I would say that there is just a little overhead as Postgres data volume is mounted inside Docker container. Running Postgres in docker container without volumes would have a bigger impact.
And Docker container is not a VM, so I don't like to use word Virtualization for containers :)
If you think that this is just new cool thing, you probably didn't read a lot about it. Can you run 20 WordPress isolated installations in 30 seconds on traditional way (20 MySQL servers and 20 Apache web servers on one host)? Or just commit changes you made on your installation push it to docker index, pull it on another host and run that? Without db and app backup and restore procedure and you have exactly the same environment on both hosts. So you don't care about MySQL, Apache all libraries on another host, their versions, etc. You are moving software, in this case WordPress with your environment. With docker environment and software are in one piece. It is all about that.
Maybe WordPress just isn't great example, but this is extremely useful for development (works locally) and for continuous integration process. Every time you install software on fresh environment, so you don't need to manage it.
Correct, Ceph is much larger than the file system. As I said at beginning of article I'm using each Ceph component, RBD, RadosGW and CephFS and I will write article for each of that. This is just some sort of getting started guide. I'm interested at running Hadoop on top of CephFS, is it stable enough?
Ok, if both load balancers go away it will crash. When master or backup is up, both webservers in background can take all load (100%), because just one loadbalancer will work in one time.
Basically this will not increase performance, just availability as the main problem with load balancing is that it becomes a SPOF, if it goes down everything is down no matter how powerful web servers you have in background. What will happen if one of those gets huge amount of traffic... I think that it will not go to the backup as CARP should work. But yes it would be nice to test that.
I didn't try keepalived on Linux (I tried it on FreeBSD) so I'm not right person to say is it better or not. Only I can say is that CARP performed really well from my perspective. I'm also interested in comparison of those two...
And Docker container is not a VM, so I don't like to use word Virtualization for containers :)