Can Graphhopper server cache its startup computation?

Open Source Routing Engine
1

Greetings!

Just started using Graphhopper; overall, I really like what I see.

One question I have about the Java server itself: when I boot up the server, it takes several minutes and about 1.6-ish gigs of RAM to precompute the 500-some-odd-MB of PBF data I feed to it for the area around me. It looks like this precomputation happens every time the server starts up.

Is there a way to cache that precomputation so that the server can start up faster if the underlying PBF data hasn’t changed since the last run? I’m experimenting with embedding Graphhopper on a Raspberry Pi in a car and the multi-minute startup time is a bit of a deal-breaker for that use case.

2

GraphHopper creates a folder typically named either graph-cache or something-gh where it stores all the precomputed data. So when the server starts a second time the pbf is not processed again. Are you using GraphHopper with Docker or something? Because what you are describing should only ever happen if this folder either is not created for some reason, cannot be found, or is deleted somehow.

3

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.