How do I avoid split brain in dual zone on 7.5?

Elastic Stack Elasticsearch
1

I have five servers in zone A and zone B each.

So I can configure 3 master nodes in zone A and 2 master nodes in zone B
a cluster.

Or I can use a cluster of 3 master nodes in zone A and zone B, and set up a tribe node to connect the two cluster

Which configuration is better for HA and will not split brain?

What happens if the connection between the two zone is interrupted.

2

You cannot achieve true HA with only 2 zones.

If Zone B loses connectivity with Zone A, it cannot tell if

  1. that means that the network failed and Zone A is actually still operating, in which case Zone B should stand down, or
  2. that all of Zone A failed, and Zone B should take over control of the cluster.

There is no possible safe configuration because there is no way for the software to distinguish between the two scenarios.

For this to work you need a 3rd zone with at least 1 node it in to act as a tie-breaker. The safe option is 2 nodes in Zone A, 2 nodes in Zone B, 1 node in Zone C. Then, as long as 2 nodes zones are available and can communicate with one another, they will have a majority and will control the cluster.

3

thank you for your help.

Does it mean there is a possibility of split brain? Or is it just that the cluster does not work?

I have another question. If I arrange the same number of data nodes in zone A and zone B

What happens when A and B lose connection?

My English isn't very good, please don't mind.
Thank you again for your time.

4

The cluster may stop working until the connection is restored, but will not lose data. Some people call this "split brain". The manual covers how to set up your cluster for high availability in considerable depth.

The one with the majority of the master-eligible nodes will carry on working, the other side will not. If neither side has a majority of the master-eligible nodes then they may both stop working.

5

I'm very grateful

It's too hard...

and I can ask again?

I have five servers in A and B, a total of ten.

Each configuration is as follows :

RAM 512GB
SSD 800G *10

So I want to set up a data node on each hard drive.

What can i do??

My original plan is as follows :

Zone A :

3 master node in 3 server (System Disk - SSD)
50 data node in 5 server (Data Disk - SSD)
kibana *1

Zone B:
2 master node in 2 server (System Disk - SSD)
50 data node in 5 server (Data Disk - SSD)
Kibana *1

and find another server to act as a client node (master node:false / data node:false)

But the above plan doesn't seem to work, can you advise me?

My English isn't very good, please don't mind.
Thank you again for your time.

6

Sounds fine to me, although running so many nodes on a single server is pretty risky as you will lose them all when the server breaks. Better to have more/smaller servers.

7

Server already bought .. T_T

Our server room experience is more often that a single or multiple hard drives are damaged

That’s why I have the above configuration ideas

So if it is to reduce the number of nodes, what can I do?

Set two or more hard disk paths in the path.data of the same node ?

But a single servo memory has 512GB.

A node can only use up to 32GB.

Will it be too wasteful of memory?

8

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