I'm thinking this is a "dumb" question but I have not been able to find the answer online. I have setup and tested using _snapshot to backup and restore my single node ES test cluster. I need to set this up for our production clusters which are made up of 3 or more nodes. Do I need to run my backup script on only one of the nodes to backup the cluster or must I run it on all 3 ? If I only run it on one node, does the NFS mounted repo need to be mounted on the other 2 nodes at time of backup?
Do I need to run my backup script on only one of the nodes to backup the cluster or must I run it on all 3 ?
You should run it on one of the nodes.
If I only run it on one node, does the NFS mounted repo need to be mounted on the other 2 nodes at time of backup?
The repository directory must be accessible to all nodes. Otherwise all data would have to be funneled through the node that does have access and that doesn't make sense for clusters of non-trivial size.
Thank you for the quick response.