Adding Nodes
Learn how to expand a VergeOS system by adding scale-out, storage-only, and compute-only nodes to your cluster using the USB installer.
Overview
Once your two-node controller cluster is operational, VergeOS supports three methods for expanding capacity: scale-out nodes (add both compute and storage to an existing HCI cluster), storage-only nodes (dedicated storage expansion), and compute-only nodes (dedicated compute expansion). Each node type follows a streamlined USB installer workflow that auto-detects the existing system on the core fabric.
This page covers the prerequisites, installation steps, and verification procedures for all three node types.
Prerequisites
Before adding any node to your VergeOS system, ensure the following conditions are met:
Operational 2-node cluster
The primary and secondary controller nodes from the Installation Guide must be fully installed and healthy
Network readiness
VLANs, physical NIC assignments, and core fabric cabling must match the configuration documented during initial setup
Matching hardware
Scale-out and storage-only nodes should have identical hardware specifications (CPU, storage, NICs, RAM) to existing nodes in the target cluster
USB installer
VergeOS installer media matching the version running on your existing cluster
Cluster created in UI
If adding a new cluster type (storage-only or compute-only) for the first time, create the cluster in the VergeOS UI before booting the installer
Create Clusters Before Installing Nodes
If this is the first storage-only or compute-only node you are adding, you must create the corresponding cluster in the VergeOS UI first:
Navigate to System → Clusters
Click New Cluster
Set the cluster's role with the independent Compute and Storage toggles — these are not a single either/or choice. Enable Storage only (Compute off) for a storage-only cluster, or Compute only (Storage off) for a compute-only cluster. (An HCI cluster has both enabled.)
Save the cluster configuration
The installer cannot add a node to a cluster that does not yet exist.
Node Type Comparison
Understanding the differences between node types helps you choose the right expansion strategy:
Installer selection
"Scale-Out"
"Scale-Out"
"Compute Only"
vSAN participation
Yes — disks added to existing tiers
Yes — disks added to existing tiers
No — installs without configuring local storage (PXE-booted compute-only nodes need no local boot drive)
Runs VM workloads
Yes
No
Yes
Requires cluster pre-creation
No (joins existing HCI cluster)
Yes (if first storage node)
Yes (if first compute node)
Typical use case
Balanced HCI growth
UCI storage expansion
GPU/ML/VDI compute expansion
Scale-Out Node Installation
Scale-out nodes are the simplest way to expand an HCI cluster. They add both compute and storage capacity proportionally, using hardware identical to the existing controller nodes.
Installation Steps
Boot from the VergeOS USB installer
Insert the USB installer into the server designated as the scale-out node
Reboot and select the USB drive as the boot device from the BIOS/UEFI boot menu
Select node type → "Scale-Out"
When the installer presents the node type menu, select Scale-Out
This tells the installer to join an existing cluster with both compute and storage roles
Enter admin credentials
Provide the admin username and password created during the primary controller installation
These credentials authenticate the new node against the existing system
Network auto-detection
The installer auto-detects the core fabric configuration by listening for UDP packets from existing nodes
Verify that detected NICs match your documented cable map
Tip: The installer may prompt you to identify which NIC is used for the External network if it cannot auto-detect. Have your NIC-to-port mapping documentation ready.
Cluster selection (if multiple clusters exist)
If your system has more than one cluster, the installer prompts you to select which cluster this node should join
Select the node in the target cluster that this new node most closely resembles in hardware
vSAN disk configuration
The installer displays detected disks and their proposed tier assignments
Confirm that storage tiers match the existing cluster configuration — uniformity is critical
Review and finalize disk selections
Complete installation
The installer formats drives, installs VergeOS, and integrates the node into the cluster
The node reboots automatically when installation completes
Remove the USB installer media when prompted
Verify in the UI
Log into the VergeOS web UI
Navigate to Infrastructure → Nodes and confirm the new node appears with an operational status
Check System → vSAN to verify new disks are integrated and tiers are healthy
Storage-Only Node Installation
Storage-only nodes expand vSAN capacity without adding compute resources. They are used in UCI (Unified Compute Infrastructure) architectures where storage and compute scale independently.
Pre-Installation: Create Storage Cluster
If this is your first storage-only node, create the storage cluster in the UI:
Navigate to System → Clusters
Click New Cluster
Enable the Storage toggle and leave Compute off
Save the configuration
Installation Steps
The storage-only node installation uses the same "Scale-Out" installer option as a regular scale-out node. The key difference is that the node joins a storage-only cluster instead of an HCI cluster.
Boot from USB installer — same as scale-out
Select node type → "Scale-Out" — the installer uses "Scale-Out" for both HCI and storage-only nodes
Enter admin credentials — authenticate against the existing system
Network auto-detection — verify NIC assignments
Cluster selection — select the storage cluster you created in the UI; select the node this new node most closely resembles
vSAN disk configuration — confirm tier assignments match the storage cluster's intended configuration
Complete installation — drives are formatted, VergeOS is installed, and the node reboots
Verify in the UI — confirm the node appears under the correct storage cluster on the Nodes page
Because storage-only nodes use the "Scale-Out" installer option, the differentiator is the cluster you select during step 5. The cluster type (storage-only) determines the node's role in the system.
Compute-Only Node Installation
Compute-only nodes provide additional CPU, RAM, and optionally GPU resources without participating in vSAN storage. They are ideal for workloads that need high processing power — machine learning, rendering, VDI, or data analytics.
Pre-Installation: Create Compute Cluster
If this is your first compute-only node, create the compute cluster in the UI:
Navigate to System → Clusters
Click New Cluster
Enable the Compute toggle and leave Storage off
Save the configuration
Installation Steps
Boot from USB installer — same as other node types
Select node type → "Compute Only" — this is a distinct installer option, different from "Scale-Out"
Enter admin credentials — authenticate against the existing system
Network auto-detection — verify NIC assignments
Cluster selection (optional) — if prompted, select the compute cluster and the node this new node most closely resembles
Complete installation — the installer configures network settings and installs without configuring local storage. Compute-only nodes can also be PXE-booted, in which case no local boot drive is required. The node reboots automatically.
Verify in the UI — confirm the node appears under the correct compute cluster on the Nodes page
Installation Order Best Practices
When deploying a multi-cluster system (UCI architecture), follow this recommended order:
Install controller nodes first — they form the foundation cluster with Tier 0 metadata storage
Add storage nodes second — ensures vSAN storage capacity is available before compute nodes need it
Add compute nodes last — they can immediately access storage over the core fabric
Within any cluster, add nodes sequentially — wait for each node to complete installation and appear healthy in the UI before starting the next. This avoids overlapping vSAN redistribution/rebuild operations as each new node joins.
Troubleshooting
If issues arise during any node installation:
Cancel installer
Press Esc
Drops to a command prompt
Resume installation
yb-install
Picks up where the installer left off
Restart from scratch
yb-install --restart
Resets and restarts the entire installation process
Common Issues
Node not appearing in the UI after reboot:
Verify network connectivity — ensure core fabric NICs are properly cabled
Check that the USB installer version matches the VergeOS version running on the controllers
Review node installation logs via the console
vSAN tier stays yellow for extended time:
This is normal during data redistribution — monitor progress in System → vSAN
Ensure sufficient bandwidth between nodes on the core fabric
Contact VergeOS support if rebuild stalls beyond expected timeframes
Installer cannot auto-detect network:
Verify core fabric switch ports are configured correctly (jumbo frames, STP disabled)
Check physical cabling matches the documented NIC-to-port mapping
Try manual network configuration if auto-detection fails repeatedly
"No clusters available" during cluster selection:
Ensure you created the target cluster (Storage or Compute) in the UI before booting the installer
Verify the cluster type matches the installer node type
Verification Checklist
After adding any node, confirm the following:
Key Takeaways
Three node types
Scale-out (HCI), storage-only, and compute-only — each with a specific role
Single installer
One USB installer media handles all node types via menu selection
Auto-detection
Nodes discover the system on the core fabric automatically
Cluster pre-creation
Storage and compute clusters must be created in the UI before adding the first node of that type
Sequential installation
Add nodes one at a time within a cluster to avoid overlapping vSAN redistribution/rebuild
Storage before compute
In UCI deployments, add storage nodes before compute nodes
vSAN rebuild
After adding storage-contributing nodes, wait for vSAN tiers to return to green before proceeding
Next Steps
With your nodes added and verified, proceed to post-installation verification to confirm overall system health: Post-Installation Verification →
Last updated
Was this helpful?