We are considering moving Elasticsearch from vsan to fnas or synology nas.
We are thinking that the IO processing speed will be reduced, but how much is the impact?
Or are there any hardware requirements for IO processing speed in the Elastic product line?
If my questions are not sufficient, please give me the information I need to make a decision.
No, Elasticsearch doesn't really have any minimum requirements for IO performance. As long as your storage behaves correctly and your application doesn't need it to run fast you should be fine. You will need to run your own experiments to determine whether you can get the performance your application needs. We recommend using Rally for these experiments.
I was under the impression that Elasticsearch was constantly reading and writing to disk, so I was under the impression that the impact of processing speed would be noticeable, but I am relieved to hear that.
I just send logs from over 30 servers via Filebeat, receive them via Logstash, and collect them via Elasticsearch. Of course I am viewing them in Kibana.
As long as your storage behaves correctly and your application doesn't need it to run fast you should be fine.
I'm not doing anything special, so the above points don't apply to me?
Indeed. I think you are right.
As for the speed, it is good to actually experience it.
Just to be sure. Is it my understanding that Elasticsearch is guaranteed to work even if the processing speed depends on disk capacity?
For example, is there any possibility of missing logs or unstable operation?
You won't miss any logs in any case and the cluster should be stable as long as the storage is not so slow that it appears to be completely broken (e.g. takes tens of seconds to perform writes).
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.