Why containers are free and easy but managing them is hard and costly
-
I believe currently the containers management/orchestration market (the only way to run in Production) has a different status than containers, sure you can easily run docker anywhere, but you can't say the same on k8s. I rarely see a production k8s cluster managed by a person/team and not handing that over to a could provider, the same can not be said to ESXi/KVM/Hyper-V.
The only propper truly open implementation for containers I really liked was Canonical LXD, really no proprietary format what so ever you actually export the container as tar.gz image, and management system is baked in using advanced filesystem like ZFS. however sadly due to politics, it is not being implemented widely and I think RHEL/Centos actually dropped library from their repository to stop allowing successful and easy install of LXD on their OS.
I feel everyone jumped on the wagon and screwed things up, or they are attracting you to only to screw you over, cause the prices of k8s cluster is huge at the moment, which begs the question you can pay the below for 3 node k8s cluster from Google/Amazon/Azure
$2,052 per month ($1,539 with discounts) $2,880 per month $2,880 per month
source:
https://www.nirmata.com/2019/03/20/comparing-cloud-hosted-kubernetes-providers-google-gke-vs-microsoft-aks-vs-amazon-eks/So the cheapest option for a 3 node is 2000 $ (disregarding other costs ), well I get the autoscaling part and everything but for that amount of money, you can get 80 of those distributed VPS at many countries, like who needs auto-scaling if you have 80 VMs world wide.
128 GB SSD
2 CPU
4 GBWhy do you need a cow when you get milk for free (alot cheaper). seems like they want to allow developers to make decisions
-
You can use something like Ansible Container to manage them.
https://docs.ansible.com/ansible-container/For those that uses Proxmox virtualization, it provides an easy web interface to manage containers.
-
You're comparing apples to oranges. Hypervisors vs k8s makes no sense. And what makes you think that running containers in production is only viable with tools like k8s? That's like saying running vm in production only makes sense on openstack. Docker for example can easily be managed with Ansible, or if you want web tool, Portainer, including managing swarm clusters.
-
@marcinozga said in Why containers are free and easy but managing them is hard and costly:
You're comparing apples to oranges. Hypervisors vs k8s makes no sense. And what makes you think that running containers in production is only viable with tools like k8s? That's like saying running vm in production only makes sense on openstack. Docker for example can easily be managed with Ansible, or if you want web tool, Portainer, including managing swarm clusters.
True docker swarm is exception, but it gets overshadowed by k8s and openshift usually and it lacks auto scaling, and not really apple to oranges cause both methods will host your software
-
@Emad-R said in Why containers are free and easy but managing them is hard and costly:
I think RHEL/Centos actually dropped library from their repository to stop allowing successful and easy install of LXD on their OS.
We run it there and haven't noticed it dropped yet.
-
@Emad-R said in Why containers are free and easy but managing them is hard and costly:
I believe currently the containers management/orchestration market (the only way to run in Production)
While K8S and such are great tech, I wouldn't say that they are the only way to run in production.
-
@marcinozga said in Why containers are free and easy but managing them is hard and costly:
You're comparing apples to oranges. Hypervisors vs k8s makes no sense. And what makes you think that running containers in production is only viable with tools like k8s? That's like saying running vm in production only makes sense on openstack. Docker for example can easily be managed with Ansible, or if you want web tool, Portainer, including managing swarm clusters.
Nomad is another popular one that does not only containers but a few different types of jobs.
