Merge pull request #19 from omnivector-solutions/matheushent/PENG-236…

…6--add-vantage-agent

PENG-2366  add vantage agent

.python-version

@@ -0,0 +1 @@
+3.11.1

README.md

@@ -4,12 +4,12 @@ This project contains the democluster image producing codebase.
 
 
 #### Tuning
-The environment variables `JG_VERSION` and `ENV` are exposed as tunables
+The environment variables `JG_VERSION`, `VTG_VERSION` and `ENV` are exposed as tunables
 to enable customizing the `democluster` for development purposes.
 
 Example
 ```bash
-ENV=dev JG_VERSION=4.3.1 \
+ENV=dev JG_VERSION=4.3.1 VTG_VERSION=2.3.0 \
   ./public-scripts/deploy-democluster.sh \
       aset-fc8b1039-faa7-47b1-967a-c1a55c418740 \
       9mWa98GbTJMcBZhinfy08aqHPyQWZUn7tH_XrAGLiYE

democluster/CHANGELOG.rst

@@ -7,6 +7,13 @@ Tracking of all notable changes to the Demo Cluster image.
 Unreleased
 ----------
 
+0.4.0 - 2024-11-25
+------------------
+
+- Install the Jobbergate Agent and the Vantage Agent from the Snap Store in classic mode (`PENG-2372`_).
+
+.. _PENG-2372: https://app.clickup.com/t/18022949/PENG-2372
+
 0.3.0 - 2024-10-13
 ------------------
 

democluster/Makefile

@@ -19,9 +19,6 @@ check-deps: ## Check deps needed to build the image
 
 .PHONY: init
 init: ## Run packer init .
-	@if [ ! -z "$JG_VERSION" ]; then\
-		sed -i 's/jobbergate-agent==[0-9]\+\.[0-9]\+\.[0-9]\+/jobbergate-agent==$(JG_VERSION)/' user-data;\
-    fi
 	${PACKER} init .
 
 .PHONY: stage0

democluster/user-data

@@ -190,6 +190,25 @@ write_files:
     Alias=jobbergate-agent.service
     WantedBy=multi-user.target
 
+- path: /usr/lib/systemd/system/vantage-agent.service
+  owner: root:root
+  permissions: '0644'
+  content: |
+    [Unit]
+    Description=vantage-agent
+    After=network.target
+
+    [Service]
+    Type=simple
+    User=root
+    Group=root
+    WorkingDirectory=/srv/vantage-agent-venv
+    ExecStart=/srv/vantage-agent-venv/bin/vtg-run
+
+    [Install]
+    Alias=vantage-agent.service
+    WantedBy=multi-user.target
+
 - path: /etc/slurm/slurmrestd.conf
   owner: root:root
   permissions: '0600'
@@ -384,6 +403,12 @@ packages:
 - apptainer
 - influxdb
 - influxdb-client
+
+snap:
+  commands:
+    0: snap install vantage-agent --channel=stable --classic
+    1: snap install jobbergate-agent --channel=stable --classic
+
 bootcmd:
 - mkdir /run/packer_backup
 - mkdir /run/packer_backup/etc
@@ -461,22 +486,3 @@ runcmd:
   EOF
 
 - systemctl daemon-reload
-
-# create jobbergate agent venv
-- /usr/bin/python3 -m venv /srv/jobbergate-agent-venv
-- /srv/jobbergate-agent-venv/bin/pip install -U pip
-- /srv/jobbergate-agent-venv/bin/pip install jobbergate-agent==5.1.0
-- |
-  cat <<EOF > /srv/jobbergate-agent-venv/.env
-  JOBBERGATE_AGENT_X_SLURM_USER_NAME=root
-  JOBBERGATE_AGENT_BASE_API_URL=https://apis.@[email protected]
-  JOBBERGATE_AGENT_OIDC_DOMAIN=auth.@[email protected]/realms/vantage
-  JOBBERGATE_AGENT_OIDC_AUDIENCE=https://apis.vantagehpc.io
-  JOBBERGATE_AGENT_OIDC_CLIENT_ID=@CLIENT_ID@
-  JOBBERGATE_AGENT_OIDC_CLIENT_SECRET=@CLIENT_SECRET@
-  JOBBERGATE_AGENT_TASK_JOBS_INTERVAL_SECONDS=30
-  JOBBERGATE_AGENT_TASK_SELF_UPDATE_INTERVAL_SECONDS=30
-  EOF
-  chown root:root /srv/jobbergate-agent-venv/.env
-  chmod 0644 /srv/jobbergate-agent-venv/.env
-- systemctl daemon-reload

image_factory/builder.py

@@ -68,7 +68,7 @@ def democluster(
     )
 
     # Mount needed directories into the instance
-    lxd_instance = lxd.LXDInstance(name=instance_name, project="image-factory")
+    lxd_instance = lxd.LXDInstance(name=instance_name, project=ctx.obj.project_name)
     lxd_instance.mount(host_source=Path(os.getcwd()), target=Path("/srv/image-factory"))
 
     # If you already have a packer cache in your home, mount it to speed things up

public-scripts/deploy-democluster.sh

@@ -62,9 +62,11 @@ launch_instance () {
 
   # Set the environment to the empty string if not supplied
   if [ -z $ENV ]; then
-      ENVIRONMENT=""
+      BASE_API_URL="https://apis.vantagehpc.io"
+      OIDC_DOMAIN="auth.vantagehpc.io/realms/vantage"
   else
-      ENVIRONMENT="${ENV}."
+      BASE_API_URL="https://apis.${ENV}.vantagehpc.io"
+      OIDC_DOMAIN="auth.${ENV}.vantagehpc.io/realms/vantage"
   fi
 
   # Create the cloud-init file and launch the demo cluster instance.
@@ -91,24 +93,25 @@ runcmd:
 
     REAL_MEMORY=\$(free -m | grep -oP '\\d+' | head -n 1)
     sed -i "s|@REAL_MEMORY@|\$REAL_MEMORY|g" /etc/slurm/slurm.conf
-  - |
-    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
-    sed -i "s|@ENVIRONMENT@|$ENVIRONMENT|g" /srv/jobbergate-agent-venv/.env
   - systemctl start slurmrestd
   - systemctl restart slurmdbd
   - systemctl restart slurmd
   - sleep 30
   - systemctl restart slurmctld
   - scontrol update NodeName=\$(hostname) State=RESUME
-  - systemctl start jobbergate-agent
+  - snap set vantage-agent base-api-url=$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 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 task-jobs-interval-seconds=30
+  - snap set jobbergate-agent x-slurm-user-name=root
+  - snap start vantage-agent.start
+  - snap start jobbergate-agent.start
 EOF
 
-  if ! [ -z "${JG_VERSION}" ]; then
-      echo "  - systemctl stop jobbergate-agent" >> /tmp/cloud-init.yaml
-      echo "  - /srv/jobbergate-agent-venv/bin/pip install -U jobbergate-agent==$JG_VERSION" >> /tmp/cloud-init.yaml
-      echo "  - systemctl start jobbergate-agent" >> /tmp/cloud-init.yaml
-  fi
   mkdir -p $HOME/democluster/tmp
 
   cat /tmp/cloud-init.yaml | multipass launch -c$(nproc) \