Elasticsearch 1.7.3 http.max_content_lenght ignored?

(Dopch) #1


On Elasticsearch 1.7.3
On Laptop with 64GB OF RAM
Number of instance 3 X 12GB
file size = 405M
My config File :
logs: /var/log/company/elasticsearch/instance1
data: /opt/company/elasticsearch/instance1

cluster.name: company
node.name: "company-1"
index.merge.scheduler.max_thread_count: 1

threadpool.bulk.type: fixed
threadpool.bulk.size: 8
threadpool.bulk.queue_size: 300

indices.memory.index_buffer_size: 30%
index.refresh_interval: 30s
index.fielddata.cache: node
indices.fielddata.cache.size: 40%
http.compression: true
index.number_of_replicas: 0
bootstrap.mlockall: true

script.indexed: on
script.dynamic: on
script.file: on
script.inline: on
script.groovy.sandbox.enabled: true
http.max_content_length: 1gb

When i curl for debugging "curl -XPOST '' -d @file
curl: (56) Recv failure: Connection reset by peer

Elasticsearch log :
Caught exception while handling client http traffic, closing connection [id: 0x18d0d451, / => /]
org.elasticsearch.common.netty.handler.codec.frame.TooLongFrameException: HTTP content length exceeded 104857600 bytes.

Now on my server
On Elasticsearch 1.7.3
On Server with 256GB OF RAM
Number of instance 6 X 12GB
file size = 405M
curl -XPOST '' -d @file return success and no error on ES log

For information on server side or on client side when i curl -XGET 'http://localhost:9200/_nodes' | grep http.max_content_length i see "http.max_content_length:":"1gb"

(system) #2