[Tilesathome] Cutover to new server? When?

[Kai Krueger]

spaetz wrote:
> On Sat, Jul 12, 2008 at 09:31:24AM +0100, Kai Krueger wrote:
> 
>> Is the tile conversion still ongoing? Do you have an estimate of how 
>> long this will still be running? If it is still going to be for a little 
>> while, it might be a good idea to add a quick check in 
>> LegacyTileset.convert if there is already a newer tileset rendered 
>> instead of simply replacing it.
> 
> I have aborted the tile conversion run. It takes 0.4 seconds per tileset when I ignore the blank database and takes 120 seconds when i use the database. Even with 0.4 secondsit would take long. I have implemented a fallback on the old tiles now and once the server is the main server, I will re-request all missing tilesets with very low priority. Do you think this is a good approach? Comparing time-stamps when converting would be an alternative, but 120 seconds per tileset literally takes forever and I fix rather bugs than invest more effort into handling legacy tilesets.

Hmm, that would indeed take quite a while. If my calculations are 
correct, then for the 0.4 sec case it would still be about 70 days for 
all of z12 tiles including the pacific to be converted.

Presumably that will also mean that re-rendering all of the tiles under 
the new server will take at least as long. So there will be an overlap 
of old and new system for quite a while. I think that is perfectly fine, 
but it does mean that the legacy system has to work well. I.e. the 
blankness lookup must be implemented correctly.

I see three possible options for the legacy blankness lookup.
1) Just like the LegacyTileset, call the old blankness db. From your 
numbers above, this seems quite inefficient though, although I could 
imagine this could be optimised quite a bit, e.g. not spanning a new 
database connection on each new lookup
2) Implement a new blankness database in some form. Not sure how this 
would be populated though.
3) Use the ocean tiles db as a fallback of the fallback. From my 
previous tests, this was actually quite efficient and easy to code at 
least for z12 and above.

What is your preferred method? If you want I could send a patch for at 
least solution 3 and possibly 1, if you think that would be useful.

Kai



> 
> spaetz
> 
> _______________________________________________
> Tilesathome mailing list
> Tilesathome at openstreetmap.org
> http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/tilesathome