Purpose of the master node

Hi all,

The conf shows the following description of the master node: “The master will perform some periodical tasks that non-masters won’t perform.”

Does somebody know in more detail what this means? Is the master for instance the only one that is allowed to write in mongodb? And what happens if the master is not available for some time? (let’s say one hour)

@spix
Hello,

What conf are you talking about?

MongoDb holds all the metadata for Graylog. So MongoDb is normally configured in a “Replica Set”.

You can find more information as follow.
MongoDb a three-member replica set.

You might find your answer here.
ES Master Node

You’re question can be answered below.

Any master-eligible node that is not a voting-only node
may be elected to become the master node by the master election process.

That was quoted from here. Elected master node

Hope that helps

Hi @gsmith,

Thanks for the reply. The config I am talking about is this one: graylog2-server/graylog.conf at master · Graylog2/graylog2-server · GitHub

We have a MongoDB replicaset with multiple nodes and we have a Elastic Cluster, no I really talk about the role “master” looking at the graylog nodes. (marked with a “star” in the nodes overview)

@spix
Ok, so you mean this?

That tell you which one is the master node.

Correct! So back to my question :wink:

See Initial Configuration — Graylog 4.0.0 documentation. The master node is responsible for things like archiving and other maintenance tasks. If the master isn’t available, then you’ll still be able to receive messages, but any sort of index maintenance, rotation, archiving, etc. won’t happen.