cancel
Showing results for 
Search instead for 
Did you mean: 

Restoring an Alternative Copy of an Oracle Backup

Auda_SD
Level 2

Hi,

I am trying to restore an Oracle database however I don't want to restore the primary copy

I found an article and added the ALT_RESTORE_COPY_NUMBER file in /usr/openv/netbackup on the Masterserver. This file contains only the number 4 which is the copy number I am interested in

It seems however that when I run my oracle rman script (see below), it uses copy 2 not copy 4. Anyone any ideas why??

I am running Netbackup version 7.5.0.4

run {
ALLOCATE CHANNEL Controlfile_channel TYPE 'SBT_TAPE';
SEND 'NB_ORA_POLICY=Oracle_database,NB_ORA_SCHED=Controlfile_Spfile,NB_ORA_CLIENT=esprd-ora01.har.uk.host.com,NB_ORA_SERV=ukprd-nbmas02.har.uk.host.com';
restore spfile;
}

Here is a section of the output which shows that copy 2 is being used......

 Info bpbrm(pid=1862) accepted connection from client        
07/01/2013 16:33:13 - Info tar(pid=16286) Restore started.          
07/01/2013 16:33:13 - connected
07/01/2013 16:34:56 - Info bptm(pid=1869) Waiting for mount of media id H00221 (copy 2) on server ukprd-nbmed01.
07/01/2013 16:34:56 - started process bptm (1869)
07/01/2013 16:34:56 - mounting H00221
07/01/2013 16:34:56 - Info bptm(pid=1869) INF - Waiting for mount of media id H00221 on server ukprd-nbmed01 for reading.

Any help appreciated
 

3 REPLIES 3

Nicolai
Moderator
Moderator
Partner    VIP   

Are you sure you got 4 copys ?

bpimagelist -backupid hostname_12345678 -L

where hostname_12345678 is the image id. Instead of using ALT_RESTORE_COPY_NUMBER you can promote a secondary copy to a primary copy see tech note:

http://www.symantec.com/docs/TECH64592

http://www.symantec.com/docs/HOWTO33348

Auda_SD
Level 2

Yes I have all the copies (see the bpimagelist command below)

I also see them as completed when I run nbstlutil list -U -backupid xxxxx

The copy promotion option is not really viable as our Oracle databases have a separate backup image for each oracle file. This means a full backup can consist of up to 1500 backups. This is why I was looking for a way of doing it with a single change

 bpimagelist -l -backupid esprd-ora01.har.uk.audatex.com_1357366287


IMAGE esprd-ora01.har.uk.audatex.com 0 0 9 esprd-ora01.har.uk.audatex.com_1357366287 Oracle_NBTEST 4 *NULL* oracle Full_DB 2 0 1357366287 20 1360044687 0 0 544 1 4 4 0 Oracle_NBTEST_1357366287_UBAK.f *NULL* *NULL* 0 2 0 0 0 NBTEST 0 0 0 0 0 0 0 *NULL* 0 0 0 *NULL* 598679 4 0 682 0 0 Platinum 4PT_ORACLE_NBTEST_FULL 3 0 0 0 *NULL* *NULL* 0 107 0 0
HISTO 0 0 0 0 0 0 0 0 0 0


FRAG 2 1 544 0 2 6 39 H00242 ukprd-nbmed02 262144 1279199 1357424778 7 64 *NULL* 1360044687 0 65539 0 0 0 1 1360044687 1357458783 1 1 37f7e6d6170c19ecf6ee590155db42873b304b0b40f7cc724bce84d63b178f11 *NULL* 0 0
 

FRAG 3 1 544 0 0 0 0 @aaaa6 ukprd-nbmed02 262144 0 0 -1 4 1;PureDisk;ukprd-nbpure01.har.uk.audatex.com;dp_Harrogate_Puredisk_Disk_Pool;PureDiskVolume;0 1360044687 0 65539 0 1 0 6 1360044687 1357458783 1 1 *NULL* *NULL* 0 0


FRAG 4 1 544 0 0 0 0 @aaaag ukprd-nbmed02 262144 0 0 -1 4 1;PureDisk;ukdr-nbpure01.rdg.uk.audatex.com;Reading_Puredisk_Disk_Pool;PureDiskVolume;0 1360044687 0 65539 0 2 0 6 1360044687 1357480060 1 1 *NULL* *NULL* 0 0


FRAG 6 1 544 0 2 6 132 R00086 ukdr-nbmed02 262144 3567354 1354892657 8 64 *NULL* 1360044687 0 65539 0 0 0 1 1360044687 1357560733 1 1 f96c7aff48759ced45b15b9e850b96c44b6be39d0d891c1c789320cfe754c39c *NULL* 0 0
 

Nicolai
Moderator
Moderator
Partner    VIP   

Humm - 1500 images - are you running with RMAN with  FILESPERSET=1. I do see the issues of promoting 1500 images. 

I took a look around but did not find any bug notes related to ALT_RESTORE_COPY_NUMBER and Netbackup 7.5.0.4

The only advice I can give you with the information present is to retry the operation assisted by a second pairs of eyes to ensure typos can be excluded.

This T/N is properly is the you are referring to:

http://www.symantec.com/docs/TECH64592