Michael Thank you for identifying this issue. There is indeed a problem. I think I have located the cause and I'm working on a fix.
80n On Sat, Nov 15, 2008 at 3:07 PM, Michael Ritzert < [EMAIL PROTECTED]> wrote: > Hi all, > > (at least) the hypercube XAPI server seems to be missing nodes. > > $ wget ' > http://osmxapi.hypercube.telascience.org/api/0.5/*[bbox=8.4571,49.4855,8.4607,49.4873]<http://osmxapi.hypercube.telascience.org/api/0.5/*%5Bbbox=8.4571,49.4855,8.4607,49.4873%5D>' > -O MA1.osm > $ wget ' > http://www.openstreetmap.org/api/0.5/map?bbox=8.4571,49.4855,8.4607,49.4873' > -O MA2.osm > $ grep '<node id' MA1.osm | wc -l > 55 > $ grep '<node id' MA2.osm | wc -l > 111 > $ grep 249819997 MA[12].osm > MA2.osm: <node id="249819997" lat="49.4866651" lon="8.4581842" > user="[HEK]the_other" visible="true" timestamp="2008-03-01T17:08:25+00:00"> > MA2.osm: <nd ref="249819997"/> > MA2.osm: <nd ref="249819997"/> > $ grep '<way id' MA1.osm | wc -l > 12 > $ grep '<way id' MA2.osm | wc -l > 21 > > When I checked to see how many nodes where added today, I found that only > two of three nodes updated during the same second today made it to the XAPI: > > $ grep 2008-11-15 /tmp/MA* > /tmp/MA1.osm: <node id='30391456' lat='49.4872431' lon='8.4599382' > timestamp='2008-11-15T14:24:47Z'> > /tmp/MA1.osm: <node id='30391457' lat='49.4869184' lon='8.4596203' > timestamp='2008-11-15T14:24:47Z'> > /tmp/MA2.osm: <node id="30391456" lat="49.4872431" lon="8.4599382" > visible="true" timestamp="2008-11-15T14:24:47+00:00"> > /tmp/MA2.osm: <node id="30391457" lat="49.4869184" lon="8.4596203" > visible="true" timestamp="2008-11-15T14:24:47+00:00"> > /tmp/MA2.osm: <node id="30391455" lat="49.4876508" lon="8.4589878" > visible="true" timestamp="2008-11-15T14:24:47+00:00"> > > Note that 30391455 is missing from the XAPI (MA1) data. From the node ids, > one should guess that they have been created within a second, also. When I > tried to verify this, I found that the creation of the missing node also > hasn't been recorded in the main history: > http://www.openstreetmap.org/browse/node/30391455/history > The other's are OK: > http://www.openstreetmap.org/browse/node/30391456/history > Maybe that's starting point for debugging this. > > I noticed this in other areas, too. > More extensive research revealed three "recent" periods of time where > timestamps of missing nodes accumulated: > On 2008-09-21, 2008-09-23, 2008-11-02. > Of course this data is only representative for the small area I checked. > > Michael > > > > _______________________________________________ > Tilesathome mailing list > [email protected] > http://lists.openstreetmap.org/listinfo/tilesathome >
_______________________________________________ Tilesathome mailing list [email protected] http://lists.openstreetmap.org/listinfo/tilesathome
