I disabled shard replication (by setting number of replica to 0) when I loaded a tons of data into elasticsearch. After the load finish I enabled the replication (by setting the number to 2). Now I can see elasticsearch started transferring data to replication shards.
At this moment when I query the index or send bulk indexing again, I got "Read timed out. (read timeout=30))". As you can see I've increased the time out to 30 seconds. I can see the cluster is transferring replication data from the head plugin. So I'm guessing it's utilizing all the thread and hence blocking any new queries.
Is there anything I can do to make rooms for searching?
Here's my cluster settings:
{
"persistent": {
"indices": {
"store": {
"throttle": {
"type": "merge",
"max_bytes_per_sec": "100mb"
}
}
}
},
"transient": {
"cluster": {
"routing": {
"rebalance": {
"enable": "none"
},
"allocation": {
"enable": "all"
}
}
},
"threadpool": {
"bulk": {
"queue_size": "1000",
"size": "16"
},
"index": {
"queue_size": "1000",
"size": "16"
}
},
"indices": {
"store": {
"throttle": {
"type": "none"
}
}
}
}
}
Thanks in advance!