3 This page contains generic installation instructions for Nominatim and its
4 prerequisites. There are also step-by-step instructions available for
5 the following operating systems:
7 * [Ubuntu 22.04](../appendix/Install-on-Ubuntu-22.md)
8 * [Ubuntu 20.04](../appendix/Install-on-Ubuntu-20.md)
10 These OS-specific instructions can also be found in executable form
11 in the `vagrant/` directory.
13 Users have created instructions for other frameworks. We haven't tested those
14 and can't offer support.
16 * [Docker](https://github.com/mediagis/nominatim-docker)
17 * [Docker on Kubernetes](https://github.com/peter-evans/nominatim-k8s)
18 * [Kubernetes with Helm](https://github.com/robjuz/helm-charts/blob/master/charts/nominatim/README.md)
19 * [Ansible](https://github.com/synthesio/infra-ansible-nominatim)
26 For larger installations you **must have** PostgreSQL 11+ and PostGIS 3+
27 otherwise import and queries will be slow to the point of being unusable.
28 Query performance has marked improvements with PostgreSQL 13+ and PostGIS 3.2+.
32 * [cmake](https://cmake.org/)
33 * [expat](https://libexpat.github.io/)
34 * [proj](https://proj.org/)
35 * [bzip2](http://www.bzip.org/)
36 * [zlib](https://www.zlib.net/)
37 * [ICU](http://site.icu-project.org/)
38 * [nlohmann/json](https://json.nlohmann.me/)
39 * [Boost libraries](https://www.boost.org/), including system and filesystem
40 * PostgreSQL client libraries
41 * a recent C++ compiler (gcc 5+ or Clang 3.8+)
43 For running Nominatim:
45 * [PostgreSQL](https://www.postgresql.org) (9.6+ will work, 11+ strongly recommended)
46 * [PostGIS](https://postgis.net) (2.2+ will work, 3.0+ strongly recommended)
47 * [Python 3](https://www.python.org/) (3.7+)
48 * [Psycopg2](https://www.psycopg.org) (2.7+)
49 * [Python Dotenv](https://github.com/theskumar/python-dotenv)
50 * [psutil](https://github.com/giampaolo/psutil)
51 * [Jinja2](https://palletsprojects.com/p/jinja/)
52 * [SQLAlchemy](https://www.sqlalchemy.org/) (1.4.31+ with greenlet support)
53 * [asyncpg](https://magicstack.github.io/asyncpg) (0.8+)
54 * [PyICU](https://pypi.org/project/PyICU/)
55 * [PyYaml](https://pyyaml.org/) (5.1+)
56 * [datrie](https://github.com/pytries/datrie)
57 * [PHP](https://php.net) (7.3+)
59 * PHP-intl (bundled with PHP)
61 For running continuous updates:
63 * [pyosmium](https://osmcode.org/pyosmium/)
65 For running the experimental Python frontend:
67 * one of the following web frameworks:
68 * [falcon](https://falconframework.org/) (3.0+)
69 * [starlette](https://www.starlette.io/)
70 * [uvicorn](https://www.uvicorn.org/)
72 For dependencies for running tests and building documentation, see
73 the [Development section](../develop/Development-Environment.md).
77 A minimum of 2GB of RAM is required or installation will fail. For a full
78 planet import 128GB of RAM or more are strongly recommended. Do not report
79 out of memory problems if you have less than 64GB RAM.
81 For a full planet install you will need at least 1TB of hard disk space.
82 Take into account that the OSM database is growing fast.
83 Fast disks are essential. Using NVME disks is recommended.
85 Even on a well configured machine the import of a full planet takes
86 around 2 days. On traditional spinning disks, 7-8 days are more realistic.
88 ## Tuning the PostgreSQL database
90 You might want to tune your PostgreSQL installation so that the later steps
91 make best use of your hardware. You should tune the following parameters in
92 your `postgresql.conf` file.
95 maintenance_work_mem = (10GB)
96 autovacuum_work_mem = 2GB
98 effective_cache_size = (24GB)
99 synchronous_commit = off
101 checkpoint_timeout = 10min
102 checkpoint_completion_target = 0.9
104 The numbers in brackets behind some parameters seem to work fine for
105 64GB RAM machine. Adjust to your setup. A higher number for `max_wal_size`
106 means that PostgreSQL needs to run checkpoints less often but it does require
107 the additional space on your disk.
109 Autovacuum must not be switched off because it ensures that the
110 tables are frequently analysed. If your machine has very little memory,
111 you might consider setting:
113 autovacuum_max_workers = 1
115 and even reduce `autovacuum_work_mem` further. This will reduce the amount
116 of memory that autovacuum takes away from the import process.
118 For the initial import, you should also set:
121 full_page_writes = off
123 Don't forget to re-enable them after the initial import or you risk database
127 ## Downloading and building Nominatim
129 ### Downloading the latest release
131 You can download the [latest release from nominatim.org](https://nominatim.org/downloads/).
132 The release contains all necessary files. Just unpack it.
134 ### Downloading the latest development version
136 If you want to install latest development version from github, make sure to
137 also check out the osm2pgsql subproject:
140 git clone --recursive https://github.com/openstreetmap/Nominatim.git
143 The development version does not include the country grid. Download it separately:
146 wget -O Nominatim/data/country_osm_grid.sql.gz https://nominatim.org/data/country_grid.sql.gz
149 ### Building Nominatim
151 The code must be built in a separate directory. Create the directory and
159 Nominatim uses cmake and make for building. Assuming that you have created the
160 build at the same level as the Nominatim source directory run:
169 The default installation no longer compiles the PostgreSQL module that
170 is needed for the legacy tokenizer from older Nominatim versions. If you
171 are upgrading an older database or want to run the
172 [legacy tokenizer](../customize/Tokenizers.md#legacy-tokenizer) for
173 some other reason, you need to enable the PostgreSQL module via
174 cmake: `cmake -DBUILD_MODULE=on ../Nominatim`. To compile the module
175 you need to have the server development headers for PostgreSQL installed.
176 On Ubuntu/Debian run: `sudo apt install postgresql-server-dev-<postgresql version>`
179 Nominatim installs itself into `/usr/local` per default. To choose a different
180 installation directory add `-DCMAKE_INSTALL_PREFIX=<install root>` to the
181 cmake command. Make sure that the `bin` directory is available in your path
185 export PATH=<install root>/bin:$PATH
188 Now continue with [importing the database](Import.md).