Hi ,
We are facing Old GC's from time to time , i like to know if the amount of replica's has a direct impact on the ES memory utilization (looks to me that the answer is yes ) ?
bottom line , if it can increase chance for Old GC's ?
Thanks
Hi ,
We are facing Old GC's from time to time , i like to know if the amount of replica's has a direct impact on the ES memory utilization (looks to me that the answer is yes ) ?
bottom line , if it can increase chance for Old GC's ?
Thanks
Yes but it's more complicated than saying it's double the impact.
Thanks Mark .
Can you elaborate on it ?
I will be more specific , on my case .
Assuming my cluster is bigger (many node’s ) one , and we don’t use query routing , so basically each query runs on all cluster nodes.
ES , will used both replica && primary shards to scale my query’s , in term of memory utilization on a given none for specific primary shard , we may have memory utilization of replica shards from other nodes ( or ,and I don’t know if its doable case , memory utilization of replica from the same node , so the memory will include duplicate data set’s , for the primary + shard of the same data set’s )
Thanks
You are likely to "pay" twice the amount for maintaining shards as you have twice the number. But when you start looking at filter/query/agg caches it really depends on what gets queried where.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.
© 2020. All Rights Reserved - Elasticsearch
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.