Restoring Oracle Database with Clones

[jouell]

Hi.


We duplicate our Oracle Databases to test restores and it works great. How do I make NW use my clones volumes instead? I'd prefer not to mark all volumes "suspect".


I tried :

run {

allocate auxiliary channel a1 type 'sbt_tape' parms 'ENV=(NSR_SERVER=XXX,NSR_CLIENT=YYY,NSR_DATA_VOLUME_POOL=Offsite Oracle)';

allocate channel t1 type 'sbt_tape' parms 'ENV=(NSR_SERVER=XXX,NSR_CLIENT=YYY,NSR_DATA_VOLUME_POOL=Offsite Oracle)';

set until sequence 85413 thread 2;
duplicate target database to mydb nofilenamecheck;

}

This seems to still pull from the main oracle pool and not the offsite one in the allocate command.

Do I need to do anything else?

Anyone else succeed in this?

-john

 

[605]

I am in doubt that this works at all although i have to admit that the RMAN script syntax allows a recovery from any pool.

By default, NW looks for the first instance (the original) of a saveset. To recover from a clone, you have these choices:
- set the original save set to "suspect" (what you did)
- you can recover from a clone by specifying the cloneid
recover -S ssid/cloneid

On the other hand, the NMO manual states that NW will use the clone "if the original media is unavailable". For more details, please read chapter 5 (nsrnmoinfo).

 

[jouell]

605,

Thanks for the info. I'll do some more looking into this.
I'll post back what I find , if it differs from what we know now.

-john

 

[jouell]

OK. So, I worked around the networker limitations as you and support suggested.

I am submitting the following script to tek-tips if others would like to pursue this:

(You'll need to grep out the name of your offsite pool in the script....see "Offsite POOL" below)

Regards,
-john

#!/bin/ksh

if [ -z "$5" ];then
  echo syntax $0 client dayofweek  month date db \(yes|no\)
  exit
fi


#Short name
client=$1

#Day
DAY=$2

#Month
MONTH=$3

#Date
DATE=$4

#Database
DB=$5

#State
STATE=$6


case $STATE in
            yes)   STATE=suspect;;
             no)   STATE=notsuspect;;
              *)   echo yes or no - exiting ; exit 1;;
esac

echo
echo
echo Current suspect savesets
echo
echo

#############################################
mminfo -r "ssid,client,cloneid,name,savetime,volume,pool" -q "suspect" | sort -r  -k 5
#############################################

echo
echo
echo Changing status of savesets 
echo
echo

for EVERY in $(

for EACH  in  $(nsrinfo -n oracle $client | grep -i "$DAY $MONTH $DATE" | grep -i $DB | cut -d "=" -f2 | cut -d " " -f1 ); do    

        mminfo -a -v -q "client=$client,savetime=$EACH" -r "client,ssid,savetime,cloneid,volume,pool" | grep -v -i "Offsite POOL" | grep [0-9] | sort; 

done | awk '{ print $2,$4; }' | sed s#' '#/#g

);do

nsrmm -o $STATE -S "$EVERY"  -y;


done

echo
echo
echo New status of savesets 
echo
echo

############################################
mminfo -r "ssid,client,cloneid,name,savetime,volume,pool" -q "suspect" | sort -r  -k 5
############################################

Disclaimer:
Don't hit enter unless you know what you're doing!