Comments on: Running out of disk space with Elasticsearch and Solr: a solution http://www.flax.co.uk/blog/2016/04/21/running-disk-space-elasticsearch-solr/ The Open Source Search Specialists Tue, 12 Feb 2019 14:44:32 +0000 hourly 1 https://wordpress.org/?v=4.9.8 By: Tom http://www.flax.co.uk/blog/2016/04/21/running-disk-space-elasticsearch-solr/#comment-37937 Wed, 27 Apr 2016 09:07:06 +0000 http://www.flax.co.uk/?p=3241#comment-37937 Hi David, thanks for the comment. We had automatic shard reallocation turned off, because this project used a hot/warm search architecture. Other than that, we had the default values for the watermarks (85% and 90% I believe). In any case, this never threw any 503s, and we kept hitting 100%. The ES documentation is unfortunately not very clear on the precise behaviour when hitting a watermark. – Tom

]]>
By: David Ecker http://www.flax.co.uk/blog/2016/04/21/running-disk-space-elasticsearch-solr/#comment-37926 Tue, 26 Apr 2016 16:29:54 +0000 http://www.flax.co.uk/?p=3241#comment-37926 How is this used in conjunction with low and high watermarks? Couldn’t you just set those to manage your disk space? When the low watermark is breached, new shards are unable to be created and a 503 is thrown IIRC. Is your solution designed to prevent new data being added to existing shards over disk utilization percentage?

]]>