Overpass API developpement

[Roland Olbricht <>]

> i'm trying to prepare a summary of what root is and is not needed for, 
> assuming that an overpass user is
> created and most of the work can be done there.

That is indeed correct. There should be a user that owns the directory 
where the database is stored. The same user should run the update 
scripts and start the dispatcher.

The user for the web server can and will usually be distinct. The user 
for the web server needs access to that directory and the right to 
execute the interpreter binary.

You may need root permissions once to configure in the web server the 
CGI directory. On the other hand, on the Rambler instance, the web 
server has been pre-configured, and I never have needed in all the years 
root access.

> the one issue i'm unsure of is setting up access to the logs for minutely
> updates. does anyone have any experience or good advice for us on this
> point?

I'm not sure what you mean here. The logging file is written to the 
database directory. Depeding on the way of updating, there might or 
might not be a second directory containing the minute diffs and a log of 
the download times.

Both directories should be, as mentioned before, under control of the 
single normal user that runs the Overpass service.

Best regards,
Roland