On this page
- Technical requirements
- Instructions for new installations
- Instructions for upgrades
- Advanced installation options
The following table lists various ways to install and use Archivematica. Further down the page, you will find more detailed information about new installations and advanced installation options.
If you need assistance or clarification regarding the installation instructions, the Archivematica user forum is a good place ask questions.
|New installations for production deployment|
|Ubuntu 16.04 64-bit Server Edition||Automated install using Ansible > Ubuntu 16.04 (Xenial)||These instructions will not work if you are using a Windows machine as the host environment. For Windows installations, see the manual install instructions for production.|
|Ubuntu 18.04 64-bit Server Edition||Automated install using Ansible > Ubuntu 18.04 (Bionic)||These instructions will not work if you are using a Windows machine as the host environment. For Windows installations, see the manual install instructions for production.|
|CentOS 7 64-bit||Installing Archivematica on CentOS/Red Hat||Archivematica versions 1.5.1 and higher support installation on CentOS/Red Hat.|
|Windows OS||Not supported|
|Test-driving the latest release|
|Archivematica sandbox||Using the sandbox||The sandbox will automatically reset daily. Any packages created will not be permanently saved. Additionally, there may be more than one demo user logged in at the same time, so you may see changes made by others while using the software.|
|Local virtual machine using Vagrant and VirtualBox||Installing on a virtual machine using Vagrant||This virtual machine is not intended to be used in production.|
|Docker and Linux||Installing Archivematica Using Docker Compose on Linux||Docker will provide instructions on how to use it as a non-root user. This may not be desirable for all.|
|Docker and Mac||Installing Archivematica Using Docker Compose on Mac||Installation of Archivematica on machines running macOS using Docker is possible but still in development and may require some extra steps.|
|Spreading Archivematica’s processing load across several machines|
|Base on your customized setup and network configuration||Scaling Archivematica||When installing Archivematica on multiple machines, the various Archivematica processes must be able to reach each other on the relevant ports. Your firewall configuration must allow for this.|
Archivematica 1.12.2 installation instructions are provided here for the following operating systems:
- Ubuntu 16.04 64-bit Server Edition
- Ubuntu 18.04 64-bit Server Edition
- CentOS 7 64-bit
Other Linux distributions should work, but will require customization of these installation instructions.
Support for macOS is possibly in theory, but is not being tested, and would require more significant deviation from these instructions.
Archivematica is unlikely to ever run directly in a Windows environment. Consider the use of a virtualization platform to run Linux VMs.
Archivematica has a long list of software it depends on. All of these dependencies are installed when following the instructions below.
Note that it is possible to install some of the components on separate machines in order to improve performance, such as:
- Elasticsearch (optional as of Archivematica 1.7, see below)
Using additional machines will require additional configuration. For more information, see Advanced.
Archivematica 1.12 has been tested with MySQL 5.5, including the Percona and MariaDB alternatives. Archivematica uses MySQL 5.7 on Ubuntu 16.04 and Ubuntu 18.04.
Some of the tools run by Archivematica require Java to be installed (primarily Elasticsearch and fits). On Ubuntu 18.04, Open JDK 8 is used, but Open JDK 11 is the default. On Ubuntu 16.04, Open JDK 8 is the default. It is possible to use Oracle Java 7 or 8 instead.
The remaining dependencies should be kept at the versions installed by Archivematica.
As of Archivematica 1.7, installing Elasticsearch is optional. Elasticsearch powers the indexes that are used for searching in the Backlog, Appraisal, and/or Archival Storage. Installing Archivematica without Elasticsearch results in reduced consumption of compute resources and lower operational complexity. Disabling Elasticsearch means that the Backlog, Appraisal, and/or Archival Storage tabs will not appear in the user interface and their functionality will not be available.
By setting the
configuration attribute, administrators can define how many things Elasticsearch
is indexing, if any. This can impact searching across several different
archivematica_src_search_enabled configuration attribute values:
transfers: Only transfers are indexed. Search is enabled on the Backlog and Appraisal tabs, but not the Archival Storage tab.
aips: Only AIPs are indexed. Search is enabled on the Archival Storage tab, but not the Backlog or Appraisal tabs.
true: Both AIPs and transfers are indexed. Search works on the Backlog, Appraisal, and Archival Storage tabs.
false: Indexless mode. Neither AIPs nor transfers are indexed. The Backlog, Appraisal, and Archival Storage tabs will be non-functional.
When Elasticsearch is used, Archivematica 1.12.2 requires version 6.x (tested with 6.5.4).
For more information on disabling Elasticsearch, please see the README for Archivematica’s ansible role,
Archivematica is capable of running on almost any hardware supported by Linux; however, processing large collections will require better hardware.
Minimum hardware requirements¶
For small-scale functionality testing using small collections (transfers with 100 files or less, total file size 1 GB or smaller), we recommend the following minimum hardware requirements:
- Processor: 2 CPU cores
- Memory: 2GB+
- Disk space (processing): 7GB plus two to three times the disk space required for the collection being processed (e.g., 3GB to process a 1GB transfer)
Recommended minimum production requirements¶
For production processing, the hardware requirements depend almost entirely on the size and number of files being processed. These recommendations should be considered the minimum for a viable production system:
- Processor: 2 CPU cores
- Memory: 4GB
- Disk space (processing): 200GB
More commonly, we deploy the following:
- Processor: 8 CPU cores
- Memory: 16GB
For processing disk space, we recommend allocating 20GB plus four times the disk space required for the largest transfer that you expect to process. If your largest transfer is 50GB, allocation at least 220GBs of disk space.
The amount of transfer source disk space needed is subjective, and depends on individual workflows.
The amount of storage disk space needed will depend on how much material you intend to store, as well as how it is stored (compressed or uncompressed).
These requirements may not be suitable for certain types of material - for example, audio-visual material requires more processing power than images or documents.
Instructions for new installations¶
Archivematica can be installed using packages or Ansible scripts in either CentOS/Red Hat or Ubuntu environments. It can also be installed using Docker. At this time, installation instructions are provided for officially tested and supported installation environments:
- Automated install on Ubuntu (16.04 and 18.04) using Ansible.
- Manual install of OS packages on CentOS/Red Hat
Note that manual install of OS packages on Ubuntu (16.04 and 18.04) is documented but not officially supported.
Installing Archivematica using Docker is not officially supported for production deployments. However, it is the preferred development environment for those who work on Archivematica’s code.
Instructions for upgrades¶
If you are upgrading from a previous version of Archivematica, please see the upgrading instructions.
Advanced installation options¶
There are many ways to install Archivematica, depending on the needs of the individual user. We have documented some common advanced installation setups.
- Installing for development
- Configure Archivematica with SSL
- Configure Archivematica with task output capturing disabled
- Scaling Archivematica