complete successfully. See messages on the Job Console Log
.ERROR: Host sort cannot be used.
.NOTE: The SAS System stopped processing this step because of errors.
.NOTE: SAS set option OBS=0 and will continue to check statements. This may caus
.NOTE: There were 1978265 observations read from the data set WORK02.SMFXCQ00.
.WARNING: The data set WORK01.SMFXCQ00 may be incomplete. When this step was st
. variables.
.ERROR: Expecting page 1650, got page -1 instead.
.ERROR: Page validation error while reading WORK01.SMFXCQ00.DATA.
.ERROR: Expecting page 1, got page -1 instead.
.ERROR: Page validation error while reading WORK01.SMFXCQ00.DATA.
.ERROR: Expecting page 1, got page -1 instead.
.ERROR: Page validation error while reading WORK01.SMFXCQ00.DATA.
.ERROR: Expecting page 1, got page -1 instead.
.ERROR: Page validation error while reading WORK01.SMFXCQ00.DATA. |
Code: |
SYSOUT:
.ICE185A 0 AN S878 ABEND WAS ISSUED BY DFSORT, ANOTHER PROGRAM OR AN EXIT (PHASE S 3) |
Could you please let me know the reason for this failure.
There was link which has a description of similar abend www-304.ibm.com/support/docview.wss?uid=swg1PK66166
but in our job it runs fine once restarted after abend, without any changes. |
|
Back to top |
|
|
expat
Global Moderator
Joined: 14 Mar 2007 Posts: 8797 Location: Welsh Wales
|
|
|
|
I have only experienced one problem with DFSORT under SAS, which with some fast and good help from Frank and Dave was quickly resolved.
Are you using predefined SORTWK DD statements.
If you are, then I suggest that you do not, but use
Code: |
//DFSPARM DD *
OPTION DYNALLOC=(,32) |
The thing that took me ages to find was that SAS always invokes its own SORT routine before deciding that DFSORT should really be doing this, but then limits the number of SORTWK's that it will use. Limit of 6 or whatever your site options are set to.
To overcome this I changed the SORT OPTION - see below - to bypass the SAS intervention and go directly to DFSORT instead.
Code: |
OPTIONS SORTPGM=SORT |
I don't know if this is related to your problem but from the benefit of experience I'm happy to share with you. |
|
Back to top |
|
|
Frank Yaeger
DFSORT Developer
Joined: 15 Feb 2005 Posts: 7129 Location: San Jose, CA
|
|
Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007 Posts: 1742 Location: Tirupur, India
|
|
|
|
Quote: |
I don't know if this is related to your problem but from the benefit of experience I'm happy to share with you. |
thanks
Quote: |
It may have been a temporary system storage shortage of some kind that disappeared by the time you reran the job. |
This abend happens like 3 times a week. So I guess its not a one off issue.
The JCL is generated by several SAS macros(30 or more), its a product by CA which generates the JCL and I think macros are overused to the point where we dont get to see actual SAS data and proc statements or JCL statements.
I would check if the SORTWK DD is hardcoded in the generated JCL and remove them and add the DFSPARM and check out if it helps.
Thanks Expat and Frank for your thoughts. |
|
Back to top |
|
|
PeterHolland
Global Moderator
Joined: 27 Oct 2009 Posts: 2481 Location: Netherlands, Amstelveen
|
|
Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007 Posts: 1742 Location: Tirupur, India
|
|
|
|
Thanks Peter that looks like a relevant link to the problem.
The link suggests,
Quote: |
1. Submit your data set in interactive. Does it run without error?
2. Run PROC SORT, which sorts the entire data set. |
the dataset being worked is temp work dataset, i would try to create a perm dataset and check the solution.
Thanks, |
|
Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007 Posts: 1742 Location: Tirupur, India
|
|
|
|
Hi,
There was a message on the JES log
Code: |
+Requested MEMSIZE=75497472 exceeded available REGION=60792832 |
and also Frank mentioned that this was a storage issue, so changed REGION=0M and there had been no failures so far for the past 4 days(touchwood).
Quote: |
the unusal thing is, this job runs fine once restarted without any changes made to the job. |
The JCL has several steps, & the JCL had REGION=48M earlier on job card. When the job was running and came to the SORT step it failed, maybe the region was insufficient.
When we restarted from "failed step" it ran successfully, maybe the region was sufficient now since it was restarted from failed step.
Thanks Expat, Peter & Frank. |
|
Back to top |
|
|
|