Lets assume , we are sending bulk request to a node and lets assume that node is not where the documents in the bulk is supposed to go , based on shard the document is mapped to.
Here what will happen first , queuing to the thread pool queue or redirecting the request ?
Also when the bulk is processed , is it again sent to the index queue or is it directly processed ?
So if i get it right , when a bulk request happens , its is divided into individual index requests and forwarded to the right nodes without pushing the data to queue right.
If this is so , i was wondering if there are huge number of bulk requests wont the machine go out of resource just processing this itself ?
I guess it could, but you'd need these bulk requests to be insanely large in order to trigger this issue, and most of them would be rejected at the shard level.
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.