From a81a4a5efef916f5fff426b3791cee25dfebc49d Mon Sep 17 00:00:00 2001 From: Martijn van Groningen Date: Thu, 20 Feb 2014 16:43:06 +0100 Subject: [PATCH] [DOCS] Included the `_percolator` index breaking change to migration docs. --- docs/reference/migration/migrate_1_0.asciidoc | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/docs/reference/migration/migrate_1_0.asciidoc b/docs/reference/migration/migrate_1_0.asciidoc index 0cc10ae06bd..b69288a71ab 100644 --- a/docs/reference/migration/migrate_1_0.asciidoc +++ b/docs/reference/migration/migrate_1_0.asciidoc @@ -362,3 +362,15 @@ in the query string. * The `custom_score` and `custom_boost_score` is no longer supported. You can use <> instead. + +== Percolator + +The percolator has been redesigned and because of this the dedicated `_percolator` index is no longer used by the percolator, +but instead the percolator works with a dedicated `.percolator` type. Read the http://www.elasticsearch.org/blog/percolator-redesign-blog-post/[redesigned percolator] +blog post for the reasons why the percolator has been redesigned. + +Elasticsearch will *not* delete the `_percolator` index when upgrading, only the percolate api will not use the queries +stored in the `_percolator` index. In order to use the already stored queries, you can just re-index the queries from the +`_percolator` index into any index under the reserved `.percolator` type. The format in which the percolate queries +were stored has *not* been changed. So a simple script that does a scan search to retrieve all the percolator queries +and then does a bulk request into another index should be sufficient. \ No newline at end of file