vault backup: 2023-08-26 11:34:11

software/Kubernetes.md

@@ -0,0 +1,113 @@
+# Vocabulary
+https://kubernetes.io/docs/tutorials/kubernetes-basics/explore/explore-intro/ - Good summary of the different parts
+## Cluster
+The largest unit, a kubernetes cluster is a collection of one or more worker machines (virtual or physical machines), and the control plane. 
+
+## Control plane
+A control plane is in charge of managing worker nodes and pods in a cluster. 
+
+## Node
+[A node](https://kubernetes.io/docs/tutorials/kubernetes-basics/explore/explore-intro/) is a virtual or physical machine that in Kubernetes, and is managed by the control plane. A Node can have multiple pods, and the control plane is in charge of scheduling pods across nodes.
+
+Each node runs at least:
+- A Kubelet, the process responsible for handling communication between the control plane and the node. It manages the pods and containers running on a machine
+- A [container runtime](https://kubernetes.io/docs/setup/production-environment/container-runtimes/) responsible for pulling the container image from a registry, unpacking the container, and running the application. Examples include containerd and docker.
+
+## Deployment
+A [deployment](https://kubernetes.io/docs/tutorials/kubernetes-basics/deploy-app/deploy-intro/) instructs kubernetes how to create and update instances of an application. It can be imagined similar to a blueprint. The term 'deployment' may also refer to the group of pods and replicas defined by one or more configuration files. 
+
+## Pod
+A [pod](https://kubernetes.io/docs/tutorials/kubernetes-basics/explore/explore-intro/) is an abstraction that represents a group of one or more application containers, and shared resources. Some shared include storage and a private network. They are meant to represent a logical host, and a rough parallel might be a physical machine running a service, although this machine may have multiple network adapters. A deployment creates pods, and not containers. 
+
+## ReplicaSet
+A [ReplicaSet](https://kubernetes.io/docs/concepts/workloads/controllers/replicaset/), often referred to as a pod replica, is an identical copy of pod. They can be used for many things, most commonly high availability and load balancing.
+
+## Service
+A [service](https://kubernetes.io/docs/concepts/services-networking/service/) in Kubernetes is is an abstraction that defines a logical set of Pods and a policy to access them. A Service is defined using YAML or JSON, and is primarily used to expose a network application that's running on one or more Pods in a cluster. They also route traffic across a set of Pods, allowing Pods to die and replicate without impacting the application.
+
+## `kubectl`
+`kubectl` is the Kubernetes command line tool used to run commands against clusters. You can use `kubectl` to deploy applications, inspect and manage cluster resources, and view logs. The general format of the command follows `kubectl [action] [resource]`
+
+## `minikube`
+`minikube` creates a local Kubernetes cluster for development. It can be configured to deploy on bare metal, containers (like docker), or a VM (QEMU).
+
+# Commands
+## Common commands
+-   **kubectl get** - list resources
+-   **kubectl describe** - show detailed information about a resource
+-   **kubectl logs** - print the logs from a container in a pod
+-   **kubectl exec** - execute a command on a container in a pod
+## Start `minikube`
+```
+minikube start
+```
+
+## Stop `minikube`
+```
+minikube stop
+```
+
+## Expose services running in minikube
+```
+minikube services
+```
+
+## View the dashboard for `minikube`(run in new terminal window)
+```
+minikube dashboard
+```
+
+## Check that `kubectl` is configured to talk to the cluster
+```
+kubectl version
+```
+
+## View nodes in the server
+```
+kubectl get nodes
+```
+
+## List deployments
+```
+kubectl get deployments
+```
+
+## Start a proxy to access the private network
+```
+kubectl proxy
+```
+
+## Access the API of a pod (through the proxy)
+```
+curl http://localhost:8001/api/v1/namespaces/default/pods/$POD_NAME/
+```
+Where `$POD_NAME` is the name of a pod, you can get this information with `kubectl get pods`
+
+## Start a Bash session in a Pod's container
+```
+kubectl exec --ti $POD_NAME -- bash
+```
+
+## Create a new Service and expose it to external traffic
+```
+kubectl expose deployment/kubernetes-bootcamp --type="NodePort" --port 8080
+```
+
+## Apply a new label to pods
+```
+kubectl label pods "$POD_NAME" version=v1
+```
+
+## See if a Service is load-balancing traffic
+```
+kubectl describe services/$SERVICE_NAME
+```
+
+## Set a different desired amount of replicas for a deployment
+```
+kubectl scale deployments/$DEPLOYMENT_NAME --replicas=$DESIRED_NUM_OF_REPLICAS
+```
+
+# Tutorials
+https://kubernetes.io/docs/tutorials/hello-minikube/
+https://kubernetes.io/docs/tutorials/kubernetes-basics/

software/Rewriting TS.md

@@ -0,0 +1,15 @@
+
+TS is currently written in Python, depending on a technically EOL version of the discord.py library. It has grown to a massive scale, and provides all of the functionality that is critical for the server, but has some downsides. 
+- The codebase is not documented very well, which makes changes to the core system difficult.
+- Python suffers from performance issues that have proven be meaningful at the scale of the server (they have been addressed)
+- The migration to a new discord.py version is proving to be extremely troublesome
+
+However, sticking with the system we currently definitely offers some pros:
+- It functions and is largely bug free, proving stable enough for our use case, with the reliability proving to be mostly dependent on the host.
+- Eff is contributing, however the level at which he will end up contributing to the project is uncertain
+- Once totally 'cleaned up', sticking with TS may be perfectly acceptable
+
+Rewriting TS offers some pros:
+- We can document it as we go and decide on design patterns and practices to stick to before implementation begins
+- We evaluate the scale that TS needs to fit, and design according to that
+- We can switch to a language that may prove to be more performant/versatile, should we decide discord.py and python are not the best tool for our needs