DB2 Backup Script

After three months of my continuous hit’s i am successful in writing a backup shell script . Its really nice feeling you get when your first script is successfully executed and is ready to go to production server

So no wasting time here is what i have written

Here are the things what i have done in this script

1. Check for the space in the file system
2. Check if the backup is already taken
3. Take the backup
4. Copy the backup taken to another server
5. Remove the 5 days old backup (According to Retention policy)

And Here Goes the script

	# What does this script do
	Check for the space in the server
	Take a backup (full online compress in this example)
	Copy it to another server
	Remove Old Backup’s
	if [ -f /home/db2inst1/sqllib/db2profile ]; then
	. /home/db2inst1/sqllib/db2profile
	fi
	#!/bin/sh
	# Check for the space in specific partition you are taking backup,in my case it’s /dev/md0 , type df –h and you will know yours and replace /dev/md0 with that
	space=`df | grep /dev/md0 | awk '{print $4}'`
	# I checked for space to < 1.1G , Change it to your requirements
	if [ $space -lt 111111 ]; then
	echo -e "No enough space to take backup , Clear some space in the filesystem for the Backup to be Taken\n" | tee –a /path/backup_logs.out
	else
	echo -e "--------------------------------------------------\n" | tee –a /path/backup_logs.out
	echo -e "Backup Summary DB-NAME for date:`date +%Y-%m-%d`\n" >> /path/backup_logs.out
	echo -e "--------------------------------------------------\n" | tee –a /path/backup_logs.out
	# Check if the backup already exists else take a backup
	if [ -f /path/DB-NAME.0.db2inst1.NODE0000.CATN0000.`date +%Y%m%d`* ] ; then
	echo -e "The backup for this `date +%Y-%m-%d` already exists " | tee –a /path/backup_logs.out
	else
	/home/db2inst1/sqllib/bin/db2 connect to db-name
	/home/db2inst1/sqllib/bin/db2 backup db db-name online to /path/ compress include logs
	# Check for the exit status and report back up successfully done or not
	if [ "$?" -eq 0 ]; then
	echo -e "Backup for DB-NAME completed successfully for the date:`date +%Y-%m-%d`\n" | tee –a /path/backup_logs.out
	else
	echo -e "Backup for DB-NAME did not complete successfully for the date:`date +%Y-%m-%d`\n" | tee –a /path/backup_logs.out
	fi
	/home/db2inst1/sqllib/bin/db2 connect reset
	# Make the server idle for some time because immediately after taking backup the load on the server would be high , so to avoid that
	sleep 90s
	# Start Copying the backup and log the output, To make this happen you need to make the two servers Paswordless , for that you need to generate a rsh pub and priv key and add pub key to the target server so that you will not be prompted for the password
	echo -e "Started copying backup for the date:`date +Y-%m-%d` to serverIP:/path-to-copy \n" >> /path/backup_logs.out
	scp /path/DB-NAME.0.db2inst1.NODE0000.CATN0000.`date +%Y%m%d`* username@serverIP:/path-to-copy
	echo -e "Copying Completed at `date`\n" | tee –a /path/backup_logs.out
	sleep 10s
	fi
	# Remove Old Backup according to retention policy , In this example I have done it for 5 days
	if [ -f /path/DB-NAME.0.db2inst1.NODE0000.CATN0000.`date +%Y%m%d -d "5 days ago"`* ]; then
	rm /path/DB-NAME.0.db2inst1.NODE0000.CATN0000.`date +%Y%m%d -d "5 days ago"`*
	echo -e "The Backup taken on `date +%Y%m%d -d "5 days ago"` has been removed\n" | tee –a /path/backup_logs.out
	else
	echo -e "The Backup for the date:`date +%Y%m%d -d "5 days ago"` was not found , so no file is deleted\n" | tee –a /path/backup_logs.out
	fi

view raw db2backupscript.sh hosted with ❤ by GitHub

 To copy the file from one server to other we need to adopt a concept called password less ssh . Please refer this how to make the transfer between 2 servers without invoking the password

http://db2administration.blogspot.com/2013/01/password-less-ssh.html

Refer this post for the Crontab scheduling to drop a mail of the output to your INBOX
http://db2administration.blogspot.com/2013/01/how-do-you-cronjob-script-to-mail-it-to.html

Download db2script.sh