Page tree
Skip to end of metadata
Go to start of metadata

WARNING: Use this Page only, if previous consistency checks had been performed already!  

Overview

Here you can find steps to perform that are either required by SAP Product Support or needed to get rid of blockcorruption!

PLEASE NOTE: General rule for corruption issues is to RESTORE and RECOVER a valid and corruption-free backup or at least parts of it.   SAP strongly recommends and informs customers about the importance of backup strategy and highlights that the existence of a valid backup is ESSENTIAL! In case there is no valid backup existing that may be used you as customer have the use one of the below available solutions.

Questions that have to be cleared and answered on customer side!

  • Is it possible to restore a database backup?
  • How many corrupted blocks do the database have?
  • What is the root cause?

Flowchart to solve Oracle Block Corruption issues

This procedure is valid for every SAP customer and clearly states the steps tp be taken. 



General Action Plan to follow

Below table shows the steps for the individual parties. Clearly visible, which step has to be performed by who. 


ActionResponsible Party
I. Test OS connection to the DB machine


II. Save the status quo
1. Perform a backup of the currently corrupt   DBCUSTOMER
2. Provide SAP the estimated time when backup   will be finished after some files have been backed upCUSTOMER


III. Check if change log between last good   backup and now is free of corruptions (other machine(s) required)
1. Restore a backup not containing   corruptions to another machine and apply the change logCUSTOMER
2. If all changes can be applied, the   corruption issue can be solved without data loss. If not either a repair of   the currently corrupt DB CUSTOMER
      or a reset to   an point in time is to be considered.
     [Optional, but   typical]:
     Because the backups   have not been checked in the past, the last good backup is not known. 
     The before   mentioned steps need to be done potentially for multiple backups: 
     the last   ASSUMED good backup, the oldest available backup as of which all changes up   to now are available
3. Provide SAP the estimated time when   restore will be finished after some files have been restoredCUSTOMER
4. Evaluate the change log volume to be   appliedCUSTOMER
5. Restore  the needed change log (in   parallel to restore of backup)CUSTOMER
6. Provide SAP the estimated time when   application of the change log will be finished after some change log files   have been appliedCUSTOMER
  
IV. Evaluate the volume of damage to decide   if the better option is to repair the currently corrupt DB or to reset the DB   to an older point in time
1. Do a full consistency check on the corrupt   productive DB to see the volume of corruptionsCUSTOMER
2. Tell SAP the estimated time when   consistency check will be finished after some files have been checkedCUSTOMER
3. Tell SAP new corruptions when loggedCUSTOMER
4. Evaluate possibilities to repair the   corrupted tables together with application support CUSTOMER with SAP
     (data needed,   cross-dependencies, etc.) (SAP DB Support/Consistency Group/Application   Developement)
  
V. Evaluate the root cause to exclude that   further corruptions will be generated
1. Check for special maintenance activity,   extraordinary errors on all levels below the DB  (OS, IO,   storage, HW, etc.) before/while corruptions have been occuredCUSTOMER with HW partner
2. Check the correctness of ALL pieces of HW   as careful as possible  ("a   damaged cable need not to write a log entry")CUSTOMER with HW partner
3. In case root cause cannot be found until   wanted go-live: CUSTOMER with HW partner
     Perform Go-live on completely different HW or at least set up a "Pseudo Standby DB"CUSTOMER with HW partner
                                     - Copy   the currently corrupt DB to another machine with COMPLETELY different    HW than production ("Pseudo Standby DB")  CUSTOMER with HW partner
                                        in case   only few corruptions occured, they can be repaired but root cause not found   until go-live
     The "Pseudo   Standby DB" recovers permanently the changes of production so that it   can be used to switch over in case of new corruptions occur after go-live


# Action Plan for SAP


I. Monitor progress, coordinate actions,   double check customer informationSAP
II. Ensure OS access to DB machine (telnet,   ssh etc)SAP
              - Setup a OS connection to the DB machineCUSTOMER

Navigating to other Chapters


General Information

Detecting Corrupted Blocks

Consistency Check with 'ANALYZE' command

Consistency Check with DBV

Consistency Check with RMAN

Consistency Check with Data Pump

Which object is stored in the corrupted block(s)