Reputation: 2021
I have a problem with an elasticsearch instance that refuses to start up and is stuck on Initializing. Elaticsearch stuck using 99% cpu and all shards are Initializing
My guess is that something is wrong with the data stored in one of the indexes. So I am thinking I might just go ahead and remove all the logs and start over. This is not a production system and the logs so far will probably never be used.
However, since the system is still stuck with trying to recover I am not able to remove the index using the REST API. So i was thinking I could probably remove it manually by removing the physical file/folder. So my questions are:
Upvotes: 1
Views: 1449
Reputation: 935
As you are not posting your ES version , I will assume that you are running version 2.x
In my local version elasticsearch-2.3.5
and based on my cluster configuration :
cluster.name: local
Data is stored at
elasticsearch-2.3.5\data\CLUSTERNAME\nodes\x\indices
Where x is the number of the node.
The folder structure here is as follows :
So if you have only one node
where replication is not present , if you backup that folder and restore it later , it is okay to do so. Now , if we're talking about more than one node . You will need to stop both nodes before you try to do such operation to avoid possible conflicts.
Steps I did :
-Stop ElasticSearch
-Backup `elasticsearch-2.3.5\data\local`
-Remove `elasticsearch-2.3.5\data\local`
-Start the server
Upvotes: 1