Nightly Vacuum analyze to resolve.

You’ll have to create a file on your appliance and run it.

The appliation needs to be up for this script to be up.

It will cause extra IO but no locking. If your using synchronize please do this off peak.

copy the script below into vi

vi vacuum_nightly.sh

chmod u+x vacuum_nightly.sh

./vacuum_nightly.sh

Let me know if there’s any issue

ymurphy@vmware.com

After running this manaually a few times then I’d suggest you run it nightly on a cron.

crontab -e

20 2 * * * /tmp/vacuum_nightly.sh

Output from the script can be seen on stats_collector.log  on the same directory as the script

#!/bin/bash
error=0
errorlog='stats_collector.log'
pbin=`ps -ea -o cmd | grep '[p]ostgres -D' | awk '{ gsub ("postgres$","",$1); print $1}'`
run_log='optimise_output.log'
dbownername=` ps -eo uname:20,pid,pcpu,pmem,sz,tty,stat,time,cmd --sort -time |grep -m 1 'postgres:' |awk '{print $1}'`
databasename=` ps -f -N --sort -pid | grep -m 1 'postgres:' |awk '{print $10}'`

function vacuumaz {

echo $dbownername database owner
echo $databasename database name

if [ $databasename = "secureall" ]; then
 $pbin/vacuumdb -U secureall -d secureall 2>>$errorlog
elif [ $databasename = "vcac" ]; then
 $pbin/vacuumdb -U postgres -d vcac 2>>$errorlog
elif [ $databasename = "VCDB" ]; then
 $pbin/vacuumdb -U postgres -d VCDB > /dev/null 2>>$errorlog
elif
 $pbin/vacuumdb -U postgres -d $databasename > /dev/null 2>>$errorlog ; then
 printf '\n No Errors on vacuumdb \n'
 printf '\n'
 else
 echo 'vacuumdb ERROR. Please review $errorlog file'
 printf '\n'
 error=1
 fi

echo `/bin/date "+%Y-%m-%dT%H:%M:%S"` Finish vacuumdb
echo `/bin/date "+%Y-%m-%dT%H:%M:%S"` Finish vacuumdb >> $errorlog
}

# Main Program

vacuumaz