Skip to main content

Installation on a Server

How to Install ABRA Flexi on a Server

Written by Hana Vršanská

Server Installation

If you want to run ABRA Flexi on your own server within a network, some configuration steps are required.

For a network installation of Flexi on Windows, the steps listed below are not necessary, as everything is handled automatically by the installation program. In the case of a local installation or if a third-party firewall is installed, you will need to make configuration changes to ensure everything works correctly.

Enabling Remote Access on the Server

If you did not select the Network Installation and Server type during the Flexi installation, you need to enable remote access (the installation prompt is not supported on Linux with RPM packages — see below). Log in to the Flexi accounting system as a company administrator and go to Tools, then Server Settings. Select Provide Flexi service on the network.

Enabling the Port on the Firewall

For the database server to communicate correctly over the network, you need to allow access to port 5434 (TCP protocol) on the firewall you are using.

Enabling Remote Access on Linux with RPM Packages

Linux distributions with RPM packages (SUSE, Fedora, CentOS, RHEL, …) do not support interactive prompts during installation. You therefore need to change the installation type manually:

  1. Edit the file /etc/default/flexibee and change:
    FLEXIBEE_CFG=local change to FLEXIBEE_CFG=server

  2. Restart the Flexi service: /etc/init.d/flexibee restart

Everything should now be working.

Creating the Initial (Administrator) Account

After installing Flexi, the system is in first-launch mode. No account exists in the system, and when the
first client connects, they will be prompted to create an account. Once the account is created, the system will be secured.

Server Configuration for More Than 5 Users

If your installation has more than 5 users, you need to configure the database system for optimal performance.

PostgreSQL Configuration

  • Windows:

    • If upgrading from a version where we replaced postgresql.conf, restore the default value there

    • In the main postgresql.conf, simply add a reference to the subfolder containing additional configuration files (include_dir = 'conf.d')

    • In the conf.d subfolder, keep your own .conf files. File names follow the format: 0xx_flexi_popis.conf (e.g. 051_flexi_main.conf)

    • Postgres loads files in alphabetical order by name, with later-loaded files taking precedence. Users should therefore create a file yxx_cokoliv.conf where y > 0 (e.g.: 101_custom.conf) if they wish to override settings. This allows them to override our configuration. Such a file will not be modified during reinstallation.

    • Users should not modify the main postgresql.conf or our .conf files, as these may be overwritten during reinstallation.

  • Addresses the same points described above, but for Linux systems:

    • in the PostgreSQL configuration, we set the jit = off option (generally intended as a performance-improving setting, but in our case it actually slows down queries)

    • the installation no longer modifies the default database configuration file postgresql.conf (the only thing it may add is the include_dir = 'conf.d' option)

    • all important settings are then placed in the conf.d subdirectory in separate files (e.g. 53-jit-off.conf)

Did this answer your question?