Subj : Projects
To   : deon
From : TassieBob
Date : Sun Jul 10 2022 21:42:50

 de> stripped to the bare bones that they only run a kernel and the docker
 de> engine.

I really should look at doing this myself.

 de> I also run multiple hosts, and mostly use "docker swarm" - so that my
 de> running docker containers float between any of the hosts.

At $dayjob we're using docker swarm for a good few production things.  I look 
after a couple of custom network automation tools and test locally on my 
machine, do a test docker build locally, and then push into GitLab and let the 
CI/CD to the push to staging.

 de> Swarm is great, I can take a host offline - say to do OS / docker
 de> runtime updates, and the containers float to another system, and then
 de> when it comes back online it'll take the next workload thats deployed
 de> (or the workload from another host when I update it).

What do you do for persistent storage?  Just an NFS mount from common storage 
onto each host machine?  The $dayjob containers have no local persistent data - 


















































































































































































































































































































































































































anything that's persisted it in a SQL database.  I think they might be using 
GlusterFS for containers that need it, but none of mine do.


--- GoldED+/LNX 1.1.5-b20220504
 * Origin: TassieBob's BBS (1337:2/106)

.