Open Source RDBMS - Seamless, Scalable, Stable and Free

한국어 | Login |Register

(click on this box to dismiss)

Does Cubrid support Master:Master replication

In my use case I require every node to be a master node, accept reads and writes is this possible with cubrid?

As it says on

"Reduced hardware costs with the configuration of multiple master database nodes and one or more slave database nodes."



link comment (0) accepted answer
asked 5 years ago
50 Answers

CUBRID HA can not support it.

CUBRID can only support master-slave and one master DB can have more than one slave node.

Only read operation can be executed on slave node.

link comment (0)
answered 5 years ago

Hello Duarte,

Please refer to CUBRID HA material we have presented at OSCON 2011 conference. There you can see a transition to "CUBRID HA & Auto Fail-over" section.

Regarding, "... multiple master database nodes..." you can actually create several master nodes in your system each of which can have multiple slave nodes. But such architecture requires each master to be independent (not synced with each other). In your application logic you can talk to all master nodes simultaneously through a single Broker (CUBRID middleware). However, each master node should have an independent role.

For example, Master1 is responsible for email server which is used in your app, while M2 is serving the Web pages data rendered by the same app, M3 is serving the data for subdomain site, etc. Each master, of course, has a list of slave (and/or replica) nodes between which HA is configured in sync/async/semi-sync mode. Fail-over will work seamlessly for each group of master-slave 1:N nodes.

Please see the "Extended - M:S=1:N" architecture displayed in the same presentation.

Let us know if you have other questions.

link comment (0)
answered 5 years ago

You are either using a very old browser or a browser that is not supported.
In order to browse you need to have one of the following browsers:

Internet Explorer: Mozilla Firefox: Google Chrome: