we have 5 data nodes and 3 master in cluster and we are on ES 1.0
2 week ago when we restarted the it leave 1 shard unassigned for some indices.
We check the Lucien index and they are fine. We create a test index and copy the Lucien index to that and it came back.
But the we don’t know what is the problem why restart leave unassigned shard. There is nothing help full in logs
The segments.gen file are missing for those shards. And we have 2 data directories.
Andy hell will be appreciated