Unassigned shard issue

Elastic Stack Elasticsearch
1

Hello Elasticusers,

I have a big problem since yesterday, when I create a new index in Elasticsearch (3 shards and 1 replica), it is unassigned and it stays in this state...

I never has had this type of issue before... Does somebody know what is the problem ?

This is more information about my Elasticsearch cluster :

I have 2 nodes, one node has 7gb of memory and the other node has 4gb of memory.

Hope you could help me, I'm losing my mind with this problem.

2

Which version of elasticsearch are you using. Did you check the logs of your cluster ? 240 indices and 728 shards seems huge for a 2 nodes setup, could you explain why you need so many indices/shards ?

3

I'm using Elasticsearch 2.2.0, I checked the logs files and I didn't see something special. I have 240 indices because some indices by day in order to see the date evolution. Do you have any idea?

4

I have found an interesting log :

[WARN ][indices.memory] failed to set shard [my_index] index buffer to [31mb]
org.apache.lucene.store.AlreadyClosedException: translog [1] is already closed