ElasticSearch Curator


(Phani Nadiminti) #1

Hi All,

I have implemented elastic search curator and i have installed on my node1 for implementations.I am specifying ip and port of elastic search since local host is not reachable.
The following is the command for curator i am running to take backup of entire cluster with all indices.

curator --host 10.0.20.2 --port 9200 --logfile /tmp/curator_snapshot.log --loglevel WARN --master-only snapshot --repository es_new indices --all-indices

so here i installed curator on node1 only.

  1. Do i need to install curator on each node? If no need to install curator on each node, what is the process of curator will it pickup master by default and perform snapshot operation?

Let say i am specifying master node ip it could be fine with out specifying --master-only parameter

I read in curator docs about --master-only feature tell cluster to create snapshot from master node only.

let say the ip specified in curator is not master node ip but i am giving create snapshot from master only using (--master-only). in this case what is the approach of curator?

please help me out to understand curator.

Thanks & Regards,
phani


(Magnus Bäck) #2

Do i need to install curator on each node?

No.

If no need to install curator on each node, what is the process of curator will it pickup master by default and perform snapshot operation?
Let say i am specifying master node ip it could be fine with out specifying --master-only parameter

The point of the --master-only option is to only run the actions if the node you're connecting to is the master node. That way you can have a cron job with Curator on each machine in the cluster but only one of the Curator processes will actually do anything (the rest will say "nope, this isn't the master node, good bye"). The point is of course that you can take down any node in the cluster without affecting your Curator jobs.


(Phani Nadiminti) #3

Hi Magnus,

Thank you for the reply. if i installed curator on node1 if node1 is goes down i observed that it is not taking snapshots to the specified node host because it is down so how can we achieve the fail over to take snapshots.

Thanks & Regards,
phani


(Magnus Bäck) #4

Run Curator on all hosts in the cluster and use the --master-only option as I explained earlier.


(Phani Nadiminti) #5

Hi Magnus,

Yes it is working fine thank you for the support.

Phani


(system) #6