depends on your job role but for my job we have 1 project that’s not containerized and each time we have issues with it I want to crush my fucking balls
containerization is incredibly wasteful but it does solve some problems
like flatpak. when you don’t build all your containers on the same base image and shared layers, then you’ll store lots of slightly different versions of the same libraries and other files, both on disk, and then in memory
but containerization does not use VMs. containers share the same kernel, but userspace and resources are separated with namespaces. it has a very little overhead
depends on your job role but for my job we have 1 project that’s not containerized and each time we have issues with it I want to crush my fucking balls
containerization is incredibly wasteful but it does solve some problems
How is it wasteful?
like flatpak. when you don’t build all your containers on the same base image and shared layers, then you’ll store lots of slightly different versions of the same libraries and other files, both on disk, and then in memory
I see. I was more thinking in terms of CPU/RAM resources where it’s far cheaper to just run a single process instead of a VM for it, etc.
but containerization does not use VMs. containers share the same kernel, but userspace and resources are separated with namespaces. it has a very little overhead