started my bash script with issues

apss_evaluator · 06-25-2015, 11:18 AM

Hi Linux Guru's

I'm still not done, but I fully assumed this should work, can you please help me to point what is wrong, I keep getting this message:

[ungas@pogi gwapo]$ ./pogi.sh
Detected /var/tmp/start_host folder is removed, created the folder and start_host.log now
/opt/pgdata is in normal permission values
./pogi.sh: line 45: [: : integer expression expected
Postgres is down
postgres needs to be checked before starting the apps

Code:

#!/bin/bash


# 06424015 - Script composition by Poging gwapo

# pogi home directory
#pogi_HOME=/opt/pogi
# Check the pogi of this box
#POGI=$(ls -r ${pogi_HOME}/pkg | grep jboss | tail -1 | cut -d'-' -f2)
# Check if postgres is running by posting the PID
POSTGRES_PID=$(ps -ef | grep postgres | grep data | /bin/awk '{print $2}')
#echo $POSTGRES_PID
# Check Postgres permission
POSTGRES_FLDR=$(stat -c a /opt/pgdata)


# Check if logging directory is available, if not create one

if [ ! -d '/var/tmp/start_host' ]; then
        mkdir /var/tmp/start_host
        sleep 5
        echo "Detected /var/tmp/start_host folder is removed, created the folder and start_host.log now"
   else
        echo "Verified /var/tmp/start_host exists, created the start_host.log now"
fi



# Check the status of postgres and create a marker in /var/tmp/start_host if the condition is okay


# Check if Postgres folder is in right permission

if [ "$POSTGRES_FLDR" != "700" ] ; then
        echo "Error: /opt/pgdata directory has wrong permissions"
  else
        echo "/opt/pgdata is in normal permission values"
        touch /var/tmp/start_host/pg_folder_ok.txt
#  exit 1
fi


# Check if Postgres is up

if [ "$POSTGRES_PID" -ne "" ]; then
        echo "Postgres up and $POSTGRES_PID PID"
        touch /var/tmp/start_host/pg_up.txt
  else
        echo "Postgres is down"
fi

# Create a counter, if the count is equal to 2 then apps should start

        L=0
        [ -e /var/tmp/start_host/pg_folder_ok.txt ] && (( L++ ))
        [ -e /var/tmp/start_host/pg_up.txt ] && (( L++ ))

                if [ L == 2 ];
                        then
                                echo "apps are good to start now"
                        else
                                echo "postgres needs to be checked before starting the apps"
                fi

the error seems to be on this part, but as far as I know this is okay and the script does not create pg_up.txt

Code:

if [ "$POSTGRES_PID" -ne "" ]; then
        echo "Postgres up and $POSTGRES_PID PID"
        touch /var/tmp/start_host/pg_up.txt
  else
        echo "Postgres is down"
fi

pan64 · 06-25-2015, 11:47 AM

insert set -xv at the beginning to see what's happening.
I assume the in line 20 the returned value was not a number, but I don't know what's really happened.

apss_evaluator · 06-25-2015, 11:50 AM

Quote:

Originally Posted by pan64

insert set -xv at the beginning to see what's happening.
I assume the in line 20 the returned value was not a number, but I don't know what's really happened.

Hi Pan64,
thanks for your reply, when you say "set -xv" was that run the script and add that as a parameter?

grail · 06-25-2015, 12:33 PM

Put set -xv as the second line of the script (the first being the shebang)

millgates · 06-25-2015, 12:42 PM

Code:

if [ "$POSTGRES_PID" -ne "" ]; then

-ne is a numeric comparison. You can't use an empty string as an argument here.

works:

Code:

if [ "$POSTGRES_PID" != "" ]; then

better:

Code:

if [ -n "$POSTGRES_PID" ]; then

even better:

Code:

if [[ -n "$POSTGRES_PID" ]]; then

apss_evaluator · 06-25-2015, 06:26 PM

Quote:

Originally Posted by grail

Put set -xv as the second line of the script (the first being the shebang)

thanks

apss_evaluator · 06-25-2015, 06:27 PM

Quote:

Originally Posted by millgates

Code:

if [ "$POSTGRES_PID" -ne "" ]; then

-ne is a numeric comparison. You can't use an empty string as an argument here.

works:

Code:

if [ "$POSTGRES_PID" != "" ]; then

better:

Code:

if [ -n "$POSTGRES_PID" ]; then

even better:

Code:

if [[ -n "$POSTGRES_PID" ]]; then

thanks millgates, I followed "even better" and issue resolved

apss_evaluator · 06-30-2015, 05:21 AM

thanks all for sharing, I made it work with the validation on the postgres condition. However (I hate to say this) after hours of googling I'm still not able to perform the loop validation

this is the validation part

Code:

#!/bin/bash
POSTGRES_Status()
{
  POSTGRES_PID=`ps -ef | grep postgres | grep data | /bin/awk '{print $2}'`
  if [ -z $POSTGRES_PID ]; then
   POSTGRES_PID=0
  fi
}
POSTGRES_Status
echo "$POSTGRES_PID"
#for i in 1 2 3
#do
   POSTGRES_Status
        if [ "$POSTGRES_PID" != "0" ]; then
                for i in 1 2 3
                        do
                                echo "postgres is now running in PID $POSTGRES_PID"
                        done
        else
                echo "detected postgres is not running, script will try to start postgres"
                        sudo -u postgres /opt/msp/pkg/postgres/bin/pg_ctl -D /opt/msp/pkg/postgres/data start
        fi

RESULT OF MY TEST SCRIPT ( I intentionally changed the permission of /opt/pgdata/PG_VERSION )

Code:

[ungas@pogi gwapo]$ ./script.sh
0
detected postgres is not running, script will try to start postgres
server starting
[ungas@pogi gwapo]$ 2015-06-30 10:13:23.604 GMT [25759][@] : [1-1]FATAL:  could not open file "/opt/pgdata/PG_VERSION": Permission denied
^C - ridiculously stuck on this part I just hit the control C

I'm trying the above part to loop this way

1> check if postgres is up with pid
2> if postgres is up loop will end and the running script as well
3> if postgres is down it will try start
4> if 1st try is not success it will try to turn up again on 2nd time
5> if 2nd try is not success it will try to turn up again on 3rd time
6> if 3rd try fail loop will end and will tell that "start posgres failed to up after three tries" and the script will end