PURPOSE
-------
 
The purpose of this bulletin is to assist support analysts in understanding and
resolving the stranded dba_2pc_entries.
 
 
SCOPE & APPLICATION
-------------------
 
The intended audience are support analysts having good experience on distributed
databases.
 
TITLE
-----
 
How to resolve stranded dba_2pc_pending entries
 
 
Contents
========
 
1. Problem Description
2. Solutions
   2.1 Dba_2pc entries without a corresponding transaction
   2.2 Distributed transaction without corresponding dba_2pc entries
 
1. Problem Description:
=======================
 
As a result of a failed commit of a distributed transaction, some entries can
be left in dba_2pc views, i.e. dba_2pc_pending and dba_2pc_neighbors. The RECO
process checks these views to recover the failed txn. However, in some cases
RECO cannot perform the recovery. One cause is that all sites involved in the
transaction not being  accessible at the same  time. Another cause is dba_2pc
views being  inconsistent with the  transaction table, which  is the topic of
this article. This cause can further be classified as follows:
 
   1. dba_2pc views have entries for a non-existent distributed transaction
 
   2. There is a distributed transaction for which there are no entries in
      dba_2pc views
 
Solutions to each subclass is provided in the rest of the article.
 
 
2. Solutions:
=============
 
2.1 Dba_2pc entries without a corresponding transaction
-------------------------------------------------------
 
In this case dba_2pc views show distributed transactions but there are no txns
in reality. If the state of the  transaction is committed,  rollback forced or
commit forced then this is normal and it can be cleaned up using
 
   dbms_transaction.purge_lost_db_entry
 
However, if the state of the transaction is PREPARED and there is no entry in
the transaction table for it then this entry can be cleaned up manually as
follows:
 
    set transaction use rollback segment SYSTEM;
 
    delete from sys.pending_trans$
      where local_tran_id = <xid>;
 
    delete from sys.pending_sessions$ where local_tran_id = <xid>;
 
    delete from sys.pending_sub_sessions$ where local_tran_id = <xid>;
 
    commit;
 
    Example:
    --------
 
    The following query reports a dist. txn. in prepared state
 
       select local_tran_id, state from dba_2pc_pending;
 
           LOCAL_TRAN_ID          STATE
           ---------------------- ----------------
           1.92.66874             prepared
 
    Given that a transaction id is composed of <rbs#, slot#, wrap#> triple,
    '1.92.66874' is located in rollback segment# 1. To find out the list of
    active transactions in that rollback segment, use:
 
      SELECT KTUXEUSN, KTUXESLT, KTUXESQN, /* Transaction ID */
             KTUXESTA Status,
             KTUXECFL Flags
      FROM x$ktuxe
      WHERE ktuxesta!='INACTIVE'
            AND ktuxeusn= 1; <== this is the rollback segment#
 
      no rows selected
 
    It is not possible to rollback force or commit force this transaction.
 
      rollback force '1.92.66874';
 
      ORA-02058: no prepared transaction found with ID 1.92.66874
 
    Hence, we have to manually cleanup that transaction:
 
      set transaction use rollback segment SYSTEM;
 
      delete from sys.pending_trans$
        where local_tran_id = '1.92.66874';
 
      delete from sys.pending_sessions$ where local_tran_id = '1.92.66874';
 
      delete from sys.pending_sub_sessions$ where local_tran_id = '1.92.66874';
 
      commit;
 
 
2.2 Distributed transaction without corresponding dba_2pc entries
-----------------------------------------------------------------
 
In this case dba_2pc views are empty but users are receiving distributed txn
related errors, e.g. ORA-2054, ORA-1591. Normally such a case should not appear
and if it is reproducible a bug should be filed. Here is the list of several
alternative solutions that can be used in this case:
 
     a. Perform incomplete recovery
     b. Truncate the objects referenced by that transaction and import them
        refer to Note 76603.1 for this.
     c. Use _corrupted_rollback_segments parameter to drop that rollback segment
        as discussed in Note 106638.1.
     d. Insert dummy entries into dba_2pc views and either commit or rollback
        force the distributed transaction
 
The first three solutions are discussed in Backup and Recovery manuals and in
the notes referred above. In the 4th solution a dummy entry is inserted into
the dictionary so that the transaction can be manually committed or rolled back.
Note that RECO will not be able to process this txn and distributed txn recovery
should be disabled before using this method. Furthermore, please take a BACKUP
of your database before using this method.
 
The following example describes how to diagnose and resolve this case. Suppose
that users are receiving
 
 ORA-1591: lock held by in-doubt distributed transaction 1.92.66874
 
and the following query returns no rows:
 
    select local_tran_id, state from dba_2pc_pending
    where local_tran_id='1.92.66874';
    
    no rows selected
 
Furthermore querying the rollback segment shows that 1.92.66874 remains in
prepared state
 
    SELECT KTUXEUSN, KTUXESLT, KTUXESQN, /* Transaction ID */
           KTUXESTA Status,
           KTUXECFL Flags
    FROM x$ktuxe
    WHERE ktuxesta!='INACTIVE'
          AND ktuxeusn= 1;  /* <== Replace this value with your txn undo seg#
                                   Which is displayed in the first part of
                                   the transaction ID */
 
      KTUXEUSN   KTUXESLT   KTUXESQN STATUS           FLAGS
    ---------- ---------- ---------- ---------------- ------------------------
             1         92      66874 PREPARED         SCO|COL|REV|DEAD
 
 
Trying to manually commit or rollback this transaction
 
    commit force '1.92.66874';
 
    ORA-02058: no prepared transaction found with ID 1.92.66874
 
raises ORA-02058 since dba_2pc views are empty. In order to use commit force or
rollback force a dummy record should be inserted into pending_trans$ as follows:
 
    alter system disable distributed recovery;
 
    insert into pending_trans$ (
        LOCAL_TRAN_ID,
        GLOBAL_TRAN_FMT,
        GLOBAL_ORACLE_ID,
        STATE,
        STATUS,
        SESSION_VECTOR,
        RECO_VECTOR,
        TYPE#,
        FAIL_TIME,
        RECO_TIME)
    values( '1.92.66874', /* <== Replace this with your local tran id */
        306206,                  /*                                         */
        'XXXXXXX.12345.1.2.3',   /*  These values can be used without any    */
        'prepared','P',          /*  modification. Most of the values are   */
        hextoraw( '00000001' ),  /*  constant.                              */
        hextoraw( '00000000' ),  /*                                         */
        0, sysdate, sysdate );
 
    insert into pending_sessions$
    values( '1.92.66874',/* <==Replace only this with your local tran id */
        1, hextoraw('05004F003A1500000104'),
        'C', 0, 30258592, '',
        146
      );
 
    commit;
 
    commit force '1.92.66874';
 
    If commit force raises an error then note the errormessage and execute the
    following:
 
      delete from pending_trans$ where local_tran_id='1.92.66874';
      delete from pending_sessions$ where local_tran_id='1.92.66874';
      commit;
      alter system enable distributed recovery;
 
    Otherwise run
 
      alter system enable distributed recovery;
 
      and purge the dummy entry from the dictionary, using
 
      connect / as sysdba
 
      alter session set "_smu_debug_mode" = 4;  /* if automatic undo management
                                                   is being used */
      commit;  /* this is to prevent the ORA-01453 in purge_lost_db_entry call */
      exec dbms_transaction.purge_lost_db_entry( '1.92.66874' )
 
http://solo.bloghome.cn/posts/182666.html
 
http://www.oraclefans.cn/forum/showtopic_tree.jsp?boardcode=o5&hit=1026&rootid=7153 
 
http://www.oraclefans.cn/forum/showtopic_tree.jsp?boardcode=o5&hit=908&rootid=7179