Installation Deployment

1. CentOS Installation overview

This article introduces the installation process and precautions of IvorySQL on Linux platform. This article mainly demonstrates rpm package installation steps, and source code installation steps of the database under the Centos 7 environment.

1.1. Software and hardware requirements

1.1.1. Introduction to software resources

operating system

Yum source download address

CentOS 7

https://github.com/IvorySQL/IvorySQL/releases

1.1.2. Hardware resource preparation

Configuration parameters

Minimum configuration

Recommended configuration

CPU

4 cores

16 cores

Memory

4GB

64GB

Storage

800MB，mechanical hard disk

Above 5GB，SSD or NvMe

Network

gigabit lan

10 Gigabit network

1.2. Environment and configuration check

Before deploying the IvorySQL database, you need to check the system environment and configuration

1.2.1. View resources

The IvorySQL database supports CentOS 7. X, operating systems. For details, refer to [_software_and_hardware_requirements].

1.2.2. View the operating system

Run the following command to view the operating system information：

cat /etc/redhat-release

1.2.3. View kernel parameters

uname -r

1.2.4. View memory and clear cache

free -g

echo 3 > /proc/sys/vm/drop_caches

1.3. Get installation package

You can install the IvorySql database or RPM package through the source code.

1.3.1. Use source code to build IvorySQL database

1.Get the source code: Run the following command to clone the source code of the IvorySQL database to your build machine:

git clone https://github.com/IvorySQL/IvorySQL.git

Note: To clone code, you need to install and configure Git first. For details, refer to Git doc.

2.Install dependent packages: To compile IvorySQL from source code, you must ensure that prerequisite packages are available on the system. Execute the following command to install related packages：

sudo yum install -y bison-devel readline-devel zlib-devel openssl-devel wget
sudo yum groupinstall -y 'Development Tools'

Note: "Development Tools" includes gcc, make, flex, bison.

3.Self-compilation and installation: The source code obtained previously is in the folder IvorySQL. Next, we will enter this folder for operation.

Configuration: Root users execute the following commands to configure:

    git checkout tags/Ivory_REL_1_17
    ./configure --prefix=/usr/local/ivorysql/ivorysql-1.17

Note: If the --prefix option is not provided, the default installation path will be /usr/local/ivorysql.

Note: We should remember the specified directory, because the system cannot find out where the compiled and installed programs are. More configure parameters can be viewed through "./configure — help". You can also view the PostgreSQL manual

Compile and install: After the configuration is completed, execute make to compile:

make

Perform a regression test before installing the newly compiled service (the following commands can be used):

make check
make all-check-world

Installation:

make install

1.3.2. Install the IvorySql database using the RPM package

Run the following command to download the IvorySQL installation package.

wget https://github.com/IvorySQL/IvorySQL/releases/download/Ivory_REL_1_17/IvorySQL-1.17-fde5539-20250326.x86_64.rpm

Note: The installation package in the example may not be the latest version. It is recommended that you download the latest installation package.

2.Run the following command to install IvorySQL.

yum install -y libicu libxslt python3                   --Install dependencies first
rpm -ivh IvorySQL-1.17-fde5539-20250326.x86_64.rpm

1.4. Initialize database service

1.4.1. Initialize database

1.Create an operating system user: under the user root session, create a new user ivorysql:

/usr/sbin/groupadd ivorysql
/usr/sbin/useradd -g ivorysql ivorysql -c "IvorySQL1.17"
passwd ivorysql

2.Environment variable: switch to user ivorysql, modify the file "/home/ivorysql/.bash_profile", and configure the environment variable：

umask 022
export LD_LIBRARY_PATH=/opt/IvorySQL-1.17/lib:$LD_LIBRARY_PATH
export PATH=/opt/IvorySQL-1.17/bin:$PATH
export PGDATA=/home/ivorysql/data

Make the environment variable effective in the current ivorysql user session:

source .bash_profile

You can also log in again or open a new user ivorysql session.

3.Set the firewall: If the firewall is enabled, port 5333 needs to be opened:

firewall-cmd --zone=public --add-port=5333/tcp --permanent
firewall-cmd --reload

Note: The default port is 5333. If the port is not opened, the external client will fail to connect via IP.

4.Initialization: Under user ivorysql, simply execute initdb to complete initialization:

initdb

Note: The initdb operation is the same as PostgreSQL. It can be initialized according to the habits of PG.

5.Start database: use pg_ Ctl starts the database service:

pg_ctl start

View the status and start successfully:

pg_ctl status

1.5. Configure service

1.Client authentication: modify /home/ivorysql/data/pg_hba.conf, add the following:

host    all             all             0.0.0.0/0               trust

Note: This is trust, which means that you can log in without password.

Execute the following command to load the configuration:

pg_ctl reload

2.Basic parameters

Connect to the database through psql:

psql

Modify listening address

alter system set listen_addresses = '*';

Note: The default is listening at 127.0.0.1. The service cannot be connected outside the host.

3.Guard service

Create a service file:

touch /usr/lib/systemd/system/ivorysql.service

The editing contents are as follows:

[Unit]
Description=IvorySQL 1.17 database server
Documentation=https://www.ivorysql.org
Requires=network.target local-fs.target
After=network.target local-fs.target

[Service]
Type=forking

User=ivorysql
Group=ivorysql

Environment=PGDATA=/home/ivorysql/data

OOMScoreAdjust=-1000

ExecStart=/opt/IvorySQL-1.17/bin/pg_ctl start -D ${PGDATA}
ExecStop=/opt/IvorySQL-1.17/bin/pg_ctl stop -D ${PGDATA}
ExecReload=/opt/IvorySQL-1.17/bin/pg_ctl reload -D ${PGDATA}

TimeoutSec=0

[Install]
WantedBy=multi-user.target

Note: There are many ways to write a service. Be careful when using it in a production environment. Please repeat the test several times.

Stop pg_ The database service started by ctl enables the systemd service and starts:

systemctl enable --now ivorysql.service

IvorSQL database service operation command:

systemctl start ivorysql.service
systemctl stop ivorysql.service
systemctl restart ivorysql.service
systemctl status ivorysql.service
systemctl reload ivorysql.service

1.6. Uninstall the IvorySQL database

1.6.1. Compile Uninstall

1.Backup data: we can protect the directory. It is better to stop the database service and make a backup.

systemctl stop ivorysql.service

2.Compile and uninstall: switch the root session to the source directory and execute the following commands respectively:

make uninstall
make clean

3.Delete residual directories and files:

systemctl disable ivorysql.servicemake                  --disable Service
mv /usr/lib/systemd/system/ivorysql.service /tmp/       --the service file can be moved                                                             to/tmp or deleted
rm -fr /opt/IvorySQL-1.17                             --remove residual installation                                                             directory

Note: There are also user ivorysql and corresponding environment variables, which can be cleaned according to the situation. Please make sure to make a backup before processing. There are also installed dependent packages, which can be uninstalled according to the situation.

2. Ubuntu Installation overview

2.1. Software and hardware requirements

2.1.1. Introduction to software resources