Merge branch 'main' into feature/etcd-cluster-size-configuration

.github/workflows/terraform.yml

@@ -4,7 +4,7 @@ on:
   pull_request:
 
 permissions:
-  contents: read
+  contents: write
 
 concurrency:
   group: project-e00pjzzrtk1fs3yavy
@@ -22,10 +22,11 @@ jobs:
       max-parallel: 2
       matrix:
         solution:
+          - name: compute-testing
+          - name: slurm
           - name: k8s-inference
           - name: k8s-training
-          - name: slurm
-          - name: compute-testing
+          - name: wireguard
 
     defaults:
       run:
@@ -61,6 +62,9 @@ jobs:
         > ~/.s3cfg
         mkdir -p tests/reports
 
+    - name: Install XMLStarlet
+      run: sudo apt install -y xmlstarlet
+
     - name: Install Nebius CLI
       run: |
         curl -sSL https://storage.ai.nebius.cloud/nebius/install.sh | bash
@@ -107,18 +111,46 @@ jobs:
 
     # Run Terraform Tests
     - name: Terraform Test
-      run: terraform test -verbose -junit-xml=tests/reports/TEST-result-${{ github.run_id }}.xml
+      run: terraform test -junit-xml=tests/reports/TEST-result-${{ github.run_id }}.xml
 
-    - name: Test Summary
-      uses: test-summary/action@v2
-      with:
-        paths: ${{ matrix.solution.name }}/tests/reports/TEST-result-${{ github.run_id }}.xml
-      if: always()
+    - name: Set date in report
+      run: |
+        xmlstarlet ed \
+        --inplace \
+        -i '/testsuites' -t attr -n timestamp -v $(date --iso-8601=seconds) \
+        -i '/testsuites/testsuite[*]' -t attr -n timestamp -v $(date --iso-8601=seconds) \
+        -i '/testsuites/testsuite[*]/testcase' -t attr -n timestamp -v $(date --iso-8601=seconds) \
+        tests/reports/TEST-result-${{ github.run_id }}.xml
 
     - name: Upload test results
       run: s3cmd sync tests/reports s3://terraform-test-reports/${{ matrix.solution.name }}/
       if: always()
 
+    - name: Load test report history
+      uses: actions/checkout@v4
+      if: always()
+      continue-on-error: true
+      with:
+        ref: gh-pages
+        path: gh-pages
+
+    - name: Build test report
+      uses: simple-elf/[email protected]
+      if: always()
+      with:
+        gh_pages: gh-pages
+        subfolder: ${{ matrix.solution.name }}
+        allure_results: ${{ matrix.solution.name }}/tests/reports
+
+    - name: Publish test report
+      uses: peaceiris/actions-gh-pages@v4
+      if: always()
+      with:
+        github_token: ${{ secrets.GITHUB_TOKEN }}
+        publish_branch: gh-pages
+        publish_dir: allure-history
+        keep_files: true
+
   cleanup-infra:
     name: 'Cleanup Infra'
     environment:

compute-testing/tests/main.tftest.hcl

@@ -1,3 +1,3 @@
-run "create_cluster" {
+run "compute_testing_plan" {
   command = plan
 }

k8s-inference/tests/main.tftest.hcl

@@ -1,4 +1,27 @@
-run "create_cluster" {
+run "k8s_inference_apply" {
+  command = apply
+  plan_options {
+    target = [
+      nebius_mk8s_v1_cluster.k8s-cluster
+    ]
+  }
+}
+
+run "k8s_node_groups_inference_apply" {
+  command = apply
+  plan_options {
+    target = [
+      nebius_mk8s_v1_node_group.cpu-only,
+      nebius_mk8s_v1_node_group.gpu
+    ]
+  }
+}
+
+run "full_inference_apply" {
+  command = apply
+}
+
+run "test_mode_k8s_inference_apply" {
   command = apply
 
   variables {

k8s-training/tests/main.tftest.hcl

@@ -1,4 +1,31 @@
-run "create_cluster" {
+run "k8s_training_apply" {
+  command = apply
+  plan_options {
+    target = [
+      nebius_mk8s_v1_cluster.k8s-cluster
+    ]
+  }
+}
+
+run "k8s_node_groups_training_apply" {
+  command = apply
+  plan_options {
+    target = [
+      nebius_mk8s_v1_node_group.cpu-only,
+      nebius_mk8s_v1_node_group.gpu
+    ]
+  }
+}
+
+run "full_training_apply" {
+  command = apply
+
+  variables {
+    enable_loki = false # TODO: Disabling Loki since not possible to delete non-empty storage bucket
+  }
+}
+
+run "test_mode_k8s_training_apply" {
   command = apply
 
   variables {

slurm/tests/main.tftest.hcl

@@ -1,4 +1,26 @@
-run "create_cluster" {
+run "slurm_master_apply" {
+  command = apply
+
+  variables {
+    cluster_workers_count = 2
+  }
+
+  plan_options {
+    target = [
+      nebius_compute_v1_instance.master
+    ]
+  }
+}
+
+run "slurm_full_apply" {
+  command = apply
+
+  variables {
+    cluster_workers_count = 2
+  }
+}
+
+run "test_mode_slurm_apply" {
   command = apply
 
   variables {

wireguard/main.tf

@@ -1,6 +1,6 @@
-resource "nebius_compute_v1_instance" "wireguard-instanse" {
+resource "nebius_compute_v1_instance" "wireguard_instance" {
   parent_id = var.parent_id
-  name      = "wireguard-instanse"
+  name      = "wireguard-instance"
 
   boot_disk = {
     attach_mode   = "READ_WRITE"
@@ -9,16 +9,18 @@ resource "nebius_compute_v1_instance" "wireguard-instanse" {
 
   network_interfaces = [
     {
-      name              = "eth0"
-      subnet_id         = var.subnet_id
-      ip_address        = {}
-      public_ip_address = var.public_ip_allocation_id != null ? { allocation_id = var.public_ip_allocation_id } : {}
+      name       = "eth0"
+      subnet_id  = var.subnet_id
+      ip_address = {}
+      public_ip_address = {
+        allocation_id = var.public_ip_allocation_id
+      }
     }
   ]
 
   resources = {
     platform = "cpu-e2"
-    preset   = "16vcpu-64gb"
+    preset   = "4vcpu-16gb"
   }
 
 

wireguard/output.tf

@@ -0,0 +1,3 @@
+output "wg_instance_pib" {
+  value = trimsuffix(nebius_compute_v1_instance.wireguard_instance.status.network_interfaces[0].public_ip_address.address, "/32")
+}

wireguard/terraform.tfvars

@@ -1,5 +1,8 @@
-#parent_id               = ""
-#subnet_id               = ""
-#ssh_user_name           = ""
-#ssh_public_key          = ""
-#public_ip_allocation_id = ""
+# parent_id = ""
+# subnet_id = ""
+# ssh_user_name = "ubuntu"
+# ssh_public_key = {
+#   key  = "put your public ssh key here"
+#   path = "put path to ssh key here"
+# }
+# public_ip_allocation_id = ""

wireguard/test-resource.tf

@@ -0,0 +1,32 @@
+locals {
+  test_wg_host = trimsuffix(nebius_compute_v1_instance.wireguard_instance.status.network_interfaces[0].public_ip_address.address, "/32")
+}
+
+resource "null_resource" "check_wireguard_instance" {
+  count = var.test_mode ? 1 : 0
+
+  connection {
+    user = "ubuntu"
+    host = local.test_wg_host
+  }
+
+  provisioner "remote-exec" {
+    inline = [
+      "set -eu",
+      "cloud-init status --wait",
+      "ip link show wg0",
+      "systemctl -q status [email protected] > /dev/null",
+    ]
+  }
+}
+
+
+resource "null_resource" "check_wireguard_web_ui" {
+  depends_on = [null_resource.check_wireguard_instance]
+  count      = var.test_mode ? 1 : 0
+
+  provisioner "local-exec" {
+    interpreter = ["bash", "-c"]
+    command     = "sleep 15 && curl ${local.test_wg_host}:5000"
+  }
+}

wireguard/tests/main.tftest.hcl

@@ -0,0 +1,11 @@
+run "wireguard_apply" {
+  command = apply
+}
+
+run "test_mode_wireguard_apply" {
+  command = apply
+
+  variables {
+    test_mode = true
+  }
+}

wireguard/variables.tf

@@ -17,7 +17,7 @@ variable "ssh_user_name" {
 }
 
 variable "ssh_public_key" {
-  description = "SSH Public Key to access the cluster nodes"
+  description = "SSH Public Key to access the cluster nodes."
   type = object({
     key  = optional(string),
     path = optional(string, "~/.ssh/id_rsa.pub")
@@ -29,10 +29,15 @@ variable "ssh_public_key" {
   }
 }
 
-
 # Access By IP
 variable "public_ip_allocation_id" {
   description = "Id of a manually created public_ip_allocation."
   type        = string
   default     = null
 }
+
+variable "test_mode" {
+  description = "Switch between real usage and testing."
+  type        = bool
+  default     = false
+}