Use Directadmin on a Kubernetes environment

[auser]

Hello,

Has someone experience in using Directadmin on a Kubernetes environment? I could not find a lot of documentation about this.

I am informing myself regarding this topic because we need to upgrade our OS due to end of life soon, and are concidering to run an new environment, inside of Kubernetes, as this was the advice of my Sysadmin, regarding scalability. Only regarding the use of directadmin is still something we are looking into. He speaks about 'using a container within Kubernetes", i think he means with docker.

I try to have a good theoritical understanding of the bigger picture, as i am an IT guy, but no sysadmin, and also try to gather as much documentation as possible to fully understand the theoretical background, and that i can provide as much documenation as possible

Thanks for advising
Kind regards

 

[factor]

Directadmin on a Kubernetes environment?

its not natively supported.

i think he means with docker.

Kubernetes

Two different things close yes same no. Neither supported.

Here are the current supported OS

Overview | Directadmin Docs

 

[sysdev]

Well, you can't 'scale' a DirectAdmin setup in any cluster environment like docker, kubernetes or simply multiple servers, but... you might go down this road:

I presume you use kubernetes with docker.

1. create a docker DA instance and make sure you have some kind of shared storage from where you can mount the /home/ dir on your directadmin instance (docker, vm, whatever).
2. move your database to a different vm or use the docker-galera instances with permanent storage. This makes sure you can scale your database platform.
3. use multiple apache and different php-fpm docker instances. This makes sure you can scale http and php whenever needed.
4. use the da instance to handle account management, ftp and everything where you use the /usr/local/directadmin/scripts/custom hooks to update configs.

We've done this to handle a few (4-5) websites with over a billion requests per month. But this is pretty complex. If you only have any number of users with 'normal' sites, I would just get an extra vm/server somewhere and move users a bit around. Clusters are complex... and even though they are meant to make life easier, every server/service added, adds up to possible errors.
Before you know it YOU might end up as the SPOF yourself if you don't have at least two sysadmins/admins who know how stuf works