Basic setup now passing initial checks

Good poc of how to get game server
connection info that we will provide to a bot/user

infra/dns/Makefile

@@ -1,24 +0,0 @@
-plan=out.plan
-
-SHELL := /bin/bash
-
-$(plan): *.tf
-	source ../secrets/set-env.sh && terraform plan -input=false -out $(plan)
-
-push: build
-	source ../secrets/set-env.sh && terraform apply $(plan)
-
-refresh:
-	source ../secrets/set-env.sh && terraform apply -refresh-only
-
-test:
-	terraform validate
-
-
-rip:
-	source ../secrets/set-env.sh && terraform destroy
-
-clean:
-	rm -f $(plan)
-
-.PHONY: test build clean push rip

infra/dns/build.sh

@@ -0,0 +1,34 @@
+#!/bin/bash
+
+set -e
+
+opt=$1
+plan=tfplan
+
+build_plan() {
+	echo Generating plan
+	set -x
+	terraform plan -var-file variables.tfvars -input=false -out $plan
+}
+
+deploy_plan() {
+	terraform apply $plan
+}
+
+init() {
+	terraform init
+}
+
+help_prompt() {
+	cat <<- EOF
+	Options: plan deploy help
+	EOF
+}
+
+# Default to building a plan
+source ./secrets.sh
+case $opt in 
+	plan) build_plan;;
+	deploy) deploy_plan;;
+	*) help_prompt;;
+esac

infra/dns/project-athens-xyz.tf

@@ -1,49 +0,0 @@
-#############################
-# project-athens.xyz DNS ZONE
-#############################
-
-# This entry is just for the sample service that is just plain nginx
-# No TLS will be placed on this just yet as we need to make sure this
-# and the load balancer are setup to receive things properly
-resource "aws_route53_zone" "project-athens" {
-  name    = "project-athens.xyz"
-  comment = "Project Athens domain zone"
-}
-
-
-locals {
-  project-athens-records = [
-    {
-      name    = "project-athens.xyz"
-      type    = "NS"
-      ttl     = 172800
-      records = [
-        "ns-806.awsdns-36.net.",
-        "ns-1881.awsdns-43.co.uk.",
-        "ns-1109.awsdns-10.org.",
-        "ns-11.awsdns-01.com.",
-      ]
-    },
-    {
-      name    = "project-athens.xyz"
-      type    = "SOA"
-      ttl     = 900
-      records = [
-        "ns-806.awsdns-36.net. awsdns-hostmaster.amazon.com. 1 7200 900 1209600 86400"
-      ]
-    }
-  ]
-}
-
-resource "aws_route53_record" "project-athens-record" {
-  for_each = {
-    for index, record in local.project-athens-records:
-      index => record
-  }
-  zone_id  = aws_route53_zone.project-athens.id
-  name     = each.value.name
-  type     = lookup(each.value, "type", "A")
-  ttl      = lookup(each.value, "ttl", 300)
-  records  = each.value.records
-}
-

infra/dns/shockrah-xyz.tf

@@ -36,8 +36,7 @@ locals {
     },
     { name = "www.shockrah.xyz",      records = [ var.vultr_host ] },
     { name = "resume.shockrah.xyz",   records = [ var.vultr_host ] },
-    { name = "immich.shockrah.xyz",    records = [ "45.32.92.196" ] },
-    { name = "git.shockrah.xyz",    records = [ var.vultr_host ] },
+    { name = "git.shockrah.xyz",      records = [ var.vultr_host ] },
   ]
 }
 

infra/dns/variables.tfvars

@@ -0,0 +1 @@
+vultr_host = "45.32.83.83"

infra/static-vultr/build.sh

@@ -0,0 +1,36 @@
+#!/bin/bash
+
+set -e
+
+opt=$1
+plan=tfplan
+
+build_plan() {
+	echo Generating plan
+	set -x
+	terraform plan -var-file variables.tfvars -input=false -out $plan
+}
+
+deploy_plan() {
+	terraform apply $plan
+}
+
+init() {
+	terraform init
+}
+
+help_prompt() {
+	cat <<- EOF
+	Options: plan deploy help
+	EOF
+}
+
+# Default to building a plan
+source ./secrets.sh
+case $opt in 
+	plan) build_plan;;
+	deploy) deploy_plan;;
+	init) init;;
+	*) help_prompt;;
+esac
+

infra/static-vultr/immich-server.tf

@@ -1,22 +0,0 @@
-# Here we define the resources for our temporary immich server
-
-resource vultr_instance immich {
-  plan    = var.host.plan
-  region  = var.host.region
-  os_id   = var.host.os
-  enable_ipv6 = true
-
-  # Enable backups for now since we're getting off of s3 as well at some point
-  backups = "enabled"
-  backups_schedule {
-    type = "weekly"
-    dow  = var.host.backups.day
-    hour = var.host.backups.hour
-  }
-
-  ssh_key_ids = [ vultr_ssh_key.immich.id ]
-  firewall_group_id = vultr_firewall_group.host.id
-  label = "Immich Server"
-}
-
-

infra/static-vultr/output.tf

@@ -14,10 +14,3 @@ output vultr_key_id {
 }
 
 
-output immich_key {
-  sensitive = true
-  # value     = tls_private_key.host.private_key_openssh
-  value     = vultr_instance.immich.default_password
-}
-
-

infra/static-vultr/ssh.tf

@@ -5,18 +5,6 @@ resource tls_private_key host {
 
 resource vultr_ssh_key host {
   name = "static_ssh_key"
-  ssh_key = tls_private_key.host.public_key_openssh
+  ssh_key = chomp(tls_private_key.host.public_key_openssh)
 }
 
-####################
-#    Immich keys   #
-####################
-resource tls_private_key immich {
-  algorithm = "RSA"
-  rsa_bits  = 4096
-}
-
-resource vultr_ssh_key immich {
-  name    = "static_ssh_key"
-  ssh_key = tls_private_key.immich.public_key_openssh
-}

infra/vultr-kubernetes/admin-services.tf

@@ -0,0 +1,62 @@
+resource kubernetes_namespace admin-servers {
+    count = length(var.admin_services.configs) > 0 ? 1 : 0
+    metadata {
+        name = var.admin_services.namespace
+    }
+}
+
+resource kubernetes_pod admin {
+    for_each = var.admin_services.configs
+
+    metadata {
+        name = each.key
+        namespace = var.admin_services.namespace
+        labels = {
+            app = each.key
+        }
+    }
+    spec {
+        node_selector = {
+            NodeType = var.admin_services.namespace
+        }
+        container {
+            image = each.value.image
+            name  = coalesce(each.value.name, each.key)
+            resources {
+                limits = {
+                    cpu    = each.value.cpu
+                    memory = each.value.mem
+                }
+            }
+            port {
+                container_port = each.value.port.internal
+                protocol       = coalesce(each.value.proto, "TCP")
+            }
+        }
+    }
+}
+
+resource kubernetes_service admin {
+    for_each = var.admin_services.configs
+    metadata {
+        name = each.key
+        namespace = var.admin_services.namespace
+        labels = {
+            app = each.key
+        }
+    }
+    # TODO: don't make these NodePorts since we're gonna want them
+    # to be purely internal to the Cluster.
+    # WHY? Because we want to keep dashboards as unexposed as possible
+    spec {
+        selector = {
+            app = each.key
+        }
+        port {
+            target_port        = each.value.port.internal
+            port               = each.value.port.expose
+        }
+        type = "NodePort"
+    }
+}
+

infra/vultr-kubernetes/backend.tf

@@ -13,7 +13,11 @@ terraform {
     }
     vultr = {
       source = "vultr/vultr"
-      version = "2.19.0"
+      version = "2.22.1"
+    }
+    kubernetes = {
+      source  = "hashicorp/kubernetes"
+      version = "2.34.0"
     }
   }
 }
@@ -31,4 +35,8 @@ provider aws {
   max_retries = 1
 }
 
+provider kubernetes {
+  config_path = "config.yaml"
+}
+
 

infra/vultr-kubernetes/build.sh

@@ -15,6 +15,10 @@ deploy_plan() {
 	terraform apply $plan
 }
 
+init() {
+	terraform init
+}
+
 help_prompt() {
 	cat <<- EOF
 	Options: plan deploy help

infra/vultr-kubernetes/cluster.tf

@@ -2,17 +2,29 @@ resource vultr_kubernetes athens {
   region = var.cluster.region
   version = var.cluster.version
   label = var.cluster.label
-  enable_firewall = true
+  # BUG: only have this set when creating the resource for the first time
+  # once the cluster is up, we should comment this out again
+  # enable_firewall = true
   node_pools {
-    # how many nodes do we want in this pool
     node_quantity = 1
-    plan = var.cluster.pool.plan
-    label = var.cluster.label
-    min_nodes = var.cluster.pool.min
-    max_nodes = var.cluster.pool.max
+    plan      = var.cluster.pools["meta"].plan
+    label     = var.admin_services.namespace
+    min_nodes = var.cluster.pools["meta"].min
+    max_nodes = var.cluster.pools["meta"].max
+    # tag       = var.admin_services.namespace
   }
 }
 
+resource vultr_kubernetes_node_pools games {
+  cluster_id = vultr_kubernetes.athens.id
+  node_quantity = var.cluster.pools["games"].min
+  plan      = var.cluster.pools["games"].plan
+  label     = var.game_servers.namespace
+  min_nodes = var.cluster.pools["games"].min
+  max_nodes = var.cluster.pools["games"].max
+  tag       = var.admin_services.namespace
+}
+
 output k8s_config {
   value = vultr_kubernetes.athens.kube_config
   sensitive = true

infra/vultr-kubernetes/dev/.gitignore

@@ -0,0 +1,4 @@
+# created by virtualenv automatically
+bin/
+lib/
+

infra/vultr-kubernetes/dev/find-server.py

@@ -0,0 +1,51 @@
+from argparse import ArgumentParser
+from argparse import Namespace
+from kubernetes import client, config
+import re
+
+def get_args() -> Namespace:
+    parser = ArgumentParser(
+        prog="Cluster Search Thing",
+        description="General utility for finding resources for game server bot"
+    )
+    games = {"reflex", "minecraft"}
+    parser.add_argument('-g', '--game', required=False, choices=games)
+
+    admin = {"health"}
+    parser.add_argument('-a', '--admin', required=False, choices=admin)
+    return parser.parse_args()
+
+def k8s_api(config_path: str) -> client.api.core_v1_api.CoreV1Api:
+    config.load_kube_config("../config.yaml")
+    return client.CoreV1Api()
+
+def get_admin_service_details(args: ArgumentParser, api: client.api.core_v1_api.CoreV1Api):
+    print('admin thing requested', args.admin)
+
+def get_game_server_ip(args: ArgumentParser, api: client.api.core_v1_api.CoreV1Api):
+    pods = api.list_pod_for_all_namespaces(label_selector=f'app={args.game}')
+    node_name = pods.items[0].spec.node_name
+
+    services = api.list_service_for_all_namespaces(label_selector=f'app={args.game}')
+    port = services.items[0].spec.ports[0].port
+
+    # Collecting the IPV4 of the node that contains the pod(container)
+    # we actually care about. Since these pods only have 1 container
+    # Now we collect specific data about the game server we requested
+    node_ips  = list(filter(lambda a: a.type == 'ExternalIP', api.list_node().items[0].status.addresses))
+    ipv4 = list(filter(lambda item: not re.match('[\d\.]{3}\d', item.address), node_ips))[0].address
+    ipv6 = list(filter(lambda item: re.match('[\d\.]{3}\d', item.address), node_ips))[0].address
+    
+    print(f'{args.game} --> {ipv4}:{port} ~~> {ipv6}:{port}')
+
+
+if __name__ == '__main__':
+    args = get_args()
+    api = k8s_api('../config.yaml')
+
+    if args.game:
+        get_game_server_ip(args, api)
+    
+    if args.admin:
+        get_admin_service_details(args, api)
+    

infra/vultr-kubernetes/dev/pyvenv.cfg

@@ -0,0 +1,8 @@
+home = /usr
+implementation = CPython
+version_info = 3.10.12.final.0
+virtualenv = 20.13.0+ds
+include-system-site-packages = false
+base-prefix = /usr
+base-exec-prefix = /usr
+base-executable = /usr/bin/python3

infra/vultr-kubernetes/dev/requirements.txt

@@ -0,0 +1,18 @@
+cachetools==5.5.0
+certifi==2024.8.30
+charset-normalizer==3.4.0
+durationpy==0.9
+google-auth==2.36.0
+idna==3.10
+kubernetes==31.0.0
+oauthlib==3.2.2
+pyasn1==0.6.1
+pyasn1_modules==0.4.1
+python-dateutil==2.9.0.post0
+PyYAML==6.0.2
+requests==2.32.3
+requests-oauthlib==2.0.0
+rsa==4.9
+six==1.17.0
+urllib3==2.2.3
+websocket-client==1.8.0

infra/vultr-kubernetes/firewall.tf

@@ -1,11 +1,32 @@
-# Inbound rules for web traffic
 resource vultr_firewall_rule web_inbound {
   for_each = toset([for port in [80, 443, 6443] : tostring(port) ])
   firewall_group_id = vultr_kubernetes.athens.firewall_group_id
-  #firewall_group_id = vultr_firewall_group.cluster.id
   protocol = "tcp"
   ip_type = "v4"
   subnet = "0.0.0.0"
   subnet_size = 0
   port = each.value
 }
+
+resource vultr_firewall_rule game-server-inbound {
+  for_each = var.game_servers.configs
+  firewall_group_id = vultr_kubernetes.athens.firewall_group_id
+  protocol = "tcp"
+  ip_type = "v4"
+  subnet = "0.0.0.0"
+  subnet_size = 0
+  port = each.value.port.expose
+}
+
+
+resource vultr_firewall_rule admin-service-inbound {
+  for_each = var.admin_services.configs
+  firewall_group_id = vultr_kubernetes.athens.firewall_group_id
+  protocol = "tcp"
+  ip_type = "v4"
+  subnet = "0.0.0.0"
+  subnet_size = 0
+  notes = each.value.port.notes
+  port = each.value.port.expose
+}
+

infra/vultr-kubernetes/game-server.tf

@@ -0,0 +1,55 @@
+resource kubernetes_namespace game-servers {
+    count = length(var.game_servers.configs) > 0 ? 1 : 0
+    metadata {
+        name = var.game_servers.namespace
+    }
+}
+
+resource kubernetes_pod game {
+    for_each = var.game_servers.configs
+
+    metadata {
+        name = each.key
+        namespace = var.game_servers.namespace
+        labels = {
+            app = each.key
+        }
+    }
+    spec {
+        container {
+            image = each.value.image
+            name  = coalesce(each.value.name, each.key)
+            resources {
+                limits = {
+                    cpu    = each.value.cpu
+                    memory = each.value.mem
+                }
+            }
+            port {
+                container_port = each.value.port.internal
+                protocol       = coalesce(each.value.proto, "TCP")
+            }
+        }
+    }
+}
+
+resource kubernetes_service game {
+    for_each = var.game_servers.configs
+    metadata {
+        name = each.key
+        namespace = var.game_servers.namespace
+        labels = {
+            app = each.key
+        }
+    }
+    spec {
+        selector = {
+            app = each.key
+        }
+        port {
+            target_port        = each.value.port.internal
+            port   = each.value.port.expose
+        }
+        type = "NodePort"
+    }
+}

infra/vultr-kubernetes/variables.tf

@@ -25,20 +25,47 @@ variable cluster {
     region  = string
     label   = string
     version = string
-    pool = object({
+    pools = map(object({
       plan = string
       autoscale = bool
       min = number
       max = number
-    })
+    }))
   })
 }
 
-variable lab_domain {
-  type = string
+variable game_servers {
+  type = object({
+    namespace = string
+    configs = map(object({
+      name  = optional(string)
+      image = string
+      cpu   = string
+      mem   = string
+      port  = object({
+        internal = number
+        expose   = number
+      })
+      proto = optional(string)
+    }))
+  })
 }
 
-variable lb_ip4 {
-  type = string
+variable admin_services {
+  type = object({
+    namespace = string
+    configs = map(object({
+      name  = string
+      image = string
+      cpu   = string
+      mem   = string
+      port  = object({
+        notes = optional(string)
+        internal = number
+        expose   = number
+      })
+      proto = optional(string)
+    }))
+  })
 }
 

infra/vultr-kubernetes/variables.tfvars

@@ -1,18 +1,51 @@
-
 cluster = {
   region = "lax"
   label   = "athens-cluster"
-  version = "v1.30.0+1"
-  pool = {
-	  plan =  "vc2-1c-2gb"
-	  autoscale = true
-	  min = 1
-	  max = 2
+  version = "v1.31.2+1"
+  pools = {
+    meta = {
+      plan =  "vc2-1c-2gb"
+      autoscale = true
+      min = 1
+      max = 2
+    }
+    games = {
+      plan =  "vc2-1c-2gb"
+      autoscale = true
+      min = 1
+      max = 3
+    }
   }
 }
 
-lab_domain = "temprah-lab.xyz"
-
-lb_ip4 = "45.32.68.232"
-
+game_servers = {
+  namespace = "games"
+  configs = {
+    # minecraft = {
+    #   image = "itzg/minecraft-server"
+    #   cpu   = "1000m"
+    #   mem   = "2048Mi"
+    #   port  = {
+    #     expose = 30808
+    #     internal = 80
+    #   }
+    # }
+  }
+}
 
+admin_services = {
+  namespace = "admin-services"
+  configs = {
+    # health = {
+    #   image = "nginx:latest"
+    #   name  = "health"
+    #   cpu   = "200m"
+    #   mem   = "64Mi"
+    #   port  = {
+    #     notes    = "Basic nginx sanity check service"
+    #     expose   = 30800
+    #     internal = 80
+    #   }
+    # }
+  }
+}

playbooks/inventory.yaml

@@ -0,0 +1,3 @@
+static-web:
+  hosts:
+    shockrah.xyz:

playbooks/manual-prerequisites.yaml

@@ -0,0 +1,9 @@
+---
+- name: Pre Pipeline Playbook for Static Hosts
+  hosts: static-web
+  remote_user: root
+  tasks:
+    - name: Import manual setup steps
+      ansible.builtin.import_role:
+        name: static
+        tasks_from: setup-webadmin.yaml

playbooks/readme.md

@@ -0,0 +1,5 @@
+# What is this
+
+Here be the ansible based workflows that we use to keep things like the static
+hosts properly setup with all the resources they need to properly host the
+services we intended on hosting.

playbooks/static-setup.yaml

@@ -0,0 +1,8 @@
+---
+- name: Static Host Maintenance and Setup
+  hosts: static-web
+  remote_user: webadmin
+  tasks:
+    - name: Import static host role
+      ansible.builtin.import_role:
+        name: static

playbooks/static/handlers/main.yaml

@@ -0,0 +1,5 @@
+- name: Restart SSH
+  become: true
+  ansible.builtin.systemd:
+    name: sshd
+    state: restarted

playbooks/static/tasks/main.yaml

@@ -0,0 +1,40 @@
+# Things that we definitely want to have are the following
+# docker docker-compose python(latest) certbot
+- name: Uhhh yea
+  become: true
+  block:
+  - name: Install base dependencies
+    ansible.builtin.apt:
+      update_cache: true
+      pkg:
+        - ca-certificates
+        - curl
+        - lsb-release
+  - name: Setup keyring directory
+    ansible.builtin.command:
+      cmd: "install -m 0755 -d {{ static_keyring_dir }}"
+      creates: "{{ static_keyring_dir }}"
+  - name: Download the docker GPG key
+    ansible.builtin.get_url:
+      url: "{{ static_docker_ubuntu }}/gpg"
+      dest: "{{ static_keyring_dir }}/docker.asc"
+      mode: "0644"
+  - name: Ensure docker.lst is present
+    vars:
+      key_path: "{{ static_keyring_dir }}/docker.asc"
+      repo: "{{ static_docker_ubuntu }}"
+      os_codename: jammy
+    ansible.builtin.template:
+      src: docker.list
+      dest: "{{ static_apt_sources_dir }}/docker.list"
+      mode: "0644"
+  - name: Install docker and python packages
+    ansible.builtin.apt:
+      update_cache: true
+      pkg:
+        - docker-ce
+        - docker-ce-cli
+        - containerd.io
+        - docker-buildx-plugin
+        - docker-compose-plugin
+        - python3

playbooks/static/tasks/setup-webadmin.yaml

@@ -0,0 +1,43 @@
+- name: Ensure sudo is available
+  ansible.builtin.apt:
+    state: present
+    update_cache: true
+    pkg:
+      - sudo
+      - zsh
+- name: Create webadmin user
+  ansible.builtin.user:
+    name: webadmin
+    state: present
+    shell: /bin/zsh
+    groups:
+      - nginx
+    append: true
+- name: Copy webadmin public key
+  ansible.posix.authorized_key:
+    user: webadmin
+    state: present
+    key: "{{ lookup('file', 'files/webadmin.pem.pub') }}"
+- name: Add webadmin to sudoers
+  ansible.builtin.copy:
+    dest: "/etc/sudoers.d/webadmin"
+    content: "webadmin ALL=(ALL) NOPASSWD: ALL"
+    mode: "0644"
+    owner: root
+    group: root
+- name: Disable Password Authentication
+  ansible.builtin.lineinfile:
+    dest: /etc/ssh/sshd_config
+    line: PasswordAuthentication no
+    state: present
+    backup: true
+  notify:
+    - Restart SSH
+- name: Disable root login
+  ansible.builtin.lineinfile:
+    dest: /etc/ssh/sshd_config
+    line: PermitRootLogin no
+    state: present
+    backup: true
+  notify:
+    - Restart SSH

playbooks/static/templates/docker.list

@@ -0,0 +1 @@
+deb [arch=amd64 signed-by={{ key_path }}] {{ repo }} {{ os_codename }} stable

playbooks/static/vars/main.yaml

@@ -0,0 +1,4 @@
+static_keyring_dir: /etc/apt/keyrings
+static_docker_ubuntu: https://download.docker.com/linux/ubuntu
+static_apt_sources_dir: /etc/apt/sources.list.d
+static_codename: jammy

runbooks/atlas-setup.md

@@ -1,45 +0,0 @@
-# What this covers
-
-The creation of Atlas as it happened in order
-
-## Commands Ran
-
-Once the infra was provisioned and verified to be configured by Terraform correctly
-we move on to the following
-
-```sh
-# Setup the machine to run docker
-ansible-playbook -i hosts.ini atlas/init/system-deps.yml
-
-# Second we copy over the contents of Alpha's mounted docker volumes
-ansible-playbook -i hosts.ini atlas/init/perma-mount-drives.yml
-
-# Next we copy over the data that we want to migrate ( if any )
-ansible-playbook -i hosts.ini -e filebrowser=/path -e clippable=/path atlas/init/migrate-clips-files.yml
-
-# Setup the services on the host that we want to run
-ansible-playbook -i hosts.ini atlas/init/setup-containers.yml
-
-# Next we put up the reverse proxy (nginx)
-ansible-playbook -i hosts.ini atlas/init/setup-reverse-proxy.yml
-
-# Finally we add TLS on top of nginx and we're done
-ansible-playbook -i hosts.ini atlas/init/setup-certbot.yml
-```
-
-Maintenance should be straight forward for this machine as TLS is automatically
-renewed every 3 months by a cron job. We can manually update the certs however
-if we really want to. They also don't require anymore manual variable injection
-like Alpha did as the only thing protected was `dev@shockrah.xyz` which is at
-this point becoming semi-public. This means while it is associated with code
-it is more of a _business e-mail_ so it can be placed in this repository with
-very little concern.
-
-System updates are now also to be fetched with a:
-
-```sh
-ansible-playbook -i hosts.ini atlas/maintain/analyze-system-deps.yml
-```
-
-Which performs purely read operations and does not affect the state of the
-machine.

runbooks/new-drive.md

@@ -1,33 +0,0 @@
-# Mounting an attached drive
-
-Assumptions:
-
-* New drive is attached(in AWS) and detected in software
-Ideally attachment is made through terraform
-
-## Mounting Instructions (Step-by-Step)
-
-1. Verify data does not have data: `sudo file -s /dev/xvdf`
-
-Should return `data` if its ok. Other wise we're probably looking at the wrong
-drive.
-
-2. Create the filesystem on the new empty drive: `sudo mkfs -t ext4 /dev/xvdf`
-
-3. Create mountpoint other wares to actaully use the drive
-`sudo mkdir /mnt/example`.
-
-Change _example_ to something that actually makes sense.
-
-4. Add a new entry to /etc/fstab for automounting
-
-`/dev/xvdf       /newvolume   ext4    defaults,nofail        0       0`
-
-Tab delimited btw.
-
-5. Mount all drives listed in `/etc/fstab` from before. `sudo mount -a`
-
-
-
-
-