Elasticsearch recovery process settings -
running es 1.5.2
i'm testing recovery settings on 4 node cluster (32 cores, 30gb jvm heap, 5tb ssd per machine) 1700 shards. cluster in state of recovery , updated values such:
"transient": { "cluster": { "routing": { "allocation": { "disable_allocation": "false", "cluster_concurrent_rebalance": "8", "node_concurrent_recoveries": "8", "node_initial_primaries_recoveries": "16" } } }, "indices": { "recovery": { "concurrent_streams": "9" } } if recovery in process , settings updated take effect or on next possible recovery?
currently see 8 primaries being initialized , 2 nodes working. i.e: 30% network utilization , barely no cpu activity.
Comments
Post a Comment