Re: [Networker] Advice on backup strategy for Oracle RMAN backups?

Hi,
> Uwe Weber wrote:

> How would NetWorker know when to back up the data on disk? In other
> words, how could you set it up so that NetWorker wouldn't attempt to
> back it up if RMAM wasn't finished sending it to disk ye

If you have set controlfile autobackup to on in RMAN it will 
backup the current cf after every succesfull backup operation.
So -depending on your setup- you will find quite a few
cf backups in the directory you specified. 

IIRC you can specify pre- and postscripts in nsrnmo. That would be
the place to put the name of a script that does something like
'save /path/mycfbackups'. This is my nsrnmo.sh template:

# Optional Variable: POSTCMD
# Default value: NONE
# Description:  This variable can be used to run a command or command 
# script after nsrnmostart has completed. It will be launched once for
# every saveset entered in the client setup.
#
POSTCMD=

> 1. Just set the start time for the NetWorker group far enough beyond
> the start time
> for the RMAN cron job on the client so as to allow enough of a
> window? In other
> words, guesstimate?

RMAN cronjob? Do you not start the Oracle backups from a NW group?
If the postcommand is no option, you can create a group that
only backups  /path/mycfbackups. And schedule this group to run after
the oracle group. For the unlikely event that RMAN is still writing to 
the cf backup while the save is running you will get the usual 
'warning: file_name changed during save'. But controlfiles will
be typically smaller than 100 MB, so if you give NW maybe 1/5 H
between the groups, this should be no problem. 

For additional security your DBA should multiplex the controlfiles
and you could send the RMAN cf autobackups to a diferent disk
or another machine.     

> 2. Create a cron job on the primary backup server that does something
> like:
>
> savegrp group1
> savegrp group2
>
> were the NSR client resource for the given client is a member of
> group1 and it has the
> Oracle RMAN script as its saveset,but instead of the script backing
> it up to tape, it instead sends it to
> disk. Next, another NSR client resource for the same client is a
> member of group2, and its saveset
> is the affected area on disk where the data gets written, so the
> second savegrp command will back that
> area up to tape but not until the first completes. Would this be
> reasonable?

Not for normal backups, I am afraid. I assume that you would 
delete the RMAN backups after transferring them to tape and
then RMAN would not know where to look for it's backuppieces,
so instead of just saying 'restore database' at the RMAN prompt
you would have to restore the necessary backuppieces from tape
to disk and then use RMAN. And you would have paid for NMO without
using it :->.

It is much easier with backing up the cf autobackups , b/c
you can use the restore controlfile from RMAN command.
 
Example RMAN config and backup session:

RMAN> CONFIGURE CONTROLFILE AUTOBACKUP ON;

new RMAN configuration parameters:
CONFIGURE CONTROLFILE AUTOBACKUP ON;
new RMAN configuration parameters are successfully stored

RMAN> CONFIGURE CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK 
TO '/opt/oracle/orabackup/%F';

new RMAN configuration parameters:
CONFIGURE CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK 
TO '/opt/oracle/orabackup/%F';
new RMAN configuration parameters are successfully stored
RMAN> backup database;
Starting backup at 19-JAN-07
[snip session log of database backup]
Finished backup at 19-JAN-07
Starting Control File and SPFILE Autobackup at 19-JAN-07
piece handle=/opt/oracle/orabackup/c-2607733149-20070119-00 comment=NONE
Finished Control File and SPFILE Autobackup at 19-JAN-07

Example restore:
RMAN> run
2> {
3> allocate channel c1 device type disk;
4> restore controlfile 
from '/opt/oracle/orabackup/c-2607733149-20070119-00';
5> }

allocated channel: c1
channel c1: sid=156 devtype=DISK

Starting restore at 19-JAN-07

channel c1: restoring control file
channel c1: restore complete, elapsed time: 00:00:02

RMAN> recover database;

Starting recover at 19-JAN-07
Starting implicit crosscheck backup at 19-JAN-07
allocated channel: ORA_DISK_1
channel ORA_DISK_1: sid=156 devtype=DISK
Crosschecked 1 objects
Finished implicit crosscheck backup at 19-JAN-07

Starting implicit crosscheck copy at 19-JAN-07
using channel ORA_DISK_1
Finished implicit crosscheck copy at 19-JAN-07

searching for all files in the recovery area
cataloging files...
no files cataloged

using channel ORA_DISK_1

starting media recovery

archive log thread 1 sequence 2 is already on disk as 
file /opt/oracle/oradata/testhome/redo01.log
archive log filename=/opt/oracle/oradata/testhome/redo01.log thread=1 
sequence=2
media recovery complete, elapsed time: 00:00:02
Finished recover at 19-JAN-07

RMAN> alter database open resetlogs;

database opened

To sign off this list, send email to listserv AT listserv.temple DOT edu and 
type "signoff networker" in the body of the email. Please write to 
networker-request AT listserv.temple DOT edu if you have any problems with this 
list. You can access the archives at 
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER