Socket time out error during bulk indexing

Hi All,

We are using 3 physical servers and each server has 1 master node, 1 data node and only one server has one client node. When we are doing bulk indexing in the cluster we are frequently getting the socket time out error. Below is the error from log.

[2016-02-05 04:51:48,599][ERROR][marvel.agent.exporter ] [tparhebfmi003_PERF_MASTER] error sending data to [http://marvel:XXXXXX@tparhebfmi003:9200/.marvel-2016.02.05/_bulk]: SocketTimeoutException[Read timed out]
[2016-02-05 04:56:44,965][ERROR][marvel.agent.exporter ] [tparhebfmi003_PERF_MASTER] error sending data to [http://marvel:XXXXXX@tparhebfmi003:9200/.marvel-2016.02.05/_bulk]: SocketTimeoutException[Read timed out]
[2016-02-05 05:02:04,717][ERROR][marvel.agent.exporter ] [tparhebfmi003_PERF_MASTER] error sending data to [http://marvel:XXXXXX@tparhebfmi003:9200/.marvel-2016.02.05/_bulk]: SocketTimeoutException[Read timed out]
[2016-02-05 05:05:38,440][ERROR][marvel.agent.exporter ] [tparhebfmi003_PERF_MASTER] error sending data to [http://marvel:XXXXXX@tparhebfmi003:9200/.marvel-2016.02.05/_bulk]: SocketTimeoutException[Read timed out]
[2016-02-05 05:14:35,195][ERROR][marvel.agent.exporter ] [tparhebfmi003_PERF_MASTER] error sending data to [http://marvel:XXXXXX@tparhebfmi003:9200/.marvel-2016.02.05/_bulk]: SocketTimeoutException[Read timed out]

I want to know the reason for getting this exception. Is it because of any network latency?

How to avoid this socket time out error?

Please help me into this.

Regards,
Ganeshbabu R