Any good tutorials on this? What are the required specs? I assume the dataset is huge?
https://wiki.openstreetmap.org/wiki/Deploying_your_own_Slipp...
https://switch2osm.org/serving-tiles/
It's small enough that I have a self hosted test tile server in my home office with about 8TB of storage, on a WD hard drive that was shucked out of a USB3 enclosure.
https://medium.com/gowombat/self-hosted-open-source-cheap-al...
"A minimum of 2GB of RAM is required or installation will fail. For a full planet import 64GB of RAM or more are strongly recommended. Do not report out of memory problems if you have less than 64GB RAM. For a full planet install you will need at least 800GB of hard disk space (take into account that the OSM database is growing fast). SSD disks will help considerably to speed up import and queries. Even on a well configured machine the import of a full planet takes at least 2 days. Without SSDs 7-8 days are more realistic." [0]
I'd also like to render tiles in the OSM Bright theme. The default OpenStreetMap just looks kind of unattractive to me.
In the end I just use an old version of Galileo Offline Maps (now Guru Maps) with tiles scraped from Google using MOBAC. [1]. I made MOBAC profiles for every country [2] so I can just re-scrape them when I have a fast Internet connection.
[0] https://nominatim.org/release-docs/latest/admin/Installation...
In my case, it's not hard for me to create an off-laptop VM as a resource... I would not want to attempt self hosting OSM on my own laptop. No matter how powerful it is.
If it's for any sort of for-profit company use (like the ISPs using openstreetmap for network monitoring maps I mentioned in the same thread), fairly powerful used 16-core, 1U and 2U rackmount servers with 128GB of RAM (dell r620, r720, etc) are available used for $430. Then add a couple of 1TB SATA3 SSDs.
I'll probably first look into running it on a dedicated SSD on the spare laptop, and if that goes well, move some subset onto my main laptop (2014 MacBook Pro 15" with 4TB SSD).
For a "slippy map", I use Leaflet with some scraped MOBAC tiles. https://github.com/peterburk/bamb
Search/geocode is missing though, and that's what would really make it more useful. Particularly if I can cut down the geographic range to squeeze it onto my 64GB phone.
I have not used it in many many years but I'd point you towards https://mapserver.org/ At one point I was one of many responsible for rending cache tile sets for all of NZ to 1:500. Sadly the whole system was scraped for a propriety solution mostly due to front-end issues, that funnily enough the propriety solution only multiplied.
https://nominatim.org/2020/10/21/photon-country-extracts.htm...