[OSM-dev] all things to all people
brett at bretth.com
Thu Apr 24 04:59:29 BST 2008
On Wed, Apr 23, 2008 at 7:54 AM, Frederik Ramm <frederik at remote.org> wrote:
> > Since the minute-diff was introduced OSMXAPI has been running, on
> > average, about 6 minutes behind the main database.
> Sigh, this project is too large already for one person to keep the
> overview. Minute diffs you say. Hadn't heard of them until now, thanks
I really should do some more advertising, this has come up a few times now
I've also created tasks for populating a PostGIS "simple" schema which is
useful for maintaining a local replica of OSM data. Latest SVN osmosis has
tasks for merging changesets into it so you could in theory have a PostGIS
database kept within 6 minutes of the main API. It should be very efficient
at serving bounding boxes although only real-world testing will confirm
this. There's still a few TODOs in terms of keeping a database in sync that
only stores a subset of the planet, at the moment I can only apply whole
> > IMHO this is good enough for all purposes, except editing, which
> > obviously needs to act on the live data.
> Have you had a chance to compare the result from a week's worth of
> minute diffs with the planet? Is it considered safe to work on the
> diffs alone, or still recommended to re-synchronize with the planet
> once a week?
I'm not aware of any issues but will be interested if people detect
problems. The reason osmxapi lags by 6 minutes is because I have osmosis
set to only read data that is at least 5 minutes old. This could be dropped
further but there's a risk of long running inserts having an old timestamp
and getting missed. Shouldn't happen most of the time but if a long running
query locks a table it is possible.
> Frederik Ramm ## eMail frederik at remote.org ## N49°00'09" E008°23'33"
> dev mailing list
> dev at openstreetmap.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the dev