Abend S813-04 during DB2 Reorg and using Vtape

Products

• Vtape Virtual Tape System

Issue/Introduction

Over the last few months the customer experienced 2 abends (S813-04) in a DB2 Reorg job $DB2xyz using Tape Work files on CA-VTAPE. The DB2 Reorg job is run every Saturday morning in parallel with another DB2 Reorg Job $PRIxyz.  At the time of the abend there is a VTAPE Cache Full condition (message SVT1V0432E 0106,G73123,$PRIxyz ,Cache Full, Mount pending and will resume when Cache become available).  The CA-VTAPE Cache Full condition has occurred several times this year, and each time during the execution of the above DB2 Jobs (on 2/26, 4/23, 5/7, 5/21, and 6/4). The abend of job $DB2xyz happened on 2 occasions this year (5/21 and 6/4). 

Sample job output:  

07:16:50.81 JOB35150 00000094  IECTMS9 010D,G73129,** WORK TAPE  **
07:16:50.82 JOB35150 00000090  IEC205I SYS00010,$DB2xyz,REORG1,FILESEQ=1, COMPLETE VOLUME LIST, 684
                               684 00000090  DSN=HSAL.DB.aaaaa.bbbbbb.cccccc.P00000.D155,VOLS=G73068,G73124,G73129, 
                               684 00000090  TOTALBLOCKS=1301909  
07:16:50.89 JOB35150 00000090  TMS014  IEC502E D 010D,G73129,SL,$DB2xyz,REORG1
07:16:50.89 JOB35150 00080094  IEC502E D 010D,G73129,SL,$DB2xyz,REORG1 
07:16:50.89 JOB35150 00000090 *TMS016  IEC501A M 010D,G73068(ACTL=QI6559),SL,COMP,$DB2xyz,REORG1,
                                                        HSAL.DB.aaaaa.bbbbbb.cccccc.P00000.D155 
07.16.50 JOB35150 *TMS016  IEC501A M 010D,G73068(ACTL=QI6559),SL,COMP,$DB2xyz,REORG1,HSAL.DB.aaaaa.bbbbbb.cccccc.P00000.D155
07.16.51 JOB35150  IEC149I 813-04,IFG0195H,$DB2xyz,REORG1,SYS00010,010D,G73068,
                          697               HSAL.DB.aaaaa.bbbbbb.cccccc.P00000.D155,                    
                          697               LBL=00000000000000000 

Environment

Release : 12.6

Component : VTAPE

Cause

Vtape is scratching some temporary volumes prematurely. 

Resolution

The Vtape scratch/sync processing was occurring at the same time that the DB2 jobs were running, with CA1 using WRKFLS set to 'YES' (the default) allowing Vtape to scratch some temporary files/volumes prematurely.  The resolution is to either schedule the Vtape scratch/sync housekeeping jobs to occur at times when the DB2 jobs are not running, or else set CA1 WRKFLS to 'NO'.