Shared storage for HA docker setup question

Hi there,

First of all, my apologies if this question has been asked before, I searched and I couldn’t find anything that matched it. I am new to GIS and to OpenRoute, coming more from an infrastructure perspective as I have been asked to provide an internal service for our products and I was wondering if I was on the right track.

We’re deep into AWS, so I am going to implement this on top of Fargate as I would like to avoid running instances (patching, compliance, etc) and offloading all that is great. I have been able to merge multiple region OSM files, point my local OSM app (running inside a docker container) to it and generate graphs for the profile my teams want to use (walking). Now, my question is:

the approach I am thinking about would be to get the graphs/* and elevation_cache/* generated locally, push those a shared storage, say something like EFS and make that available to two or more Fargate tasks (you could see that as multiple docker containers accessing the same shared storage), so that we have some high availability. If I understand correctly, once the graphs are generated (in my local docker setup, the application would not need to write to any of these?). Is this a known and acceptable use case? or is there documentation on how I could provide a solution that gives me a bit more redundancy than a single instance? or should each docker (fargate) task should have their own storage?

I hope my question makes sense. Looking forward to any pointers that would help me get this implemented.

And thank you in advance.


Hi @bruno ,
yes this makes perfectly sense and I believe this should work. You would need to configure the graphs folder in ORS and correctly link it to your container. If ORS finds an existing graph in the graphs folder, it just loads it without building a new one. The elevation cache is only needed to build the graphs but not for routing. You can find the backend-documentation here: Getting started | openrouteservice backend documentation.

Best regards

1 Like