PostgreSQL Installation types

Postgres database installation 3 types, they are

1. GRAPHICAL Installation (./ postgresql.run).
2. BINAY INSTALLATION (RPM).
3. SOURCE CODE INSTALLATION (./configure)

2. POSTGRES BINARY INSTALLATION :
     ============================

Step 1. As root user install the below packages 

postgresql-infrastructure	Basic Clients and Utilities for PostgreSQL
postgresql113	PostgreSQL client programs and libraries
postgresql113-contrib	Contributed source and binaries distributed with PostgreSQL
postgresql113-devel	PostgreSQL development header files and libraries
postgresql113-docs	Extra documentation for PostgreSQL
postgresql113-libs	The shared libraries required for any PostgreSQL clients
postgresql113-plperl	The Perl procedural language for PostgreSQL
postgresql113-plpython	The Python procedural language for PostgreSQL
postgresql113-pltcl	The Tcl procedural language for PostgreSQL
postgresql113-server	The programs needed to create and run a PostgreSQL server
postgresql113-test	The test suite distributed with PostgreSQL

rpm -ivh postgresql-infrastructure-1-6.noarch.rpm
rpm -ivh postgresql113-libs-11.3-1.x86_64.rpm
rpm -ivh postgresql113-server-11.3-1.x86_64.rpm
rpm -ivh postgresql113-contrib-11.3-1.x86_64.rp
rpm -ivh postgresql113-devel-11.3-1.x86_64.rpm
rpm -ivh postgresql113-docs-11.3-1.x86_64.rpm
rpm -ivh postgresql113-plperl-11.3-1.x86_64.rpm
rpm -ivh postgresql113-plpython-11.3-1.x86_64.rpm
rpm -ivh postgresql113-pltcl-11.3-1.x86_64.rpm
rpm -ivh postgresql113-server-11.3-1.x86_64.rpm
rpm -ivh postgresql113-test-11.3-1.x86_64.rpm

Step 2. Now create a linux  user to access the postgres database  and create  data directory for initializing database cluster. Owner of this data directory should be postgres user and permissions should be 700 and also set path for postgresql binaries for our ease.

# useradd nijam
# passwd nijam
# mkdir -p /data_11.3/
# chown -R nijam. /data_11.3/

Step 3. Initialize the postgres cluster as user nijam.

$ /opt/11.3/bin/initdb -D /data_11.3/ -U nijam

->>> Where -D is location for this database cluster or we can say it is data directory where we want to initialize database cluster, -U for database superuser name.

Step 4. put the Web/Apps details on pg_ha.conf file.

# TYPE  DATABASE        USER            ADDRESS                 METHOD

# "local" is for Unix domain socket connections only
local   all             all                                     trust
# IPv4 local connections:
host    all             all             127.0.0.1/32           trust
# IPv6 local connections:
host    all             all             ::1/128                 trust
# Allow replication connections from localhost, by a user with the
# replication privilege.
#local   replication     nijam                                  trust
#host    replication     nijam          127.0.0.1/32            trust
#host    replication     nijam          ::1/128                 trust
host    all             all             127.0.0.1/32            trust
host    all             all             54.99.18.56/32          trust
host    all             all             54.99.17.86/32          trust

Step 5 . Allocate the  memory and connection as per the  CPU and RAM.

For Example if you have 100 GB  RAM.

1. Max_connection =1000×16mb(work_mem)=16 GB
2. Effective_cache_size=40gb optimized execution path.the query planner also requires some space 
3. Shared_buffer=25GB,
4.Wall_buffer=3% of shared buffer,if lot of insert, update, delete set maximum 8mb
5.Maintenance_work_mem=4GB to 6GB(RAM/8) for vacuum,create,alter, reindex, backup/restoring dump.
6.Temp_buffer=8mb default session-local buffers used only for access to temporary tables.

Including this change the below parameter :

$ cat postgresql.conf
listen_addresses = '*'         
port = 5420  
logging_collector=on
log_filename = 'postgresql-%a.log'
log_directory = 'log'
log_rotation_age = 1d 
archive_mode=on
archive_command = 'gzip < %p > /srv/postgresql/var/inst1/log/archive_logs/%f'

Step 6. Startup the new cluster as user nijam .

/opt/11.3/bin/pg_ctl -D /data_11.3/ restart

Step 7. Set the bash profile to access the postgres utility on user nijam home path

$ cat /home/nijam/.profile
export PATH=/opt/11.3/bin:$PATH
export PGHOME=/opt/11.3/
export PGDATA=/data_11.3/
export LD_LIBRARY_PATH=/opt/11.3/lib
export PGDATABASE=nijam
export PGUSER=nijam
export PGPORT=5420
export PGLOCALEDIR=/opt/11.3/share/locale
export MANPATH=$MANPATH:/opt/11.3/share/man

Step 8. Run the bash profile and try to  connect the database 

$ ..profile
$ psql -p 5420
postgres=# \l --- to list all databases in cluster
postgres=# \q --- to quit form postgres console

3. SOURCE CODE INSTALLATION :
     =======================

1. First install required prerequisites such as gcc, readline-devel and zlib-devel using package manager as shown.

# yum install gcc zlib-devel readline-devel [On RHEL/CentOS]
# apt install gcc zlib1g-dev libreadline6-dev [On Debian/Ubuntu]
# Zypper in gcc zlib1g-dev libreadline6-dev [On SUSE Linux ]

2. Download the source code tar file from the official postgres website using the following wget command directly on system.

# wget https://ftp.postgresql.org/pub/source/v11.3/postgresql-11.3.tar.gz

3. Use tar command to extract the downloaded tarball file. New directory named postgresql-11.3 will be created.

# tar -xvzf postgresql-11.3.tar.gz

# ll
Sample Output
total 26328
drwxr-xr-x 2 nijam nijam 4096 Sep 21 2014 bin
-rwxr-xr-x 1 nijam nijam 347 Nov 7 2017 init-user-db.sh
drwxr-xr-x 6 nijam nijam 4096 May 7 04:59 postgresql-11.3
-rw-r--r-- 1 nijam nijam 25868246 Jul 8 12:04 postgresql-11.3.tar.gz
drwxr-xr-x 2 nijam nijam 4096 May 17 2017 public_html

4. Next step for installation procedure is to configure the downloaded source code by choosing the options according to your needs.

# cd postgresql-11.3
# ls -l
total 764
-rw-r--r-- 1 nijam nijam 730 May 7 04:46 .dir-locals.el
-rw-r--r-- 1 nijam nijam 1622 May 7 04:46 .gitattributes
-rw-r--r-- 1 nijam nijam 504 May 7 04:46 .gitignore
-rw-r--r-- 1 nijam nijam 1192 May 7 04:46 COPYRIGHT
-rw-r--r-- 1 nijam nijam 3848 May 7 04:46 GNUmakefile.in
-rw-r--r-- 1 nijam nijam 284 May 7 04:46 HISTORY
-rw-r--r-- 1 nijam nijam 74257 May 7 04:59 INSTALL
-rw-r--r-- 1 nijam nijam 1682 May 7 04:46 Makefile
-rw-r--r-- 1 nijam nijam 1212 May 7 04:46 README
-rw-r--r-- 1 nijam nijam 522 May 7 04:46 aclocal.m4
drwxr-xr-x 2 nijam nijam 4096 May 7 04:58 config
-rwxr-xr-x 1 nijam nijam 561752 May 7 04:46 configure
-rw-r--r-- 1 nijam nijam 84451 May 7 04:46 configure.in
drwxr-xr-x 56 nijam nijam 4096 May 7 04:58 contrib
drwxr-xr-x 3 nijam nijam 4096 May 7 04:58 doc
drwxr-xr-x 16 nijam nijam 4096 May 7 04:59 src

Note : use ./configure --help to get help about various options.

5. Now create a directory where you want to install postgres files and use prefix option with configure.

# mkdir /opt/11.3/ 
# ./configure --prefix=/opt/11.3/ (or) ./configure --prefix=/opt/11.3/ --without-readline
# ./configure --prefix=/opt/11.3/ --without-readline --without-zlib

Sample output 

6. After configuring, next we will start to build postgreSQL using following make command.

# make (or) # make world (additional modules (contrib), type instead PostgreSQL, contrib, and documentation
# make install (or) # make install-world (if you want contribution extension)

7. Postgresql 11 has been installed in “/opt/11.3/” directory. now create a postgres user and directory to be used as data directory for initializing database cluster. Owner of this data directory should be postgres user and permissions should be 700 and also set path for postgresql binaries for our ease.

# useradd nijam
# passwd nijam
# mkdir -p /data_11.3/
# chown -R nijam.nijam /data_11.3/

8. Now initialize database using the following command as postgres user before using any postgres commands.

# su nijam
$ /opt/11.3/bin/initdb -D /data_11.3/ -U nijam

Where -D is location for this database cluster or we can say it is data directory where we want to initialize database cluster, -U for database superuser name.

For more info and options we can refer initdb –help.

9. After initializing database, start the database cluster or if you need to change port or listen address for server, edit the postgresql.conf file in data directory of database server.

$ /opt/11.3/bin/pg_ctl -D /data_11.3/ -l logfile start

10. After starting database, verify the status of postgres server process by using following commands.

$ ps -ef |grep -i postgres
$ netstat -apn |grep -i 5420

We can see that database cluster is running fine, and startup logs can be found at location specified with -l option while starting database cluster.

11. Now connect to database cluster and create database by using following commands.

$ psql -p 5420
postgres=# \l --- to list all databases in cluster
postgres=# \q --- to quit form postgres console

12. Create the environment variable file then only you can able to access the postgres utility directly on home path without going to bin path.

$ cat /home/nijam/.profile
export PATH=/opt/11.3/bin:$PATH
export PGHOME=/opt/11.3/
export PGDATA=/data_11.3/
export LD_LIBRARY_PATH=/opt/11.3/lib
export PGDATABASE=nijam
export PGUSER=nijam
export PGPORT=5420
export PGLOCALEDIR=/opt/11.3/share/locale
export MANPATH=$MANPATH:/opt/11.3/share/man