We have 2 indices each one 5 shards and parent child indexed . each index
around 150gig size.
our hw is : 128 gig ram 28 core .
1- starting the ES takes too long , around 20 min+ and when its started
and we can see the green in the head plugin, ES using around 60gig heap !
(big desk). it is just starting to ES . There is no query and index .
I gave 64 gig heap to open the indexes.. below this size I couldnt open.
Is there any idea why ES using such big amount of memory ?
ES 0.90.7 . 150 GB is the folder size . We are also storing the document
. I think the segment merging process using a lot memory. We have 12
index each one is around 150 GB. We will use 3 or 4 machine cluster for
pre-production test.
Thx
On Wednesday, December 11, 2013 9:42:39 PM UTC+2, Jörg Prante wrote:
What ES version is this? Can you check the file sizes in the data folder?
150g = ~30g per shard is quite a lot.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.