Closed brusic closed 6 years ago
Can you disable bulk mode, so then we can see the exception logged on per node/document basis and paste the logs here then.
Thanks
Was not even aware there was a bulk mode. After disabling bulk updates, the error is now clear: "Rejecting mapping update .. more than 1 type". I guess I need to look into the Json Mapper.
Was about to work on a fix and I started reading the Json Mapper documentation. The plugin requires a mapping file, which is to be placed inside the conf
directory, however I am using GrapheneDB, which does not allow adding files. The server extension documentation states: Only files with jar format or neo4j-server.properties are extracted, the rest will be ignored.
It would appear that the two are not compatible. Before I go down a possible fool's errand, is it possible to use GrapheneDB and Elasticsearch 6+?
HI @brusic We already deployed neo4j-to-elasticsearch with custom mapping.json file on GrapheneDB, You can just contact them and they will explain what is the procedure to follow.
Feel free to put me in cc of your conversation with GrapheneDB support : christophe (at) graphaware (dot) com
Thanks again. Closing this issue once again. :)
Will test locally and then reach out to support. Wish it was more automated.
When indexing to elasticsearch via graphaware, the nodes index is being update, but fails when attempting to index the relationships:
2018-10-03 19:39:12.759+0000 WARN [c.g.m.e.ElasticSearchWriter] 6 operations could not be replicated to Elasticsearch. These updates got lost.
I can see the changes in the -node index, so connectivity is working. No errors on the elasticsearch side.
Using the standard settings (com.graphaware.module.ES.* settings omitted)
index settings
Connecting to Elastic Cloud via GrapheneDB with the GraphAware extension using the three jars provided. Using the appropriate jars for Neo4j 3.4.5