RMAN recovery incomplete issue while recovering in DG

From: Shastry(DBA) <"Shastry>
Date: Wed, 2 Mar 2011 22:57:40 +0530
Message-ID: <AANLkTimcuC3pcPjMyRswradVfUGNQ42qT6dXUAjXaTfs_at_mail.gmail.com>



Hi All,

We are having the issue with dataguard lag issue from production due to heavy generation of archives and had a challenge to keep in synch of DG with production database. We later tried to use the manual recovery option to see if we can make the recovery faster rather than script which ships the archive and then get applied, but it did not help. Finally we tried with RMAN option like creating a catalog database and registering the DG as level0 (since its 9.2.0.8 version) and taking level 1 at production side and later have the backup piece in DG host to apply RMAN with "recover database no redo".
We are facing challenge in this where RMAN did not errored out in any stage/phase but only few datafiles got updated with the latest SCN and rest were left untouched by RMAN.
Also if the Oracle is on 10g we just need a level1 backup and apply the changes in DG database without any level0 catalog.

What we observed was, there were 27 total backup pieces were generated in level1 backup, but first 10 backup pieces were not cataloged and looks missing (wierd).

  1. The recovery noredo has not applied any changes to the files which are dated 22-FEB and few 28-FEB.
  2. When you start recovery in manual mode, it asks for archive files generated on Feb22
  3. The backupsets which are not listed in RMAN catalog but exists in physical media is of *Size 0* in the rman catalog. *The first 10 backupsets are kicked out*. But the backup piece is in available status and the files returns no rows.

RMAN> *list backupset 198935;*

List of Backup Sets


BS Key Type LV Size Device Type Elapsed Time Completion Time

  • ---- -- ---------- ----------- ------------ ---------------

198935 Incr 1 0 DISK 00:50:01 01-MAR-11

        BP Key: 198569   Status: AVAILABLE   Tag: SCPSTBY
        Piece Name: /oracle/product/9.2.0.8/rman/2pn63c7m


RMAN> *list backup of datafile "/dev/vgssoprd7/rv4000-63";*

returns No rows.

Looks like some threshold value of control is getting hit and the first files/backupsets are kicked out when new files/backupsets are generated.

  1. Checked if itís the file size which is causing issue but it is not as other big files were taken and applied.

Need your help in showing some ideas and light on the same as its 9i version and having challenges.

Database version : 9.2.0.8
Platform : HP-UX

Thanks,
Ananth

--

http://www.freelists.org/webpage/oracle-l Received on Wed Mar 02 2011 - 11:27:40 CST

Original text of this message