Re: RMAN 8i question - Disaster Recovery and online redo logs...

Comments embedded...

On 19 Jan 2006 11:06:47 -0800, "BD" <bobby_dread_at_hotmail.com> wrote:

>Hi, all.
>
>I'm still (gak) working through various permutations of rman
>functionality (8i on AIX).
>
>At this point, I'm trying a full-on disaster recovery: backup the db
>with rman, blow it away, and restore/recover with rman.
>
>My basic process is listed below. The symptom I am finding is that when
>I do the restore, rman fails to recover because it cannot find the
>online redo log file. I see that rman does not include the online redo
>logs in its backup (no redo*.dbf files are listed in the backup set).

This has been gone over ad nauseam, in this forum, in Howard Rogers backup course, and on his site http://www.dizwell.com Please let's NOT start it AGAIN!!!!!!!!!!!! You don't NEED the online redolog!

>
>Obviously my process is flawed and my understanding of rman is lacking,
>but I do not see how I can do a full disaster recovery when not all the
>required files are being included in the rman backup. Can someone
>quicly scan my process below, and suggest what I am missing?
>
>Thanks for all feedback!!
>
>NOTE: The backup listed below is an incremental level 0; I retried with
>"full" and the redo*.dbf files were once again not included.

As was to be expected. You don't need them.

>________
>________
>
>*STEP 1 - confirm status of archive logging
>
>SID1> archive log list
>Database log mode Archive Mode
>Automatic archival Enabled
>Archive destination /rs02/oradata/SID1/archivelogs/arch
>Oldest online log sequence 1
>Next log sequence to archive 2
>Current log sequence 2
>
>
>*STEP 2 - Commit change to db
>*STEP 3 - alter system switch logfile
>*STEP 4 - Check archive log folder
>
>SID1:server:/bkup01/rman_backup/SID1> ls -lrt
>/rs02/oradata/SID1/archivelogs/arch
>total 12544
>-rw-r----- 1 oracle dba 6336512 Jan 19 09:41 1_SID1_1.dbf
>
>**DB SHOULD BE READY**
>
>*STEP 5 - set nls_date_format in shell
>
>set nls_date_format='YYYY-MM-DD HH24:MI:SS'; export nls_date_format
>
>
>*STEP 6 - backup db with rman
>
>rman nocatalog target /
>
>run {allocate channel d1 type disk format
>'/bkup01/rman_backup/${ORACLE_SID}/${ORACLE_SID}_full_%s.bak';
> sql "alter session set nls_date_format=''YYYY-MM-DD HH24:MI:SS''"; <------- unnecessary!!!
> backup incremental level 0 database;
> backup archivelog all;}
>
>*STEP 7 - backup control file and password file externally

Should be included in the RMAN script, no need to do that externally, you are also risking 'gaps' in your recovery (what happens if your redolog switches just when you get out of RMAN? The switch will appear in the controlfile, but you don't have included that redolog in your backup.

>
>*STEP 8 - shutdown db, delete all dbf files and the archive log file
>(simulate disaster)
>
>*STEP 9 - recover control file to all locations identified in init.ora
>
>*STEP 10 - restore db with rman
>
>startup mount
>rman nocatalog target /
>run {allocate channel ch1 type disk;
> sql "alter session set nls_date_format=''YYYY-MM-DD HH24:MI:SS''"; <---- yet again unnecessary.

>
> restore database;
> switch datafile all;
> recover database;}

What does this mean? You have *COMPLETE* recovery, which assumes you have the CURRENT controlfile etc. However, you BLEW AWAY the controlfile AND the online redolog!!!! So is this the CURRENT controlfile? You need to specify *INCOMPLETE* recovery, so you need to limit the restore/recovery by specifying a set until scn|time|logseq
command *prior* to running restore and
recover database using backup controlfile.

>
>**predictable parts of rman log clipped**
>RMAN-08054: starting media recovery
>RMAN-08059: media recovery failed
>RMAN-03026: error recovery releasing channel resources
>RMAN-08031: released channel: ch1
>RMAN-00571: ===========================================================
>RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
>RMAN-00571: ===========================================================
>RMAN-03002: failure during compilation of command
>RMAN-03013: command type: recover
>RMAN-03006: non-retryable error occurred during execution of command:
>recover(3)
>RMAN-07004: unhandled exception during command execution on channel
>default
>RMAN-10032: unhandled exception during execution of job step 1:
>ORA-00283: recovery session canceled due to errors
>RMAN-11003: failure during parse/execution of SQL statement: alter
>database recover if needed
>start
>RMAN-11001: Oracle Error: ORA-00283: recovery session canceled due to
>errors
>ORA-00313: open failed for members of log group 1 of thread 1
>ORA-00312: online log 1 thread 1: '/fs04/oradata/SID1/redo_log_1b.dbf'
>ORA-27037: unable to obtain file status
>IBM AIX RISC System/6000 Error: 2: No such file or directory
>Additional information: 3
>ORA-00312: online log 1 thread 1: '/fs03/oradata/SID1redo_log_1a.dbf'
>ORA-27037: unable to obtain file status
>IBM AIX RISC System/6000 Error: 2: No such file or directory
>Additional information: 3

--
Sybrand Bakker, Senior Oracle DBA