V$BACKUP_PIECE vs LIST BACKUP SUMMARY [message #567970] |
Mon, 08 October 2012 10:42 |
|
Karlia
Messages: 40 Registered: May 2011 Location: Algiers
|
Member |
|
|
Hi everybody,
I tried to compare between the outputs of the RMAN COMMAND :
and the results of, SQL command :
SELECT * FROM V$backup_piece where deleted='NO'
I found much more rows in SQL command than in RMAN.
Can someone tell me what could mean the rows given by "V$BACKUP_PIECE" but not by RMAN?
Thanks.
|
|
|
|
|
Re: V$BACKUP_PIECE vs LIST BACKUP SUMMARY [message #568245 is a reply to message #567970] |
Wed, 10 October 2012 06:09 |
|
Karlia
Messages: 40 Registered: May 2011 Location: Algiers
|
Member |
|
|
thank you John for the reply
Ok,I'll try to explain more ,
when i issue :
SELECT COUN(*) FROM V$BACKUP_PIECE where DELETED='NO';
it gives me :
but when i go to RMAN and issue :
i think that i shall the same number of rows , isn't it ?
but here, it gives me only : 181
where are the 56 row missed ?
for example , to be more precise , I will consider the backups that were done on august 31th, so
in sql :
select recid, set_stamp, set_count-7,status, completion_time, device_type , deleted from v$backup_piece
where deleted='NO' and completion_time between sysdate-41 and sysdate-40;
it gives :
1900 792724610 1438 A 31/08/12 DISK NO
1901 792735310 1439 A 31/08/12 SBT_TAPE NO
1902 792735567 1440 A 31/08/12 SBT_TAPE NO
1903 792735872 1441 A 31/08/12 SBT_TAPE NO
1904 792735888 1442 A 31/08/12 SBT_TAPE NO
that means that there are five (05) BACKUPS completed on August the 31th and not yet deleted
but when i go to RMAN and issue :
RMAN> list backup summary completed between 'sysdate-41' and 'sysdate-40';
it gives :
List of Backups
===============
Key TY LV S Device Type Completion Time #Pieces #Copies Compressed Tag
------- -- -- - ----------- --------------- ------- ------- ---------- ---
1438 B F A DISK 31-AUG-12 1 1 NO TAG20120831T011650
1440 B F A SBT_TAPE 31-AUG-12 1 1 NO TAG20120831T041926
1442 B F A SBT_TAPE 31-AUG-12 1 1 NO TAG20120831T042448
only three (03) rows !!! so, where are the tow (02) miessed rows ?
|
|
|
|
|
|
|
Re: V$BACKUP_PIECE vs LIST BACKUP SUMMARY [message #568277 is a reply to message #568275] |
Wed, 10 October 2012 08:12 |
John Watson
Messages: 8964 Registered: January 2010 Location: Global Village
|
Senior Member |
|
|
I can't simulate what you are seeing. The primary key to foreign key join columns of v$backup_set to v$backup_piece are SET_STAMP and SET_COUNT, do you really have rows in v$backup_piece that do not have a parent row in v$backup_set?
|
|
|
|
Re: V$BACKUP_PIECE vs LIST BACKUP SUMMARY [message #568288 is a reply to message #568287] |
Wed, 10 October 2012 09:18 |
John Watson
Messages: 8964 Registered: January 2010 Location: Global Village
|
Senior Member |
|
|
Quote:But , i don't know why do you insist on V$BACKUP_SET! Your RMAN command LIST BACKUP SUMMARY is almost certainly running a query that joins v$backup_set to v$backup_piece, but we don't know what that query is. So in order to reconcile what RMAN tells you with what the views against the repository tell you, you have to guess at what the query is: try to reverse engineer it. Otherwise you will never be able resolve the apparent contradiction. Because that is probably what your problem is: understanding what those views really mean and how RMAN is presenting the information.
It is unlikely (though always possible) that you have a corrupted RMAN repository.
|
|
|
|
|
|
|
|
|
|