Question about Elasticsearch version 1.7
I have been trying to use bulk queue size as 1000 using transient setting, which seems to be propogated on all the hosts and I verified by looking at /_nodes
thread_pool:
bulk:
type: "fixed"
min: 2
max: 2
queue_size: "1k"
As our persistent setting is defined as
threadpool:
bulk:
size: "2"
queue: "8"
Considering transient settings take preference the max bulk queuing on each data node should be 1000, but in our cluster there are nodes running more than 1000 requests getting queued. Any quick ideas on debugging if the setting is defauting to -1 (no bound ) ?