Index not replicated but split up in ES 1.4.2

Hi we recognised that something is wrong with one of our indices.
curl -XPOST 'http://user:password@localhost:9200/indice/devices/_search' -d '{ "query": { "query_string": { "query": "panel:abcdefg AND is:fixed AND lastInteraction:{now-2d TO now}" } }, "sort": [ { "id": "asc" } ], "from": 0, "size": 25, "fields": [] }'
We have 2 data nodes. if i do the search above twice (each search ran against an other node : checked with _search_shards), i get lets say 500 results for the first and 1000 results for the 2nd search. The expected Resultset should contain 1500 items. So our result set is split up. we have more indices like this but this is the only one that has this behavior.

How to get that repaired ?
What could cause this?

Might you have a cluster that is suffering from a split-brain? If both your nodes are master eligible, do you have minimum_master_nodes correctly set to 2 as per these guidelines (still applies to version 1.4.2, even though it is very, very, very old...).

You should also really look to upgrade...

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.