Multiple Index Data directories


#1

We recently suffered a power outage and the 2 nodes become split brained. I managed to bring it back to green, but suffered a bit of data loss. When I was review the index data directories, I noticed that one of the nodes had more than one shard index.
.
├── 0
│ ├── indices
│ ├── node.lock
│ └── _state
└── 1
├── indices
├── node.lock
└── _state

While the other node only had one
.
└── 0
├── indices
├── node.lock
└── _state

Looking for guidance as to why this duplicated and how can I clean this up properly? Thanks.


(Mark Walkom) #2

That would be because more than one instance of ES started on the same host.
You can start a second instance so that it recovers that shard data, then reallocate it off, then shutdown the second process.


#3

So after checking into this. the correct folder on the node with two indices is folder "1". how do I go abouts cleaning this up? or does it matter if I remove the "0" folder?


(Mark Walkom) #4

I wouldn't just delete things unless you want data loss.


(system) #5

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