Restore database

I need to test and create a procedure to restore databases.
I followed the steps based on the site, but I couldn't finish succesfully.

I did:

  1. Put the database on Backup Mode and copy datafiles.
     /pg/bin/psql cresoldev -c "SELECT pg_start_backup('/pg/backup/');"
     tar -cvf /pg/backup/bk_base.tar /pg/data/base/*
     /pg/bin/psql cresoldev -c "SELECT pg_stop_backup();"

     File .conf: archive_command = 'cp -i %p /pg/backup/xlog/%f </dev/null'


  2. Created a new table and populated with data, to simulate the recovery:
     create table test (
      aa integer,
      bb varchar(50)
     );

     insert into test values (1,'aaa');
     ...
     insert into test values (5,'aaa');

     Data inserted successfully!!!


  3. Shutdown on database;

     Last log transactions copied to the directory archived log;
     cp /pg/data/pg_xlog/* /pg/backup/xlog/


  4. Configuring the recovery.conf file:

     restore_command = 'cp /pg/backup/xlog/%f %p'
     recovery_target_time = '2006-07-06 16:33:52 BRT'


  5. Simulate the lost directories, deleting... :

     rm -r /pg/data/base/*

  6. Recreating the directories exploding the tar file:

     tar -xvf bkp_base... .tar

  7. Starting the database for applying the log transactions.
     Supposing recove the table "test" located on log transactions.

     LOG:  database system was shut down at 2006-07-06 16:47:18 BRT
     LOG:  starting archive recovery
     LOG:  restore_command = "cp /pg/backup/xlog/%f %p"
     LOG:  recovery_target_time = 2006-07-06 16:33:52-03
     cp: cannot stat `/pg/backup/xlog/00000001.history': No such file or directory
     LOG:  restored log file "000000010000000000000001" from archive
     LOG:  record with zero length at 0/1122880
     LOG:  invalid primary checkpoint record
     LOG:  restored log file "000000010000000000000001" from archive
     LOG:  record with zero length at 0/1122844
     LOG:  invalid secondary checkpoint record
     PANIC:  could not locate a valid checkpoint record
     LOG:  startup process (PID 3989) was terminated by signal 6
     LOG:  aborting startup due to startup process failure


  8. There was an error and the table was lost.
     What's the procedure that I doing mistakes????????


Thanks for helping me, in advance.

Regards
Alexander