[Tilesathome] Why sleep when you could render?

[Matthias Julius]

Dirk-Lüder Kreie <osm-list at deelkar.net> writes:

> Dermot McNally schrieb:
>> On 04/11/2007, OJW <streetmap at blibbleblobble.co.uk> wrote:
>> 
>>> AFAIK, it was considered wasteful to download OSM data and process it, if you
>>> already know that you wouldn't be able to upload the result.
>> 
>> Generally, though, you expect to be able to upload the result (which
>> is why it currently sleeps rather than bombing out). It's just a
>> question of what you do while you wait to try again.
>
> So why would you want to render a new tile before uploading the current
> one? this would result in an even bigger upload backlog as it is already
> and basically generate stale map tiles.

This would only help in the (unlikely) case when the upload queue
emties while the clients render new tiles. With all the TIGER and AND
data rendering sometimes takes a long time.

>
> The bottleneck of t at h is the server, not the rendering clients, but we
> have been promised a "new" server, so that should change to some extent.

What happended to the server that it now only processes half as many
tilesets than it used to some time ago? Is it just that tilesets get
bigger and more complex?

Matthias