Unable to tesselate shape

[mhugo]

Elasticsearch Version

7.17.6

Installed Plugins

icu

Java Version

bundled

OS Version

Linux ThinkPad-P15s-Gen-1 5.19.0-46-generic #47~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Wed Jun 21 15:35:31 UTC 2 x86_64 x86_64 x86_64 GNU/Linux

Problem Description

With some geo shapes, the indexing fail with an "Unable to tesselate" error, whereas the geo shapes seem to be valid polygons.

Example of such a shape:
sample.geojson.txt

This shape is valid according to gdal/ogr/geos tools.
I suspected initially a difference between GEOS-based tools that work with a default precision model of 32bit floating points when ES works in 16bit. I've then tested for validity using JTS with the FLOATING_SINGLE precision model and it is still valid ...

Steps to Reproduce

• Create an index with a column mapped to a geo_shape type
• Index the supplied example as a new doc using the geo_shape column

ES should respond with an "unable to tesselate" error.

Logs (if relevant)

No response

[elasticsearchmachine]

Pinging @elastic/es-analytics-geo (Team:Analytics)

[iverase]

Hello @mhugo,

The version 7.17.6 is not the latest 7. 17.x version and there has been some fixes in the Lucene tessellator on later versions. In particular there has been fixes introduced in 7.17.9 and 7.17.11.

I tried to index the provided example in 7.17.6 and indeed it failed, but after upgrading to 7.17.9, I was able to ingest the polygon successfully. Therefore I recommend to upgrade to 7.17.11+ to get the fixes.

I will close this issue now, please feel free to open an issue if you have issues after the upgrade.

[mhugo]

@iverase Thank you very much for your feedback, I should have tried the last version first