[Geocoding] osm2pgsql / Gazetteer import errors

[Brian Quinion]

On Mon, Jan 4, 2010 at 1:29 PM, Frans Hals <fhals7 at googlemail.com> wrote:
> I'm glad to hear that the import should be done faster on another
> machine. It's just the more you wait for completion, the less you
> won't stop it before.

Indeed.

Can I suggest that you might want to try a much smaller file a do a
test import.  For instance I often test new code on the Isle of Man:

http://downloads.cloudmade.com/europe/isle_of_man/isle_of_man.osm.bz2

This should be done completely in about 30 minutes and will confirm
everything is installed and working.

> I started the process on a standard Amazon cloud 32-bit 1-Processor Fedora unit.

Not any immediate problem but I'd guess due to the size of node IDs a
32 bit machine has a limited time left - I'd guess around 2 years.

> So this may be a warning example for everybody else trying it.
> When you talk about another 35 days, will this be the current
> after-import-indexing or the indexing mentioned in step 6 of the
> Readme?
>> 6) Index the database - this will take a VERY long time!
>> cat gazetteer-index.sql | psql gazetteer

Yes, 35 hours to complete all steps (at a guess)

> If the current indexing is done, I will try to clone the machine image
> to a bigger one and
> start with paragraph 6. If you think, the current one will last again
> some days, I'll stop it and restart on a bigger block.

I assume you will do steps 2-5 first? or step 6 will be very short :)

If you are doing a full planet I'd suggest using util.index.php and
commenting out the update placex set indexed = true... lines in
gazetteer-index.sql

> By now I am here:

>From the numbers in there I'd guess you have about another 5-10 hours
to go on the osm2pgsql import.  Just a guess though.

--
 Brian