I have configured big cluster in my production env. just wanted to confirm do we need to allocate half of memory in master node also. Suppose i have 16G node and do i need to allocate 8G in jvm options file what if i allocate let say 6G does this will effect my reindexing process which is happening in master nodes.
Actually i have configured it with 6G in prod env. but my reindexing is taking so much time to complete i have data in GB's.
Also i want to clarify one do i need to set half of memory as recommended by elasticsearch, in kibana server also. Could you please help me out suitable configuration for kibana servers if i have 16G node.
You can be a bit more nuanced with master only nodes, as they don't hold data. This means you can use more than 50%, eg 75% or so depending on how much you have on the host.
Thanks for your support. is there possibility that if we not define only 50% of memory to master node as it does not contain data then our reindexing process will become slow??
Thanks warkolm for your revert. I am doing re-indexing and primary and replica size is 300 GB each but re-indexing is in progress from last 15 days. Is it expected to take this much time for re-indexing of 300 GB data? If no, what should I check for this slowness.
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.