Configuration round robin for elasticseach

I have some questions about the output configuration of winlogbeat.

Today, I have this for a windows server :

hosts: ["serverA:9200"]

I want to add another server in case of failure of serverA :

hosts: ["serverA:9200", "serverB:9200"]

According to the doc, it's a round robin mode, so if serverA is up, all datas will be sent to him, and if he's down, the data are sent to serverB.

But if serverA and serverB are in the same cluster, is it working ? No problem with that ?
Today, my configuration is serverA as a central point to host the data (it's a data node) and serverB is another data node in the same cluster.

So the data are replicated in this way : serverA => serverB

If server A is down, is serverB ships data to serverA when this one is up again ?

Thank you for your answers.

That means it will send to one serverA and then serverB, then repeat.

As long as you have replicas and as long as they form a cluster again, yes.

Thank you for your answer.

It works very well, I have done a test by stopping elasticsearch on serverA, the health of the cluster is yellow but it's normal, and I see the data shipped to serverB.

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