We are testing out release 1.1.1 and during our indexing performance testing, we seemed to get significantly slower throughput, the document/second rate is about 30% slower. We used the exact same yml file and startup settings. The code is also identical except for the breaking changes in the java client api (in this case minor naming changes) and different elasticsearch/lucene jars.
We have a 4 node test cluster. The test basically creates an index with 4 shards and no replicas/refreshing. We are indexing documents that are about 2KB each using the bulk api (500 documents per request). Below is some environment info and some settings that we changed from the default. MemTotal: 99016988 kB ES_HEAP_SIZE=24g MAX_OPEN_FILES=65535 */etc/elasticsearch/elasticsearch.yml :* cluster.name: estest node.name: "es1" node.rack: rack2 bootstrap.mlockall: true network.host: 1.2.3.4 gateway.recover_after_nodes: 3 gateway.expected_nodes: 3 discovery.zen.minimum_master_nodes: 1 discovery.zen.ping.multicast.enabled: false discovery.zen.ping.unicast.hosts: ["es1", "es2", "es3", "es4"] indices.memory.index_buffer_size: 50% index.translog.flush_threshold_ops: 50000 threadpool.search.type: fixed threadpool.search.size: 20 threadpool.search.queue_size: 100 threadpool.index.type: fixed threadpool.index.size: 60 threadpool.index.queue_size: 200 threadpool.bulk.type: fixed threadpool.bulk.size: 50 threadpool.bulk.queue_size: 1000 cluster.routing.allocation.awareness.attributes: rack -- 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 [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/edea9e8f-e2f3-424e-b381-3d6dc4b96979%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
