ES client "drops requests" randomly (multi-client nodes setup)

Hi all,
we are running a cluster (ES v2.3.1) with two client nodes (namely C1,C2).
Production users send requests to C1 only while C2 is used for testing purposes only.
It happens very rarely that C1 is not responding to a request (we know for sure that C1 received the request but it is not replying). From the end-user point of view, the request remains pending (i.e., no timeouts are expiring nor exceptions appear on the ES side).
Any idea?
Is this possibly due to a bad interaction between C1 & C2?

Thanks in advance for your help.


do you have any hunch if this is a special kind of request? How do you know C1 received the request?

Hi Christoph,
This problem is not related to a specific kind of request. It changes almost everytime.
We know that C1 received the request because we captured the network packets by means of tcpdump.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.