PI51870: ADS STATUS IN RECON IS AVAIL EVEN IF SEVERE ERROR AT DEDB ALTER

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• ADS status of RECON and /DIS AREA command didn't match when
  shadow ADS I/O error has occured during DEDB Alter.
  At completion of DEDB ALTER, /DIS AREA shows ADS is SEVERE ERROR
  but LIST.RECON shows as AVAIL.
  < Scenario >
    Data share between IMS1 and IMS2
    Two active ADSs and two shadow ADSs + two shadow IC data sets.
    Current OLDS is OLP00.
    Reserve OLP01 from other CEC.
    Run DEDB Alter.
      DEDB alter almost completed
      OLDS switch requested from ALTER_ADIS
       ==> Start pending on OLP01
      Force shadow ADS2 offline.
      release OLP01
       ==> Write error on 2nd CI of shadow ADS2.
    DEDB alter completed with RC=00 even if I/O error on shadow
    ADS.
    /DIS AREA in IMS1
      => ADS2 was not closed and is SEVER ERROR
    /DIS AREA in IMS2
      => ADS2 was remain open without SEVERE ERROR
    LIST.RECON
      => Old shadow ADSs became active and status is AVAIL
  -------------------------------------------------------------
  Same problem happens even if I/O error on ADS1, or both ADS1 and
  ADS2.
  < from SYSLOG in IMS1 >
   DFS3705I AREA=DEPO240  DD=DEPO24I1 CLOSED IMP1
   DFS2500I DATASET  DEPO24I1 SUCCESSFULLY DEALLOCATED IMP1
   DFS3705I AREA=DEPO240  DD=DEPO24I2 CLOSED IMP1
   DFS2500I DATASET  DEPO24I2 SUCCESSFULLY DEALLOCATED IMP1
   DFS2500I DATASET  DEPO2401 SUCCESSFULLY DEALLOCATED IMP1
   DFS2500I DATASET  DEPO2402 SUCCESSFULLY DEALLOCATED IMP1
   DFS3713A (00201110) AREA=DEPO240  DD=DEPO24S2,WRITE ERROR,
   CI-RBA=00004000,SEVERE ERROR IMP1
   DFS2500I DATASET  IMSACB   SUCCESSFULLY DEALLOCATED IMP1
   DFS552I    IFP  REGION ALTADEPO STOPPED ID=00001 TIME=1704
   ...
   @DIS AREA DEPO240.
   +    AREANAME  EQECT   TOTAL UNUSED  TOTAL UNUSED      DBNAME E
      CONDITIONS
         DDNAME   REMAIN SEQ DEPENDENT DIR ADDRESSABLE
        DEPO240     N/A     N/A     N/A      35     N/A  JBX0DEPO
         DEPO24S1    10     N/A     N/A     N/A     N/A      N/A
   +     DEPO24S2   N/A     N/A     N/A     N/A     N/A      N/A
      SEVERE-ERROR
        *15299/170657*
  < from SYSLOG in IMS2 >
   DFS3705I AREA=DEPO240  DD=DEPO24I1 CLOSED IMP5
   DFS2500I DATASET  DEPO24I1 SUCCESSFULLY DEALLOCATED IMP5
   DFS3705I AREA=DEPO240  DD=DEPO24I2 CLOSED IMP5
   DFS2500I DATASET  DEPO24I2 SUCCESSFULLY DEALLOCATED IMP5
   DFS2500I DATASET  DEPO2401 SUCCESSFULLY DEALLOCATED IMP5
   DFS2500I DATASET  DEPO2402 SUCCESSFULLY DEALLOCATED IMP5
   DFS2500I DATASET  IMSACB   SUCCESSFULLY DEALLOCATED IMP5
   ....
   @DIS AREA DEPO240.
   +    AREANAME  EQECT   TOTAL UNUSED  TOTAL UNUSED      DBNAME E
      CONDITIONS
         DDNAME   REMAIN SEQ DEPENDENT DIR ADDRESSABLE
        DEPO240     N/A     N/A     N/A      35     N/A  JBX0DEPO
         DEPO24S1    10     N/A     N/A     N/A     N/A      N/A
         DEPO24S2    10     N/A     N/A     N/A     N/A      N/A
        *15299/170700*
  ----------------------------------------------------------------
  I found the reason why DBFMEQE0 skipped ADS close at 2nd CI
  write error.
  ALTER_ADIS routine in DBFDAU70 set both EPSTOREQ=EPSTALTR and
  EPSTCREQ=EPSTAREC at line#3270, then called DBFMLCL0.
  DBFMLCL0 did not clear EPSTCREQ if EPSTOREC=EPSTALTR and
  EPSTCREQ=EPSTAREC.
  With this EPSTCREQ, ALTER_ADIS called DBFMOCI0 at line#3344.
  DBFMOCI0 set DMHRI2OC if EPSTCREQ=EPSTAREC around line#329.
  At I/O error on 2nd CI, DBFMEQE0 was called from DBFMER00.
  DBFMEQE0 checked DMHRI2OC and set ADSCF2IO at line#359, and
  called MEQERTN without setting DMHRI4SP. Then DBFMEQE0 skips
  calling of MEQECADS (Close ADS routine).
  And failed ADS was not closed in both active and data share IMS.
  I feel that we need to set EPSTCREQ=0 before calling DBFMOCI0.
  Or, DBFMLCL0 clears EPSTCREQ just after lable SKIP_AREA_CLOSE.
  Then DBFMEQE0 will handle ADS unavailable and internal stop ADS.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * All IMS V14 Fast Path DEDB Alter utility users               *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * ADS status of RECON and /DIS AREA command didn't match when  *
  * shadow ADS I/O error has occured during DEDB Alter.          *
  *                                                              *
  * At completion of DEDB ALTER, /DIS AREA shows ADS is SEVERE   *
  * ERROR,  but LIST.RECON shows as AVAIL.                       *
  *                                                              *
  * DFS4651E was not issued even if ALTERGRP is not specified    *
  ****************************************************************
  * RECOMMENDATION:                                              *
  * INSTALL CORRECTIVE SERVICE FOR APAR/PTF                      *
  ****************************************************************
  ADS status of RECON and /DIS AREA command didn't match when
  shadow ADS I/O error has occured during DEDB Alter.
  
  At completion of DEDB ALTER, /DIS AREA shows ADS is SEVERE
  ERROR,  but LIST.RECON shows as AVAIL.
  
  DFS4651E was not issued even if ALTERGRP is not specified
  

Problem conclusion

• DBFDAU70
  
  Clear EPSTCREQ before calling DBFMOCI0
  Compare AREA name with 8 byte length
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

  UI33826

Modules/Macros

• DBFDAU70
  

Fix information

• Fixed component name

  IMS V14

• Fixed component ID

  5635A0500

Applicable component levels

• R400 PSY UI33826

     UP15/12/19 P F512

Fix is available

• Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.