[Tile-serving] [openstreetmap/osm2pgsql] Full Planet over inflated db size (#579)

[MyPC8MyBrain]

Hi everyone,
Thank you for supporting osm2pgsql, 

I’ve been trying the past few weeks to complete full planetosm import with very mixed and confusing results,
by any means i am not an expert and fairly new to working with pgsql and osm2pgsql but not to with osm data in general,
I have completed one import which wasn’t too bad after 29h, I did get allot of “Standard exception processing relation id=…: TopologyException: side location conflict at …”
The import completed, db was around 20GB (very odd size) slim was 32gb + hchange; through accessing the data didn’t work for our application need; default projection wasn’t right for us,
I dropped the db, i changed a single flag in my string from -E 4326 to -l
This is where everything went wrong,
First I wasn’t sure about the errors I got the first run; I verified md5 all was intact; I decided to grab the latest pbf,
With the latest  pbf posted I didn’t get as many errors as i got the first time; but I also changed the projection to latlong,
I did however kept getting these two repeatedly 
“Standard exception processing relation id=2563291: TopologyException: side location conflict at 18.659793799999999 44.570205299999998
and
Standard exception processing relation id=3151399: TopologyException: side location conflict at -74.521022599999995 48.374912999999999”
But that’s it, import kept going through every time; what didn’t work out is the db size!
All I changed was -E to -l and my db size went through the roof, but I mean through the roof blowing up 405GB and failing in my one before last run,

This may be important:
Initially I noticed I wasn’t getting parallel processing, I found the right build for my OS and was finally getting 8 pthread as expected,
The first run where helper processes are used seems to be going well, all 8 cores spike up for the duration of the first batch run,
The next batch though doesn’t! strangely reduces itself to use 6-1 core all the way to the end while not utilizing more than half of the memory or processors allocated,
I also get this throughout “WARNING:  there is no transaction in progress”

Any idea why I’m getting over inflated db size with -l Flag?
 Using 1.0.132 build with PG 9.5

TIA
Chris




---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/openstreetmap/osm2pgsql/issues/579
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tile-serving/attachments/20160504/a12b3fe5/attachment.html>