MySQL Forums
Forum List  »  NDB clusters

Re: Number of machines required for cluster redundancy
Posted by: Jeremy Smyth
Date: March 05, 2015 02:09AM

To avoid a split brain (two functioning cluster partitions) if the network between the machines goes down, only one of the management nodes acts as arbitrator.

If the machine hosting the arbitrator fails, then the remaining data node cannot reach the arbitrator and so it shuts down.

A better solution from a HA perspective would be to have three machines: one for each data node, and a third for the management node, and then you can distribute API nodes on those machines.

Options: ReplyQuote


Subject
Views
Written By
Posted
Re: Number of machines required for cluster redundancy
813
March 05, 2015 02:09AM


Sorry, you can't reply to this topic. It has been closed.

Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.