Help on ---- remote server bulk load

Yes please. Will be happy to hear news from your tests.

About Virtual machines, I definitely prefer having one single machine per hardware box. You can start a JVM with 30Gb RAM instead of 10 VM with smaller RAM.
Also, don't forget that in case of a crash (hard disk failure for example), you will probably lost all your replicas as well (as they are on the same physical box - even if it's another virtual node)…

--
David Pilato | Technical Advocate | Elasticsearch.com
@dadoonet | @elasticsearchfr | @scrutmydocs

Le 14 mai 2013 à 11:40, rafi me.mrafi.en@gmail.com a écrit :

Hi, David Pilato
Thanks a Lot for Replying. and your posts are very useful to me.

IMHO, the best you can give to ES is to have one shard per node and one node per server.
So, with 8 shards, you can have up to 8 nodes so up to 8 servers.
Don't run more than one node on a single server (or only if you have more than 64Gb of RAM).

             Ok David., Now I got some clear idea on it.

Actually., The Server Machine which we are using has 64Gb of Ram but as of Our Works we have Splited our Server into approximately 10 Virtual Machines
And now, our server has 10Gb which we are using for ES Activities

That said, you can try to tune it and start with only two shards as you have two nodes. Then increase the number of shards, reindex and see where it goes.

       So now we will use Only One Node[that means we will run only one ES Instance] and we will prepare some statistics for "having 1shard, 2 shards, ..  for testing the speed of index operation".
 And I will Let U Know what happen? 

Thanks again for Replying.

Best Regards
Mohammad Rafi.

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.