Ed Finnell wrote:
In a message dated 6/8/2010 8:00:40 A.M. Central Daylight Time,
mzel...@flash.net writes:
Reading the description, it looks like a problem that would only affect
a single data set. How did it affect so many data sets and volumes
in your environment?
Yes. The problem happens one data set at a time (or maybe it can affect
all OPEN-for-output PDSEs when a job is canceled).
As I understand it, PDSE modules get called by CATALOG/DADSM if DS1PDSE
is ON in the FMT-1 or FMT-8 DSCB and if there is an abend in PDSE code
at an inopportune time, it can be disruptive to those operations.
As near as I can tell, in one case someone renamed one of the corrupted
PDSEs and that somehow led to overlapping extents with 22 other data
sets. (These are big PDSEs used to hold ADATA from product builds.) That
overlap led to a lot of heartache and a reINITed volume.
After getting current with recommended maintenance and IPLing, we had
issues with CATALOG errors trying to delete another one of the corrupted
PDSEs via IDCAMS:
DELETE EJES.PRODGEN.ADATA PURGE
0IDC3014I CATALOG ERROR
IDC3009I ** VSAM CATALOG RETURN CODE IS 102 - REASON CODE IS IGG0CLFM-2
IDC0551I ** ENTRY EJES.PRODGEN.ADATA NOT DELETED
0IDC0001I FUNCTION COMPLETED, HIGHEST CONDITION CODE WAS 8
And DADSM errors trying to delete via ISPF 3.2 or TSO/E DELETE command
with ISPF error messages and accompanying SVC dumps:
| An internal service failed with return code = 12 decimal, reason code
| X'280F01D5'. ISPF may be unable to obtain information about an HFS file or
| PDSE. More information may be available in the ISPF log.
IEC331I 042-006(043D57D3),EDJXADM ,$IKJTEST,SCRT,IGG0CLH0
IEC331I VOL,MVSEV0,NAME,EJES.PRODGEN.ADATA
IEC614I SCRATCH FAILED - RC 008, DIAGNOSTIC INFORMATION IS (043D57D3),
$IKJTEST,MVSEV0,EJES.PRODGEN.ADATA
***
IEA045I AN SVC DUMP HAS STARTED AT TIME=10.39.19 DATE=06/05/2010 085
FOR ASIDS(0009,0071)
ERROR ID = SEQ00044 CPU00 ASID0071 TIME10.39.19.2
QUIESCE = YES
IEA794I SVC DUMP HAS CAPTURED: 086
DUMPID=001 REQUESTED BY JOB (EDJXADM )
DUMP TITLE=COMPID=DF115,CSECT=IGWDACND+1AFA,DATE=04/06/09,MAINT
ID= NONE ,ABND=0F4,RC=00000024,RSN=01045AF1
IEF196I IGD17070I DATA SET SYS3.DUMP.D100605.T103919.EDJXADM.S00001
IEF196I ALLOCATED SUCCESSFULLY WITH 1 STRIPE(S).
IEF196I IGD17160I DATA SET SYS3.DUMP.D100605.T103919.EDJXADM.S00001
IEF196I IS ELIGIBLE FOR COMPRESSION
IEF196I IGD101I SMS ALLOCATED TO DDNAME (SYS00002)
IEF196I DSN (SYS3.DUMP.D100605.T103919.EDJXADM.S00001 )
IEF196I STORCLAS (SVCDUMP) MGMTCLAS (DUMPS) DATACLAS (SVCDUMP)
IEF196I VOL SER NOS= MVSEV0
IEC331I 042-006(043D57D3),EDJXADM ,$IKJTEST,SCRT,IGG0CLH0
IEC331I VOL,MVSEV0,NAME,EJES.PRODGEN.ADATA
IGD17040I ERROR IN DADSM PROCESSING ON VOLUME MVSEV0 FOR DATA SET 097
EJES.PRODGEN.ADATA
HISTORIC RETURN CODE IS 8 DIAGNOSTIC INFORMATION IS 043D57D3
IGD306I UNEXPECTED ERROR DURING IGGDAS02 PROCESSING 098
RETURN CODE 4 REASON CODE 211
THE MODULE THAT DETECTED THE ERROR IS IGDVTSDA
SMS MODULE TRACE BACK - VTSDA VTSCU VTSCT VTSDL SSIRT
SYMPTOM RECORD CREATED, PROBLEM ID IS IGD00000
IEF196I IGD104I SYS3.DUMP.D100605.T103919.EDJXADM.S00001 RETAINED,
IEF196I DDNAME=SYS00002
IEA611I COMPLETE DUMP ON SYS3.DUMP.D100605.T103919.EDJXADM.S00001 102
DUMPID=001 REQUESTED BY JOB (EDJXADM )
FOR ASIDS(0009,0071)
INCIDENT TOKEN: PHXHQ MVS60 06/05/2010 17:39:19
ERROR ID = SEQ00044 CPU00 ASID0071 TIME10.39.19.2
IEF196I IEF237I 8112 ALLOCATED TO IPCSDDIR
The "fix" is to follow the somewhat convoluted process I described
earlier zapping VTOC, deleting the PDSE, allocating a dummy sequential
file, etc. for each PDSE so affected. Then recovering from backup or
rebuilding.
I suppose the pervasiveness of this issue depends entirely on how many
PDSEs you have in your shop and how many of them get corrupted before
you notice there is an issue.
FWIW, IBM says they are looking into adding DATALOSS/HIPER flags. I'll
keep you posted.
--
Edward E Jaffe
Phoenix Software International, Inc
831 Parkview Drive North
El Segundo, CA 90245
310-338-0400 x318
edja...@phoenixsoftware.com
http://www.phoenixsoftware.com/
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
