Size on Disk increased after reindexing documents in ES2.2

(Pir Abdul Rasool Qureshi) #1

We recently created an index with ElasticSearch 2.2. We repopulated the index with all the documents we have in our old ElasticSearch 1.7.2 cluster. The size on disk for each shard on ES 2.2 is approx. 2 times bigger than what it was on ES 1.7.2. Total number of documents in both clusters is same.

I am confused as I do not find any reason for that increase in size. Is there something which I am missing?

It is worthy to mention here that we use "_cat/shards" to find out size on disk for each shard.

(Christian Dahlqvist) #2

Elasticsearch 2.x has doc values enabled by default, which can cause a larger on disk footprint. Having these data structures on disk helps reduce the heap usage, which is generally beneficial and a worthwhile tradeoff.

(system) #3