RE: Help...RMAN and control file

  • From: "Storey, Robert \(DCSO\)" <RStorey@xxxxxxxxxxxxxxxxxx>
  • To: "Martin Brown" <martinfbrown@xxxxxxxxxxx>, <oracle-l@xxxxxxxxxxxxx>
  • Date: Mon, 30 Aug 2010 12:34:28 -0500

I've not fully implemented my retention schedule. I'm still learning
RMAN.  RMAN got started before I was ready and understood all of these
parameters I needed to do.

 

Does that fact that I don't purge my archive logs make entries into the
control file?

 

From: Martin Brown [mailto:martinfbrown@xxxxxxxxxxx] 
Sent: Monday, August 30, 2010 12:32 PM
To: Storey, Robert (DCSO); oracle-l@xxxxxxxxxxxxx
Subject: RE: Help...RMAN and control file

 

I think you have a couple of options. Depending on your retention
policy, you may want to try "crosscheck archivelog all;" followed by
"delete expired archivelog all completed before 'sysdate-30';" (or some
number you feel comfortable with). 
 
You should add "crosschecks archivelog all;" "crosscheck backup of
database;" and "delete noprompt obsolete;" statements to the start of
the rman scripts. But be sure you check your retention policy and don't
step on stuff you'll need later. 
 

________________________________

Subject: Help...RMAN and control file
Date: Mon, 30 Aug 2010 11:53:37 -0500
From: RStorey@xxxxxxxxxxxxxxxxxx
To: oracle-l@xxxxxxxxxxxxx

Okay,

 

I'm pretty new to RMAN and apparently I've been pretty stupid and not
run some sort of maint to clean up my backups. I bag only that I got
pulled to higher priorities and did not fully understand RMAN.

 

Anyway, I have a RMAN backup job being executed each night at 10PM using
Symantec. So, my backup goes straight to tape. The backup runs fine.
But, since the 7th, I've been getting tons of this in my log:

kccrsz: denied expansion of controlfile section 15 by 6481 record(s)

  the number of records is already at maximum value (65535)

krcpwnc: following controlfile record written over:

 RECID #182457 Recno 3109 Record timestamp 

 07/19/10 22:13:00

 Thread=1 Seq#=19687

  Backup set key: stamp=724803170, count=3409

  Low

scn: 0x0007.5a0bac52

 06/04/10 14:58:59

 Next 

scn: 0x0007.5a0c851a

 06/04/10 15:48:35

  Resetlogs scn and time 

scn: 0x0004.e497f06c

 06/12/08 11:47:44

  Block count=20478  Blocksize=512

kccrsz: denied expansion of controlfile section 15 by 6481 record(s)

  the number of records is already at maximum value (65535)

krcpwnc: following controlfile record written over:

 RECID #182458 Recno 3110 Record timestamp 

 07/19/10 22:13:00

 Thread=1 Seq#=19688

  Backup set key: stamp=724803170, count=3409

  Low

scn: 0x0007.5a0c851a

 06/04/10 15:48:35

 

I have my control_file_record_keep value set to 45 days. Where I 've
been stupid is in not running the Obsolete(?) command to delete and
purge the old backups.  

 

My issue now is that I suspect I will have a bunch of this generated
again tonight at 10PM and I think it interfered with another job I tried
to run.  The tapes containing the older backups are no longer available.

 

So, in my panic scan through the documentation, I'm trying to find a
command that will purge all the obsolete records from my control file
without having the backupsets available on tape.    I've opened a SR
with Oracle, but, this list normally responds faster.  What is the
proper command string I need to issue to get my control file cleaned up,
and not crash my system.

 

So..in a slight bit of a panic till I get his taken care of......

 

THANKS!!

Other related posts: