Disaster Recovery with Galera Cluster 2021 Edition webinar wrap-up

We might have had one of our largest, most well attended webinars in recent times, for both EMEA and North American timezones, and we thank you for attending. We had many good questions, and we realise that this is a popular topic, so expect to get more blog posts around this. It also helps us drive the development of Galera Manager further, to know what you want (a new release, is just around the corner).

The video recording  of the webinar is available and we encourage you to take a gander.

We did demo a 9-node Galera Cluster running across San Francisco, London and Singapore, and also discussed latency for local nodes, a local Galera Cluster, as well as what penalty you might pay for running across a large wide area network (WAN).

We know that you would like to hear more about proxies; and it is clearly time for us to also evaluate third party solutions that do talk about us, from time to time.

We realised that we have to remind all why we focus on 3 data centers and not 2, and would refer you to quorum components. We do cover two node clusters, but we don’t recommend this as a solution overall.

A good followup is also likely showing how to have Galera Cluster with asynchronous replication across two data centres (this is quite easy to setup, once you turn on binary logging, but we will be thorough with some documentation).

When did streaming replication come along? With Galera 4 (so, in the current MySQL 8 release).

It is also clear while we offer options that are dangerous (like ignoring split brain checking via pc.ignore_sb or even ignoring quorum calculations via pc.ignore_quorum), you have to know when it is right to use these options, and good defaults, truly do matter.

Understanding segments (gmcast.segment) is also something we should further document via blogs. We also had good questions around the Galera Arbitrator Daemon (garbd) and we should focus on some practical use cases for this as well.

All in all, very popular topic, people are interested in disaster recovery, and High Availability, so expect a lot more posts on this topic. Thank you for attending the webinar! Don’t forget to watch the video recording. And if you have more questions, please do not hesitate to drop us a note so we can have them answered.