[OSM-dev] Taginfo (tagstats) data processing error

[Jochen Topf]

On Tue, Apr 16, 2013 at 07:53:51AM -0600, Martijn van Exel wrote:
> I am trying to get Taginfo up and running for the US. 
> I had my share of trouble getting to this point, especially getting this to run on Ubuntu Lucid (don't try) but now I'm on Precise, and the update-all.sh script does not complete. Here's the last few lines:
> 
> processing nodes took 1756 seconds (about 29 minutes)
> string_store: chunk_size=10MB chunks=4 memory=40MB bytes_in_last=9888kB
> VmPeak:   241916 kB
> VmSize:   241916 kB
> gridcells_all: 0
> sum of location image sizes: 416416
> dumping images took 0 seconds (about 0 minutes)
> string_store: chunk_size=10MB chunks=4 memory=40MB bytes_in_last=9888kB
> VmPeak:   242960 kB
> VmSize:   242960 kB
> tagstats: /usr/include/google/sparsetable:1346: google::sparsetable<T, GROUP_SIZ
> E, Alloc>::const_reference google::sparsetable<T, GROUP_SIZE, Alloc>::get(google
> ::sparsetable<T, GROUP_SIZE, Alloc>::size_type) const [with T = short unsigned i
> nt; short unsigned int GROUP_SIZE = 48u; Alloc = google::libc_allocator_with_rea
> lloc<short unsigned int>; google::sparsetable<T, GROUP_SIZE, Alloc>::const_refer
> ence = const short unsigned int&; google::sparsetable<T, GROUP_SIZE, Alloc>::siz
> e_type = long unsigned int]: Assertion `i < settings.table_size' failed.
> Aborted (core dumped)
> 
> This is on a m3.2xlarge EC2 instance.
> Any idea what is the matter here?

Can you provide a stack trace?

With something like this:
    ulimit -c unlimited
    run the update script
    gdb core tagstats
    > bt

Jochen
-- 
Jochen Topf  jochen at remote.org  http://www.remote.org/jochen/  +49-721-388298