These changes upgrade the democluster to use slurm 23.11 and ubuntu noble 24.04.

* use linuxproc instead of cgroup
* remove slurmrestd refs and update local script
* add oidc-domain config for jobbergate and vantage agent snaps
* add --enable to jobbergate and agent start

Author: jamesbeedy

democluster/democluster.pkr.hcl

@@ -29,7 +29,7 @@ locals {
 
 variable "ubuntu_series" {
   type        = string
-  default     = "jammy"
+  default     = "noble"
   description = "The codename of the Ubuntu series to build."
 }
 
@@ -69,8 +69,8 @@ source "qemu" "stage0" {
     ["-cpu", "${lookup(local.qemu_cpu, var.architecture, "")}"],
     ["-serial", "stdio"],
     ["-device", "virtio-gpu-pci"],
-    ["-drive", "if=pflash,format=raw,id=ovmf_code,readonly=on,file=/usr/share/${lookup(local.uefi_imp, var.architecture, "")}/${lookup(local.uefi_imp, var.architecture, "")}_CODE.fd"],
-    ["-drive", "if=pflash,format=raw,id=ovmf_vars,file=${lookup(local.uefi_imp, var.architecture, "")}_VARS.fd"],
+    ["-drive", "if=pflash,format=raw,id=ovmf_code,readonly=on,file=/usr/share/OVMF/OVMF_CODE_4M.fd"],
+    ["-drive", "if=pflash,format=raw,id=ovmf_vars,file=/usr/share/OVMF/OVMF_VARS_4M.fd"],
     ["-drive", "file=seeds-cloudimg.iso,format=raw"],
     ["-drive", "file=output-stage0/packer-stage0,format=qcow2"]
   ]
@@ -86,7 +86,7 @@ build {
 
   provisioner "shell-local" {
     inline = [
-      "cp /usr/share/${lookup(local.uefi_imp, var.architecture, "")}/${lookup(local.uefi_imp, var.architecture, "")}_VARS.fd ${lookup(local.uefi_imp, var.architecture, "")}_VARS.fd",
+      "cp /usr/share/OVMF/OVMF_VARS_4M.fd OVMF_VARS_4M.fd",
       "cloud-localds seeds-cloudimg.iso user-data meta-data"
     ]
     inline_shebang = "/bin/bash -e"

democluster/helpers/deploy_local_democluster.sh

@@ -2,6 +2,8 @@
 
 CLIENT_ID=$1
 CLIENT_SECRET=$2
+BASE_API_URL="https://apis.vantagecompute.ai"
+OIDC_DOMAIN="auth.vantagecompute.ai/realms/vantage"
 
 cat <<EOF > /tmp/cloud-init.yaml
 #cloud-config
@@ -30,13 +32,26 @@ runcmd:
     sed -i "s|@CLIENT_ID@|$CLIENT_ID|g" /srv/jobbergate-agent-venv/.env
     sed -i "s|@CLIENT_SECRET@|$CLIENT_SECRET|g" /srv/jobbergate-agent-venv/.env
   - |
-  - systemctl start slurmrestd
   - systemctl restart slurmdbd
-  - systemctl restart slurmd
   - sleep 30
   - systemctl restart slurmctld
+  - systemctl restart slurmd
   - scontrol update NodeName=\$(hostname) State=RESUME
-  - systemctl start jobbergate-agent
+
+  - snap set vantage-agent base-api-url=$BASE_API_URL
+  - snap set vantage-agent oidc-client-id=$CLIENT_ID
+  - snap set vantage-agent oidc-client-secret=$CLIENT_SECRET
+  - snap set vantage-agent oidc-domain=$OIDC_DOMAIN
+  - snap set vantage-agent task-jobs-interval-seconds=30
+  - snap set jobbergate-agent base-api-url=$BASE_API_URL
+  - snap set jobbergate-agent oidc-client-id=$CLIENT_ID
+  - snap set jobbergate-agent oidc-client-secret=$CLIENT_SECRET
+  - snap set jobbergate-agent oidc-domain=$OIDC_DOMAIN
+  - snap set jobbergate-agent task-jobs-interval-seconds=30
+  - snap set jobbergate-agent x-slurm-user-name=ubuntu
+  - snap set jobbergate-agent influx-dsn=influxdb://slurm:rats@localhost:8086/slurm-job-metrics
+  - snap start vantage-agent.start --enable
+  - snap start jobbergate-agent.start --enable
 EOF
 
 mkdir -p $HOME/democluster