The number of uses and posts related to Docker is huge. The problem is the clutter is also big so finding the relevant posts is hard. Bellow is a selection of posts that got my attention.
First I have to mention the official Docker blog where the curated list of posts about docker are posted Curated. If you look there you can find some of my posts :). Lots of traffic on this blog is still originating from them.
Next is a post that does a retrospective of 2017 April Docker events
A very short and relevant post about the new multi stage Docker builds feature Multi-Stage Docker Builds for Creating Tiny Go Images. The post introduces us to the multi stage builds where you construct a “compile image” to construct your application and a final “production” image that contains only the minimal libraries needed by the final application.
Next is an amazing practical guide of using Docker Swarm in production Tips for using Docker Swarm mode in production. The author lists the issues he encountered in using Docker Swarm in a real environment. I have to mention that I had some of the same issues when using Docker Swarm but this is a very nice list and I found new things I never knew. Check out the other posts of this guy he knows what he is talking about.
The next posts describe real production environments that use the full power of Docker Swarm. MySQL on Docker: ClusterControl and Galera Cluster on Docker Swarm. This is a very nice HA deploy of a MySQL Galera Cluster over Docker Swarm. It is so easy to deploy something that was so complicated that I predict that Docker will make more and more people use the true power of HA application environments. This will no longer be restricted to big enterprise application environments, average Joe will do it for his home made application.
In the same context of HA the next post Using GlusterFS with Docker swarm cluster talks about using GlusterFS (a scale-out network attached storage file system) to replicate volumes across docker swarm nodes. This will make even the containers that need persistent data truly HA.
Nest based on the above someone even implemented a HA WordPress deployment using Docker Swarm and GlusterFS Architecting a Highly Available and Scalable WordPress Using Docker Swarm, Traefik & GlusterFS [Part I]
If you feel that GlusterFS is not OK for you and maybe you want also geographically distributed HA then LinBit to the rescue. Persistent and Replicated Docker Volumes with DRBD9 and DRBD Manage
Wow so many interesting things you can do this days without spending one month in trying to find the right combination of cluster and data replication configurations.
My idea from the above is to combine them in an experimental setup:
Geographical distributed WordPress deployment using:
– Docker Swarm across main site and remote site servers
– MySQL Galera Cluster on Docker Swarm on main site
– Active-Active WordPress deployment on main site swarm nodes with load-balancing
– Asynchronous DRBD replication of MySQLdata and WordPress data to a remote site
– GlusterFS for main site WordPress files to have the same data across all nodes
– Finally find an automatic way to cope with failure ideally across main and remote site without any manual intervention.