How Much Server Does MetaKarta Need, and Where Can It Run?
The Deployment Guide is for administrators and platform engineers who install, size, and maintain MetaKarta. It covers the application server, the repository database (PostgreSQL, Microsoft SQL Server, or Oracle), Solr indexing, MIMB metadata harvesting setup, SSL and authentication, backup, upgrade, and migration. When you're done, you can stand up a server on Linux or Windows, connect it to a supported database, and run harvesting agents on premises while the server runs in the cloud.
Note
Some of the features detailed in this document may not apply and/or be available for the particular edition/version you are using.
Get started
Frequently asked questions
What are the minimum hardware requirements for MetaKarta?
The application server's minimum is a 2 GHz quad-core processor, 8 GB of RAM, and 10 GB of disk. That covers the tutorial and small use cases only. For enterprise use, the guide's capacity planning model (200 light and 100 heavy users) recommends at least 64 cores and 300 GB of memory.
Which operating systems does the application server support?
Most popular 64-bit Linux distributions, such as Red Hat, and Microsoft-supported 64-bit Windows, including Windows Server 2012, 2016, 2019, and 2022. Support follows the bundled Java runtime and Tomcat.
On headless Linux, install fontconfig so diagrams render. Windows installs need administrator rights and .NET Framework 3.5 or later.
Does MetaKarta support PostgreSQL, SQL Server, and Oracle?
Yes. The repository database can run on PostgreSQL, Oracle, or Microsoft SQL Server, on premises or in the cloud.
The database server minimum is four cores, 8 GB of RAM, and 20 GB of disk, and setup needs database administrator privileges. For small deployments and proofs of concept, the Windows package bundles PostgreSQL on the same machine as the application server.
How do I migrate the repository database to another database platform?
Use the migration option in the Setup utility. Clean up old model versions, create an empty database on the new platform, stop the application server, run Setup -dm with the new connection details, then restart.
Changing the connection string alone isn't supported. Moves within one platform use that platform's backup and restore.
How do I harvest on-premises sources from a cloud-hosted server?
Install MIMB as a remote harvesting agent on a machine that can reach your sources, then set M_SERVER_URL and M_AGENT_NAME in agent.properties so it connects to the application server. The same setup lets a Linux server use Windows-only bridges, such as those built on COM-based SDKs.
Do I need to install Solr separately?
A new install needs no setup. When the shared metadata repository is empty, MetaKarta starts the bundled Solr indexing server automatically on the same Tomcat server and stores the index under $MM_HOME/data/solr. The guide also covers running an external Solr server and migrating to it from bundled Solr or legacy Lucene index files.
What should I check before upgrading the application server?
Back up the installation and the repository database, and make sure the database has at least 20% free space. Delete unused configurations and old model versions, test every model import against the new bridges, and plan full re-imports. Then follow the 13-step procedure, and don't run anything else against MetaKarta during the automatic migration.
Related documentation
Return to the MetaKarta documentation home. For bridge coverage and known limits, see the MIMB bridge documentation; for release notes and current patches before an upgrade, see Releases & Roadmap.