RE: IPL problems -- MIDRANGE-L


        -----Original Message-----
        From:   S. Kalyana Sundaram [SMTP:stgsys@squared.co.in]
        Sent:   Monday, November 24, 1997 4:39 AM
        To:     'MIDRANGE-L@midrange.com'
        Subject:        IPL problems

        We recently had a disk unit failure on our AS/400 (V2 R3).
Before the 
        crash, it took hardly 20 minutes for IPL. After disaster
recovery,  we 
        accidentally applied all PTFs permanently. (All these PTFs had
been applied 
        even before but only temporarily). Now the system takes nearly 2
hours for 
        IPL.. The IPL proceeds normally until SRC code C900 2AA0. At
this code, it 
        takes nearly 1 and a half hours (normal time should be less than
a minute). 

        In the manual, this SRC code C900 2AA0 indicates damage object 
        notification.

        We are not able to understand the reason for this abnormal delay
in IPL. 
        Can anybody please help us  in figuring out the problem ?

        TIA,

        GopiPriya,


        Just some guesses here.  Maybe your machine pool is set too
small ?

        There are several APAR's relating to possible problems with this
IPL step that have had PTF's issued.  Of course the standard advice here
is to get off of V2R3 and install V3R2 with the latest PTF cum package.
You don't mention what PTF level you are at for V2R3.


        See below APARs that may address the issue:



        Item SA48972

          APAR Identifier ...... SA48972       Last Changed..97/02/24  
          OSP-F90 IPL RUNNING FOR OVER 90 HOURS.
         
         
          Symptom ...... DD LOOP              Status ........... CLOSED
DOC
          Severity ................... 2      Date Closed .........
97/02/24
          Component .......... 5763SS100      Duplicate of ........
          Reported Release ......... 310      Fixed Release ............
          Component Name 5763 OS/400 ROC      Special Notice
          Current Target Date ..              Flags
          SCP ...................
          Platform ............
         
          Status Detail: APARCLOSURE - APAR is being closed.
         
          PE PTF List:
         
          PTF List:
         
         
          Parent APAR:
          Child APAR list:
         
         
          ERROR DESCRIPTION:
          * EQUIVALENT ABSTRACT:
          OSP-F90 IPL RUNNING FOR OVER 90 HOURS_
          *
          See problem summary.
         
         
          LOCAL FIX:
          force manual ipl ,change major system values at ipl and
          increase qmchpool value.
         
         
          PROBLEM SUMMARY:
          F90 with 512MB main store installed from D70 192MB savsys and
          nonsys and ipl'd which failed to complete aftre 90 hours,
          srcc9002AA0 displayed for over 24 hours,cpu start/stop shows
          different tdes running, DST via F21 on panel fails with 0000
          0000 in panel lights.  also suspect qdbsrvxr attemping to
update
          the cross reference files after the nonsys restore.  MSSD
shows
          the machine pool too small and thrashing as the cause of the
          long IPL.  The system values qmchpool and qbaspool restored
from
          D70 savsys were 35,000KB and 50,000KB which caused the machine
          pool to be set too small on the F90.Even qpfradj was page
          faulting thus unable to tune the system by adjusting pool
sizes.
          Backup Recovery Guide sc41-3304-01 page 11-12 needs to include
          system values qmchpool and qbaspool to be changed if restoring
          onto a different CPU with qmchpool set to 15% of main storage
          size.  This will prevent the incorrect resources in the
machine
          pool, while base pool will still get all unallocated storage.
         
         
          PROBLEM CONCLUSION:
          The Basic B&R book will be changed for the next release to
tell
          the customer that in the given circumstance the machine pool
          should be set to 15% of main storage.
         



        Item SA36355

          APAR Identifier ...... SA36355       Last Changed..95/03/15  
          OSP-SRCC9002AA0-LOOP IPL WITH DAMAGED OR DELETED RECOVERY OBJ.
         
         
          Symptom ...... LP SRCC9002AA0       Status ........... CLOSED
PER
          Severity ................... 2      Date Closed .........
94/08/30
          Component .......... 5738SS1DB      Duplicate of ........
          Reported Release ......... 230      Fixed Release ............
999
          Component Name 5738 OS/400 DAT      Special Notice
          Current Target Date ..94/09/30      Flags
          SCP ...................
          Platform ............
         
          Status Detail: SHIPMENT - Packaged solution is available for
                                    shipment.
         
          PE PTF List:
         
          PTF List:
          Release 230   : SF18234 available 94/09/27 (5038 )
         
         
          Parent APAR:
          Child APAR list:
         
         
          ERROR DESCRIPTION:
          ******* (Do NOT alter/erase this or next 3 lines) *******
          * EQUIVALENT ABSTRACT:
          OSP-SRCC9002AA0-LOOP IPL WITH DAMAGED/DELETED RECOVERY OBJ_
          *
          THIS APAR WAS SYSROUTED TO R305 SA36786.
          Customer initiated a normal ipl after completing an entire
          system save. The IPL looped at step SRCC9002AA0. Restarting
the
          IPL nor installing the system resolved the problem. Processor
          light was solid during this step.
         
         
          LOCAL FIX:
         
         
          PROBLEM SUMMARY:
          Customer experienced an unexpected power failure but did
          successfully IPL and ran a couple of days without problems
after
          the IPL. The customer was running under commitment control. A
          job that used a file that was active during the power failure
          was started and ran two days before being cancelled since it
was
          in a loop. After completing an entire system save the customer
          initiated a normal IPL. The IPL looped at step SRCC9002AA0.
          Neither restarting the IPL nor re-installing the system
          resolved the problem. Processor light was solid during this
          step.
         
         
          PROBLEM CONCLUSION:
          The system encountered a damaged/deleted recovery object
related
          to commitment control during IPL. The delete file code was not
          properly handling this situation and an infinite loop
occurred.
          The code has been changed to prevent the infinite loop.
         

        Item SA37155

          APAR Identifier ...... SA37155       Last Changed..94/10/12  
          IPL- SRCC9002AA0 LOOPING FOR HOURS IN SCPF
         
         
          Symptom ...... AB SRCC9002AA0       Status ........... CLOSED
CAN
          Severity ................... 2      Date Closed .........
94/10/12
          Component .......... 5763SS1DB      Duplicate of ........
          Reported Release ......... 305      Fixed Release ............
          Component Name 5763 OS/400 DAT      Special Notice
          Current Target Date ..              Flags
          SCP ...................
          Platform ............
         
          Status Detail: APARCLOSURE - APAR is being closed.
         
          PE PTF List:
         
          PTF List:
         
         
          Parent APAR:
          Child APAR list:
         
         
          ERROR DESCRIPTION:
          9406 300 scratched installed with os400 and user data and
          running ok, the wsc controller resource names where incorrect
          so srm data cleared and system re ipl, and looped at
          srcc9002AA0 for 4 hours in SCPF.
          reload of vlic and os400 gave same symptom.
          cpu stop start gave modules #ixmain and #ixxindx.
          mssd shows scpf as current task with cre stack
          #cfszrel,qdbapmgr ,  icb stack #pminpr2 qwciscfr qrcimpln
          qdbrcips #ixmain .
          user has to scratch install again to recover.
         
         
          LOCAL FIX:
         
         
          PROBLEM SUMMARY:
          A 9406 was scratch installed with OS400 and user data.  It
came
          up and appeared to be running fine.  However, it was
discovered
          that the workstation controller resource names where
incorrect.
          The SRM (System Resource Management) data was cleared and the
          system was re-IPLed.  The system appeared to be hung at
          SRCC9002AA0 for 4 hours in SCPF.
         
         
         
         
         
          PROBLEM CONCLUSION:
         
         
          TEMPORARY FIX:
         
         
          COMMENTS:
          During an IPL the system can be at SRCC9002AA0 for long
periods
          (many, many hours) of time.  A list of access paths that have
to
          be rebuilt is being generated.  The most common reasons for an
          access path to have to be rebuilt are that the access path was
          open when an abnormal termination occurred or a file was
          restored but its access path has not yet been rebuilt.  The
list
          generation can be very time consuming because the algorithm
used
          requires sequential searches.  In V3R1M0 the algorithm was
          changed.  This change will improve the performance at the
          SRCC9002AA0.
         
          Also in V3R1M0, a change was made to greatly improve the
          performance of changing all of the sequence numbers or
changing
          them to *HLD or *OPN for rebuilding the access paths at
          SRCC9002AB0.  These changes are made on the EDTRBDAP screen
          during a manual IPL.
         
         

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to "MIDRANGE-L@midrange.com".
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---