I've started the es cluster on 2 EC2 instances (medium, 2 core, 3.75G memory) using Amazon Linux, and Using java TransportClient to connect to the cluster (See my other email for why I didn't using native client. Which has no answer yet)
when the cluster start everything is ok but when it running for couple of days or a week it sometimes freeze when I want to update the index. In my business logic we have some small frequently updates. After I restart the es cluster it went back ok again.
Can anyone give me some advice to avoid the problem? At lease it can keep running for a month without problem.