How To Install OpenRefine on a web server with Ubuntu 20.04

OpenRefine is intended to be installed locally as a desktop application. Due to the client-server architecture it is also possible to install OpenRefine on a web server to share it with multiple users. This can be useful despite the missing user administration, e.g. temporarily for a workshop or permanently in a protected network.

Security warning

Can I somehow host OpenRefine for others to access ?

OpenRefine has no built-in security for multi-user or multi-tenant scenarios. OpenRefine has a single data model that is not shared, so there is a risk of columnar data operations being overwritten by other users, so care must be taken by users. Having said that, if you are inclined to proceed at your own risk, you can get some security by using a proxy.

https://github.com/OpenRefine/OpenRefine/wiki/FAQ

Installation (without Authentication)

Tested on Ubuntu 20.04 LTS

1. Install Java 11 and Unzip

apt install openjdk-11-jre-headless unzip

1. Configure Java to be behave better on cloud infrastructure

sed -i 's/securerandom.source=file:\/dev\/random/securerandom.source=file:\/dev\/urandom/' /etc/java-11-openjdk/security/java.security

1. Install OpenRefine 3.5.0

mkdir /opt/openrefine
wget https://github.com/OpenRefine/OpenRefine/releases/download/3.5.0/openrefine-linux-3.5.0.tar.gz
tar -xzf openrefine-linux-3.5.0.tar.gz -C /opt/openrefine --strip 1

1. Create a systemd service for OpenRefine (adjust memory if needed with parameter -m 3g)

adduser --system openrefine

echo "[Unit]
Description=OpenRefine
[Service]
User=openrefine
ExecStart=/opt/openrefine/refine -i 0.0.0.0 -m 3G
TimeoutStopSec=3600s
Restart=always
RestartSec=10
[Install]
WantedBy=default.target
" > /etc/systemd/system/openrefine.service

systemctl enable openrefine.service
systemctl start openrefine.service

1. Redirect port 3333 to 80

iptables -t nat -A PREROUTING -p tcp --dport 80 -j REDIRECT --to-port 3333
apt install iptables-persistent

See also