It depends on what you want to cache. In most cases with ES, cache is
overkill because ES uses internal caching, and it easily ends up in the
double caching problem which affects performance.
Saving ES query results? Depends on your queries, their runtime,
concurrency, length of result set, network latency...
Saving Java objects on the heap? Look at Guava
Or a generic, volatile key/value memory cache? Look at memcached http://memcached.org/
We are planning to store the 2 Elasticsearch cluster.
Cluster 1 will store the data using neoFs file-system.
Cluster 2 will store the data using mMap file-system and will store the
result of queries fired on cluster1. Later, we would search on cluster 2 to
apply some more filtering/condition on fetched records.
Regards,
Ankit Jain
On Wednesday, 20 November 2013 16:25:23 UTC+5:30, Jörg Prante wrote:
It depends on what you want to cache. In most cases with ES, cache is
overkill because ES uses internal caching, and it easily ends up in the
double caching problem which affects performance.
Saving ES query results? Depends on your queries, their runtime,
concurrency, length of result set, network latency...
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.