How to get the correct dbase name in UNIX

[Suman_7]

I have multiple databases in the HP UX box.
for example

box name : HPUX1
dbase1: dev1
dbase2:dev2
dbase3:dev3
dbase4:qa1
dbase5:qa2
dbase6:qa3


I have to promote a shell script which basically tests whether the dev database is up if it is being executed with dev env file
if the script is being executed with qa env file it has to check for qa databases.


Presently I have hardcoded the dbase name in my script for examplce:
dbase1=dev1
dbase2=dev2
dbase3=dev3


but my concern is if this script gets promoted and someone wants to execute it with qa environments then it will still check for the dev baseses dev1, dev2, dev3. How do I dynamically change these parameters if thescript is being executed in qa to qa1, qa2, qa3.


Thank You very much.

[KapilRaj]

We keep such in a local file in every node not in the script and the script does a "cat" to this file to get the correct file.

Something like , hv a /std/local/DBS/env which has these and script looks for this file on every node.

Regds,

Kaps

[chapaya_1]

Hi ,

Try with ps -ef |grep -i pmon ,
then use awk to cut the db name .

BYE

[Victor Fridyev]

Hi,

According to Oracle rules, the DBs which work on a computer, should be written to a file called /etc/oratab.
The format of each entry of the file is:
SID:ORACLE_HOME:Y
for the DB, which should be started by dbstart, or
SID:ORACLE_HOME:N
if you don't want to start the DB by dbstart.
e.g.:
QA1:/home/oraprod/ORA8174:Y
QA2:/home/oraprod/ORA9i:N

Reading /etc/oratab file, you can obtain all SIDs and check the DB status with
ps -fuoracle|grep ora_pmon_$SID

HTH

[Arturo Galbiati]

Hi,
for OraSid in $(grep -E -v "^#|^$" /etc/oratab|cut -d: -f1);do
# setup the rigth Oracle environment
export ORACLE_SID=$OraSid
ORAENV_ASK=NO; . oraenv

# put your checks here

done

HTH,
Art

[Ross Barton]

Write a script that gets run whenever you want to administer any of your DB's. This script should read a config file with your DB names - and depending on the way it is called set up your DB name for you (using the info on the command line and the info in the config file).

Setup the config file so it exists on all the boxes that you run DB's on - then NFS mount it (if you use NFS) so you only have to maintain and publish one copy.

You can even call the script from within the .profile (at login) of any user that is used to administer a single DB. If your admin user is used to administer multiple DBs (such as is typical on ORACLE and INGRES) then you can still just call it from the command line each time you login to adminsiter a set of DB's.

[Alexander M. Ermes]

Hi there. Perhaps this kicks off some idea :

cat /etc/oratab | while read LINE
do
case $LINE in
\#*) ;; #comment-line in oratab
*)
# Proceed only if third field is 'Y'.
if [ "`echo $LINE | awk -F: '{print $3}' -`" = "Y" ] && [ "${Db_Name}" = "none" ]
then
ORACLE_SID=`echo $LINE | awk -F: '{print $1}' -`
oracle_sid=`echo $ORACLE_SID|tr "[A-Z]" "[a-z]"`

.....


Rgds
Alexander M. Ermes