[SOLVED] Unknown 'cluster/nodes/info' messages on EC2/Mesos


I have elasticsearch (the official elasticsearch 2.2.0 docker image) deployed on a mesos cluster on EC2 using marathon.

For some reason, I keep getting the following errors:

[2016-04-13 14:37:18,338][WARN ][transport.netty          ] [Celestial Madonna] exception caught on transport layer [[id: 0xa8faa383, / => /]], closing connection
java.lang.IllegalStateException: Message not fully read (request) for requestId [145], action [cluster/nodes/info], readerIndex [39] vs expected [57]; resetting

Even though the only thing running in my mesos cluster is a single elasticsearch instance, I get this error coming from all the machines in the cluster (maybe this is related to the discovery protocol?).

I don't have any health checks running on marathon and I'm running the vanilla elasticsearch docker image.

Any clues on what might be causing this?

Update: I believe I found out the problem. I had an old version of logstash installed on the AMI.