[Tilesathome] TRAPI problem? - maplint crashes

Jeremy Adams milenko at king-nerd.com
Thu Oct 29 18:45:26 GMT 2009 

On Thu, Oct 29, 2009 at 2:36 PM, Jeremy Adams <milenko at king-nerd.com> wrote:

> On Thu, Oct 29, 2009 at 1:17 PM, <a_a at gmx.de> wrote:
>
>> Hi Andre,
>>
>> Andre Hinrichs wrote:
>> > Hi List!
>> >
>> > Viewing the job list I saw some tiles which were permenantely in the
>> > queue while they seem to be very simple with few content.
>> >
>> > So, I just tried to run the client on one of that tiles directly and
>> > found that maplint crashed for that tile with an error. Running it again
>> > for maplint only shows the following behaviour:
>> >
>> >> $ ./tilesGen.pl --Debug=1 xy 2348 2125 12 maplint
>> >> - Using working directory /tmp/client-no-auto-1
>> >> - Using process log file tah-process.log
>> >> - Pngcrush version 1.6.15
>> >> - pngnq version 0.5
>> >> - Java version 1.6 is available
>> >> - ImageMagick version 6.5.2 (lowzoom enabled)
>> >> - rendering using or/p
>> >> - Using http://server.tah.openstreetmap.org/Request/take/ for Requests
>> >> This is version 14692 (Ulm) of tilesgen running on linux, ID: 3734
>> >> - rasterizing using SVG::Rasterize::Engine::Inkscape
>> >> * Take care to manually backup your inkscape user preferences
>> >>   if you have knowingly changed them.
>> >>   Some tilesets will cause inkscape to clobber that file!
>> >> [#1   0% ] Tileset (12,2348,2125) around -6.80,26.41
>> >> [#1   0% Download] Downloading map data for maplint from TRAPI...
>> >> Downloading:
>> http://api1.osm.absolight.net/api/0.6/map?bbox=26.323242,-6.854006,26.499023,-6.737060
>> >> [#1   0% Download] Download of data complete
>> >> [#1   0% Download] Checking for UTF-8 errors...
>> >> [#1   0% Download] Download in 1 sec
>> >> [#1   0% maplint] Running maplint...
>> >> [#1   0% maplint] Creating tags from maplint...
>> >> [#1   0% maplint-z12] Generating SVG file...
>> >> [#1   0% maplint-z12] Transforming zoom level 12 with or/p...
>> >> Modification of non-creatable array value attempted, subscript -1 at
>> /home/tilesathome/client-no-auto-1/osmarender/orp/orp-drawing.pm line
>> 399.
>> >> [#1   0% maplint-z12] orp produced an error, aborting render.
>> >> [#1   0% maplint-z12] xml2svg failed
>> >> removing job dir /tmp/client-no-auto-1/12_2348_2125_Luha2
>> >
>> >
>> >
>> > Examing the osm data file I could not find any errors, but running the
>> > same tile with API as data server just worked fine. So, I assume that
>> > this is a TRAPI problem but it could also be a problem of maplint. I
>> > created a list of tiles which hung in my process queue. But also saw
>> > that there are many more as active jobs for other users... Here is the
>> > list, interestingly all in a specific area:
>> >
>> > 2332 2121
>> > 2332 2124
>> > 2333 2115
>> > 2333 2116
>> > 2333 2119
>> > 2334 2113
>> > 2334 2117
>> > 2334 2120
>> > 2334 2125
>> > 2334 2126
>> > 2335 2117
>> > 2335 2125
>> > 2335 2126
>> > 2336 2116
>> > 2336 2126
>> > 2338 2118
>> > 2339 2122
>> > 2340 2116
>> > 2341 2116
>> > 2341 2125
>> > 2342 2123
>> > 2343 2123
>> > 2343 2125
>> > 2347 2122
>> > 2347 2126
>> > 2348 2123
>> > 2348 2126
>> > 2349 2124
>> > 2351 2123
>> >
>> > The jobs of the other users will stay in the queue unless they get to a
>> > renderer which does not load from TRAPI or the problem is fixed. Second
>> > case would be better...
>> >
>> > Can someone please have a look at it?
>> It seems that mat's TRAPI db is corrupted.
>> If you compare the output of
>>
>> http://api1.osm.absolight.net/api/0.6/map?way=33598566
>>
>>
>> <osm version="0.5" generator="Trapi 0.3 mat">
>> <!-- some nodes from z0 0 0 -->
>> <!-- some nodes from z11 1174 1062 -->
>> <node id="435333843" lat="-6.7741543" lon="26.48248"/>
>> <node id="435662873" lat="-6.7578006" lon="26.4925016"/>
>> <node id="435950080" lat="-6.8008796" lon="26.4632973"/>
>> <node id="435967399" lat="-6.8071957" lon="26.4625141"/>
>> <!-- some ways from z11 1174 1062 -->
>> <way id="33598566">
>> <nd ref="436000366"/>
>> <nd ref="435967205"/>
>> <nd ref="435951365"/>
>> <nd ref="436060715"/>
>> <nd ref="435738062"/>
>> <nd ref="435967399"/>
>> <nd ref="435950080"/>
>> <nd ref="435359028"/>
>> <nd ref="435589020"/>
>> <nd ref="435333843"/>
>> <nd ref="435843476"/>
>> <nd ref="435662873"/>
>> <nd ref="435683994"/>
>> <tag k="highway" v="road"/>
>> <tag k="ID" v="28558"/>
>> </way>
>> </osm>
>>
>>
>> and
>>
>> http://api.openstreetmap.org/api/0.6/way/33598566/full
>>
>>
>> <osm version="0.6" generator="OpenStreetMap server">
>> <node id="435585572" lat="-6.8168601" lon="26.4588445" version="1"
>> changeset="1756085" user="tmcw" uid="957" visible="true"
>> timestamp="2009-07-06T20:06:58Z"/>
>> <node id="383826997" lat="-6.8168332" lon="26.4588445" version="1"
>> changeset="932219" user="tmcw" uid="957" visible="true"
>> timestamp="2009-04-24T19:50:58Z"/>
>> <node id="382129524" lat="-6.8141755" lon="26.4607324" version="1"
>> changeset="913788" user="tmcw" uid="957" visible="true"
>> timestamp="2009-04-23T16:01:34Z"/>
>> <node id="381350181" lat="-6.811628" lon="26.4619548" version="1"
>> changeset="906941" user="tmcw" uid="957" visible="true"
>> timestamp="2009-04-22T21:43:50Z"/>
>> <node id="435967399" lat="-6.8071957" lon="26.4625141" version="1"
>> changeset="1758543" user="tmcw" uid="957" visible="true"
>> timestamp="2009-07-07T03:20:33Z"/>
>> <node id="435950080" lat="-6.8008796" lon="26.4632973" version="1"
>> changeset="1758478" user="tmcw" uid="957" visible="true"
>> timestamp="2009-07-07T02:57:57Z"/>
>> <node id="383788447" lat="-6.7880346" lon="26.4728462" version="1"
>> changeset="931940" user="tmcw" uid="957" visible="true"
>> timestamp="2009-04-24T19:33:36Z"/>
>> <node id="382152797" lat="-6.7792926" lon="26.4795131" version="1"
>> changeset="913918" user="tmcw" uid="957" visible="true"
>> timestamp="2009-04-23T16:11:20Z"/>
>> <node id="435333843" lat="-6.7741543" lon="26.48248" version="1"
>> changeset="1753826" user="tmcw" uid="957" visible="true"
>> timestamp="2009-07-06T16:00:05Z"/>
>> <node id="383669934" lat="-6.7659956" lon="26.4868391" version="1"
>> changeset="930324" user="tmcw" uid="957" visible="true"
>> timestamp="2009-04-24T17:28:07Z"/>
>> <node id="435662873" lat="-6.7578006" lon="26.4925017" version="1"
>> changeset="1756781" user="tmcw" uid="957" visible="true"
>> timestamp="2009-07-06T21:14:28Z"/>
>> <node id="382074731" lat="-6.7525679" lon="26.4956324" version="1"
>> changeset="913426" user="tmcw" uid="957" visible="true"
>> timestamp="2009-04-23T15:30:06Z"/>
>> <way id="33598566" visible="true" timestamp="2009-10-27T12:38:24Z"
>> version="3" changeset="2964814" user="tiger" uid="16558">
>> <nd ref="435585572"/>
>> <nd ref="383826997"/>
>> <nd ref="382129524"/>
>> <nd ref="381350181"/>
>> <nd ref="435967399"/>
>> <nd ref="435950080"/>
>> <nd ref="383788447"/>
>> <nd ref="382152797"/>
>> <nd ref="435333843"/>
>> <nd ref="383669934"/>
>> <nd ref="435662873"/>
>> <nd ref="382074731"/>
>> <tag k="highway" v="road"/>
>> <tag k="ID" v="28558"/>
>> </way>
>> </osm>
>>
>>
>> there are some significant differences: The TRAPI output contains wrong
>> nd_refs and also doesn't include all referenced nodes.
>> It seems that changeset # 2964814 caused this corruption.
>> I already sent this error report to mat as his TRAPI server is concerned.
>>
>>
>> Best regards
>>
>> Alfons
>>
>>
> It doesn't look like it's a problem with just mat's server since my TRAPI
> output is identical to mat's.
>
> If the data is the problem, then it must be something in the TRAPI code
> since both of our servers return the same thing.
>
> Here's the output from my TRAPI:
>
> <osm version="0.5" generator="Trapi 0.3 - Milenko">
>
> <!-- some nodes from z0 0 0 -->
> <!-- some nodes from z12 2349 2125 -->
>
> <node id="435333843" lat="-6.7741543" lon="26.48248"/>
> <node id="435662873" lat="-6.7578006" lon="26.4925016"/>
> <node id="435950080" lat="-6.8008796" lon="26.4632973"/>
> <node id="435967399" lat="-6.8071957" lon="26.4625141"/>
> <!-- some ways from z12 2349 2125 -->
> -
>
> <way id="33598566">
> <nd ref="436000366"/>
> <nd ref="435967205"/>
> <nd ref="435951365"/>
> <nd ref="436060715"/>
> <nd ref="435738062"/>
> <nd ref="435967399"/>
> <nd ref="435950080"/>
> <nd ref="435359028"/>
> <nd ref="435589020"/>
> <nd ref="435333843"/>
> <nd ref="435843476"/>
> <nd ref="435662873"/>
> <nd ref="435683994"/>
> <tag k="highway" v="road"/>
> <tag k="ID" v="28558"/>
> </way>
> </osm>
>
> -Jeremy
>

Sorry to reply to myself, but that changeset is massive [1].  It took over a
half hour to run [2] which means it was missed by the minute-slow diffs that
the TRAPI servers use.  We'll both have to roll back to the 27th and apply
updates from there.

I see Mat's server is down atm, so I assume that's what he's doing.  Once
his server comes back online, then I'll take mine down and do the same.

-Jeremy

[1]    http://www.openstreetmap.org/browse/changeset/2964814
[2]    http://api.openstreetmap.org/api/0.6/changeset/2964814
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tilesathome/attachments/20091029/77a39994/attachment.html>

More information about the Tilesathome mailing list