STRATISYSTEM INSTALLATION GUIDE
General installation guide for StratiSTOR, StratiSERV, and HyperSERV.
STRATISYSTEM OS — INSTALLATION GUIDE — REV 4.10
This guide provides the standard installation, preparation, provisioning, initial service activation, validation, and management-access workflow for StratiSYSTEM OS 4.9 deployments. It applies to StratiSTOR, StratiSERV, and HyperSERV unless a section is explicitly product-specific.
Terminology: StratiSTOR refers to the distributed storage and data-services layer, StratiSERV refers to the virtualization and compute layer, and HyperSERV refers to converged StratiSTOR and StratiSERV deployments. Low-level paths, executable names, and configuration keys are shown only where they are required for an exact installation procedure.
IMPORTANT: Use the release-approved StratiSYSTEM ISO, node-specific licenses, and an approved network and storage plan. Some steps in this guide are destructive. Verify device identity, network addressing, and cluster topology before proceeding.
DEPLOYMENT OVERVIEW
A StratiSYSTEM deployment follows a consistent sequence regardless of whether the target environment is StratiSTOR, StratiSERV, or HyperSERV:
- Plan: Validate hardware, firmware, BIOS settings, networking, addressing, licenses, and the intended deployment architecture.
- Install: Boot the StratiSYSTEM OS ISO and install the operating system on every node.
- Define: Edit and validate the product-specific current_deployment.json file.
- Prepare: Apply licensing, configure optional hardware support, review networking, and run the StratiSYSTEM Preparation Engine on every node.
- Provision: Initialize Node 1, join peer nodes, and complete cluster integration.
- Activate: Enable storage devices, SDCache, VM services, Kubernetes, data protocols, and other services required by the deployment.
- Validate and secure: Confirm health, connectivity, time synchronization, licensing, access, services, and production security settings.
A single node is supported for lab, evaluation, edge, and other non-HA use cases. Three or more nodes are preferred for highly available production clusters. Hardware does not need to be identical in every respect, but firmware, BIOS policy, network design, and intentional hardware differences should be documented and consistent with the deployment plan.
PRE-DEPLOYMENT PLANNING AND REQUIREMENTS
Hardware and OS Storage
- Use supported 64-bit Intel or AMD x86 hardware.
- Provide at least 200 GB of OS capacity per node.
- SSD-based OS storage is recommended.
- RAID is optional and supported. RAID 1 is recommended for production OS storage using hardware RAID or installer-configured software RAID.
- Do not use intended StratiSTOR data devices, cache devices, SAN LUNs, or previously deployed cluster disks as the OS target.
Firmware and BIOS Consistency
- Use a consistent boot mode across all nodes: legacy BIOS or UEFI.
- Use a consistent Secure Boot policy across all nodes. Supported deployments may use Secure Boot enabled or disabled, but the policy should remain consistent throughout the cluster.
- Set the system power profile to High Performance.
- Disable CPU C-states when required by the approved performance policy.
- Enable CPU virtualization extensions and IOMMU capabilities when virtual machines, PCIe passthrough, GPUs, or other accelerators will be used.
- Record BIOS, BMC, NIC, storage-controller, HBA, drive, and accelerator firmware versions before deployment.
Networking
Network design has a direct impact on availability and performance. Prepare the switch configuration, VLANs, bonds, LACP groups, MTU values, IP addresses, hostnames, DNS records, and routing before beginning the installation.
- Front-end network: 10 Gb or faster is strongly recommended. Production systems should use redundant interfaces where practical.
- Cluster-private network: 10 Gb or faster is strongly recommended. A dedicated, redundant, non-routed high-speed network is preferred for inter-node communication and applicable migration/storage traffic.
- Management network: May share the front-end network, although a dedicated 1/10 Gb management interface is preferred where available.
- Higher-performance environments may use 25/40/100 Gb and faster interfaces according to workload requirements.
- Ensure MTU settings match across the entire network path before enabling jumbo frames.
Recommended deployment information to record before installation:
- Node hostname and management IP address
- Front-end interface or bond, IP/prefix, gateway, DNS, and MTU
- Cluster-private interface or bond, IP/prefix, and MTU
- Service VIPs and DNS names
- OS disk and data-device mapping
- Node machine IDs and corresponding licenses
OPERATING SYSTEM INSTALLATION
Each StratiSYSTEM node requires an initial installation of the StratiSYSTEM OS. The recommended method is to mount the approved StratiSYSTEM ISO through the server's out-of-band management interface, such as Dell iDRAC, HPE iLO, Supermicro BMC/IPMI, or an equivalent remote-management platform.
For larger deployments, approved imaging workflows such as Clonezilla or other enterprise deployment tools may also be used when they are incorporated into the deployment design. Systems without remote virtual-media support can be installed from a bootable USB device.
Boot the Installation Media
Mount the StratiSYSTEM ISO, set the temporary boot device to the virtual optical drive or USB media, and boot the node. The following screen should appear:

Select option #1: Install SteelDome StratiSYSTEM OS.
After the installer loads, the main graphical installation screen will appear.

Select the required language and keyboard settings, then click CONTINUE.
The installer summary screen contains the major configuration areas required before installation. Begin with INSTALLATION DESTINATION.

SINGLE DISK OR HARDWARE RAID INSTALLATION
If the server presents a single OS disk or a hardware RAID virtual disk, select only that device under Local Standard Disks. Verify the device by controller, capacity, model, and identifier before continuing. Select CUSTOM under Storage Configuration and click DONE.

DESTRUCTIVE OPERATION: Removing partitions permanently destroys data on the selected disk. Never select StratiSTOR data disks, cache devices, SAN LUNs, or previously deployed cluster disks as the OS destination.
If pre-existing or foreign partitions exist on the selected OS target, remove them using the (-) control. For physical hardware, use Standard Partition unless a release-specific design explicitly requires another layout, then select Click here to create them automatically.

Review the automatically created partitions:
- Leave /boot and /boot/efi, when present, at the installer defaults.
- Remove the automatically created /home partition.
- Remove the automatically created swap partition for the standard StratiSYSTEM physical-node layout.
- Expand the root / filesystem to consume the remaining OS-disk capacity. Enter a value larger than the physical disk, such as 1 PiB, if necessary to force the installer to allocate all available remaining capacity.

Click DONE, review the proposed storage changes carefully, and click Accept Changes.

MULTIPLE DISK SOFTWARE RAID 1 INSTALLATION
If hardware RAID is not present and two OS disks are available, StratiSYSTEM can be installed using software RAID 1. Select only the intended OS disks, choose CUSTOM storage configuration, and click DONE.

Remove any stale partitions from the intended OS disks, choose Standard Partition, and allow the installer to create the initial partition layout.

As with a single-disk installation, retain the boot partitions, remove /home and swap, and allocate the remaining capacity to /.

For each OS partition that must be protected by software RAID, select RAID as the Device Type and RAID1 as the RAID Level. Confirm both intended OS devices are members of the RAID set.

OPTIONAL OS ENCRYPTION
OS-disk encryption may be enabled when required by the approved security design. Do not encrypt /boot. Select LUKS2 for supported encrypted OS partitions and ensure the operational key-unlock procedure is documented before production deployment.

When the partition layout is complete, click DONE. If encryption was enabled, the installer will request an encryption passphrase.
IMPORTANT: Store the encryption passphrase in the customer-approved secret-management system. The passphrase may be required during subsequent preparation or recovery operations.

Click Save Passphrase to continue.
HOSTNAME AND INITIAL NETWORK CONFIGURATION
After the installation-disk configuration is complete, open Network & Host Name.

Assign the final hostname for the node before installation. Use the approved site naming standard; common examples include stratistor1-1, stratistor1-2, and stratistor1-3.
Enable the interface that will provide initial management or front-end access.

Click CONFIGURE and enter the approved IPv4 address, prefix, gateway, and DNS settings. Configure at least one reachable management/front-end IPv4 interface during OS installation. Additional interfaces, bonds, VLANs, bridges, and private-cluster networking can be completed during preparation or through the management interface where appropriate.
Disable IPv6 when required by the current StratiSYSTEM deployment policy. Before continuing, verify the final hostname and confirm that the intended management connection is active.
COMPLETE THE OS INSTALLATION
Return to the installer summary screen and click BEGIN INSTALLATION.

Installation time depends on the hardware and installation-media path. When installation completes, unmount the ISO or USB installation media and click Reboot System.

Repeat the installation on every intended StratiSYSTEM node using the same boot, partitioning, firmware, and network policies.
INITIAL NODE VALIDATION
After each node reboots, connect by console or SSH and verify the basic configuration before proceeding:
hostnamectl
ip -4 address show
ip route
lsblk -o NAME,SIZE,TYPE,FSTYPE,MOUNTPOINTS,MODEL,SERIAL
ping -c 4 <default-gateway>
ping -c 4 <peer-node-management-ip>
Do not continue until every node has the correct hostname, is reachable over the approved management path, can reach required gateway/DNS/time services, and shows the expected OS and non-OS device inventory.
DEPLOYMENT CONFIGURATION
Before running the Preparation Engine, configure the product-specific deployment definition. The correct file depends on the target deployment:
- StratiSTOR:
/stratisystem-prep/stratistor/config/deployment/current_deployment.json - StratiSERV:
/stratisystem-prep/stratiserv/config/deployment/current_deployment.json - HyperSERV: Use the StratiSTOR deployment definition as the storage foundation. StratiSERV/HyperSERV compute services are enabled after the storage foundation is provisioned and healthy.
IMPORTANT: The validated deployment file must be identical on all nodes participating in the deployment. Do not begin preparation until JSON syntax is valid and the file checksum matches across all nodes.
Cluster Configuration Fields
- FrontendAccessSubnet — Defines the front-end/management subnet.
- ClusterPrivateNetwork and ClusterPrivateSubnet — Define the private inter-node cluster network.
Node Configuration Fields
- FrontendAccessInterface — Physical interface, bond/team, bridge, or VLAN interface used for front-end access.
- FrontendAccessIpAddr — Static front-end IPv4 address.
- FrontendAccessGwAddr — Front-end/default gateway. Use when this network is routed; avoid multiple default gateways unless explicitly designed.
- FrontendAccessDnsAddrs — Reachable DNS server address or addresses.
- ClusterPrivateInterface — Physical NIC or bond used for cluster-private communication. A dedicated interface is preferred where possible.
- ClusterPrivateBridge — Optional bridge for the cluster-private network. Normally left empty because VMs do not typically attach to this network.
- ClusterPrivateMtu — MTU used by the cluster-private network. This must match the complete switch and host path.
- ClusterPrivateIpAddr — Static IPv4 address used for private inter-node communication.
- ClusterPrivateGwAddr — Normally blank for a non-routed private cluster network. Use only when required by an approved routed or stretch design.
- ClusterPrivateDnsAddrs — Normally blank for the private cluster network.
- SmbPrimaryInterface — Interface or bridge used to present SMB services; commonly br0.
Useful interface-discovery commands:
ip -br a
ip link
ip route
Validate and Distribute the Configuration
Before copying the deployment file to peer nodes, validate the JSON and create a checksum:
python3 -m json.tool <path>/current_deployment.json >/dev/null
sha256sum <path>/current_deployment.json
Copy the approved file using a secure method, then compare checksums on every node:
scp <path>/current_deployment.json root@<peer-node>:<path>/current_deployment.json
sha256sum <path>/current_deployment.json
Verify that the NodeConfig array contains one entry per intended node, hostnames match hostnamectl, interface names match ip link, IP addresses are unique, MTUs are consistent, and service VIPs or DNS names do not conflict.
TIME SYNCHRONIZATION
Accurate and consistent time is critical to cluster operation. Use a reliable NTP source that is reachable from every node. For disconnected environments, one StratiSYSTEM node or another internal system can be configured as the local authoritative source.
If a StratiSYSTEM node will provide local time service, edit /etc/chrony.conf on that node. The example network below must be changed to match the deployment:
allow 192.168.1.0/24
bindaddress 0.0.0.0
local stratum 10
Restart or reload Chrony after making changes, then use that node's reachable IP address as the NTP source during preparation. Verify synchronization before provisioning.
PREPARATION PHASE
Run the StratiSYSTEM Preparation Engine on every node after the OS installation and deployment configuration have been validated.
Start the Preparation Engine:
cd /stratisystem-prep
source start.sh
LICENSE ACTIVATION
A valid SteelDome license is required for each node. Use the Licensing function in the Preparation Engine to obtain or display the node Machine ID, then apply the SteelDome-provided license that matches that specific node.
LICENSE MATCHING: A license generated for one Machine ID must not be applied to another node. Confirm node identity before importing each license.

NODE PREPARATION SELECTION
Select the workflow that matches the node's intended role. Menu numbering may vary slightly by release build; when in doubt, follow the displayed workflow label.
- New cluster: use the new-cluster preparation workflow on every node participating in the new cluster.
- Add node to existing cluster: use the add-node workflow.
- Add quorum node: use the quorum-node workflow.
- Reintegrate previously operational node: use the reintegration workflow and preserve existing data devices unless the recovery plan explicitly requires otherwise.
- Hardened deployment: use the hardened-mode workflow when required by the approved security design.
- Advanced networking: use the advanced-networking/Cockpit workflow when bonds, VLANs, bridges, or other network changes must be completed before cluster preparation.

ROOT ACCOUNT AUTHORIZATION
The Preparation Engine will request the root account password used by the current installation so it can perform privileged configuration tasks on the node.

NVIDIA AND AMD GPU SUPPORT
The Preparation Engine can install supported NVIDIA or AMD GPU frameworks when the deployment will use GPUs for AI, VDI, visualization, accelerated compute, CUDA/ROCm, DOCA, vGPU, or other approved accelerator workloads.

NOTE: NVIDIA vGPU functionality may require separate NVIDIA licensing obtained through an authorized NVIDIA licensing channel.
NTP TIME SOURCE
Enter a reachable network time source or accept the approved default presented by the workflow. All cluster nodes must use consistent, healthy time synchronization before provisioning.

NETWORK CONFIGURATION REVIEW
The Preparation Engine displays network values obtained from the deployment configuration. Review every value carefully. The workflow can override configuration-file values, so confirm the final settings before continuing.
Press ENTER to accept a displayed value or enter a replacement when a correction is required.

DRIVE WIPE — NON-BOOT DEVICES ONLY

The preparation workflow may display attached physical block devices and offer to wipe eligible non-boot devices.
WARNING: Wiping a device permanently destroys its contents. Only wipe a non-boot device after positively identifying it as disposable.
- Do not wipe storage from a previously operational node that is being reintegrated.
- Do not assume every visible non-boot device is local or disposable.
- Fibre Channel, iSCSI, NVMe-oF, multipath, or other externally presented LUNs may appear in the block-device inventory.
- When device ownership is uncertain, skip the automated wipe operation and validate the device outside the wizard before performing any destructive action.
BEGIN SYSTEM PREPARATION

Review the final summary and begin preparation. Actual duration varies by hardware, storage media, network access, firmware, and selected services.
The preparation stage normally performs two automated reboots. Do not begin provisioning until every node has completed the preparation workflow.

After the final reboot, verify that the preparation-complete marker exists:
ls -l /.prepcomplete
If the marker is missing, inspect the preparation log before attempting provisioning:
less /.preplog
tail -n 200 /.preplog
GATE TO PROVISIONING: Every node must have /.prepcomplete before the Provisioning Engine is started. Resolve preparation errors first; do not attempt to bypass incomplete phases manually.
PROVISIONING PHASE
After every node has completed preparation, start the product-specific Provisioning Engine.
- StratiSTOR / HyperSERV storage foundation:
cd /stratistor ./start.sh - StratiSERV:
cd /stratiserv ./start.sh
The provisioning menu will appear.

STANDARD NEW-CLUSTER WORKFLOW
For a standard new multi-node cluster:
- Node 1: Select the workflow to initialize the first node of a new cluster. In the current workflow this is typically option 1.
- All remaining nodes: Select the workflow to install the node as a peer in the new cluster. In the current workflow this is typically option 2.
Other workflows may include adding nodes to an existing cluster, adding quorum-only nodes, reintegrating previously operational nodes, single-node deployment, attaching existing StratiSERV nodes, or adding HyperSERV compute-only resources. Follow the workflow label presented by the installed release if menu numbering differs.
Validate the hostname, addresses, and intended node role before starting each workflow.

PHASE 1 — INITIAL NODE AND PEER PROVISIONING
Begin provisioning on Node 1 and all peer nodes as directed by the wizard. Allow every peer node to complete Phase 1 successfully.

PHASE ORDERING IS MANDATORY: Keep Node 1 at the Phase 1 completion point until all peer nodes have completed successfully. Do not start Phase 2 while any peer node is still running Phase 1 or has failed.
PHASE 2 — CLUSTER INTEGRATION
After every peer node has completed Phase 1, continue with Phase 2 on Node 1. This stage integrates the participating services across the cluster.

Confirm successful completion before moving to product-specific activation.
- StratiSTOR: Activate storage devices and SDCache as required, then enable the approved block, file, object, and other data services.
- StratiSERV: Activate VM services, Kubernetes/container services, or both according to the deployment plan.
- HyperSERV: Complete StratiSTOR storage activation first, verify the data foundation is healthy, and then enable StratiSERV/HyperSERV compute services.
SDCACHE AND STORAGE ACTIVATION
This section applies to StratiSTOR and the storage portion of HyperSERV.
SDCache provides RAM- or disk-based acceleration for supported storage devices. The optimal mode and allocation depend on workload characteristics, storage media, node memory, and the approved design.
Use the SDCache configuration workflow under the Storage Management section of the Provisioning Engine to generate and validate the configuration for the node hardware.
IMPORTANT: SDCache configuration files describe currently unactivated devices. After activation, regenerate or update the configuration before using it to activate additional devices.

If validation succeeds, confirm the device mapping before proceeding. Do not automatically add devices when encryption, device groups, or another specialized storage layout requires a different activation workflow.
After all intended storage devices have been activated, verify that the storage cluster is healthy before enabling production data services.

ACTIVATING ADDITIONAL SERVICES
After the base cluster is healthy, activate only the services required by the approved deployment. Depending on the selected product role and installed build, services can include:
- SMB file services — Managed through the main StratiSYSTEM console on TCP 3199; storage-specific administration is also available directly on TCP 3198 where enabled.
- NFS file services — Managed through the main StratiSYSTEM console on TCP 3199; storage-specific administration is also available directly on TCP 3198 where enabled.
- iSCSI block storage — Service configuration is performed through StratiSYSTEM management on TCP 3199 or the direct StratiSTOR administration interface on TCP 3198.
- NVMe-oF block storage — Service configuration is performed through StratiSYSTEM management on TCP 3199 or the direct StratiSTOR administration interface on TCP 3198.
- S3 object storage — Configure the service through StratiSYSTEM management on TCP 3199 or direct StratiSTOR administration on TCP 3198. The actual S3 data-service endpoint is defined separately by the deployed service configuration.
- Virtual machine services — Managed from the StratiSERV module embedded in the main console on TCP 3199, including the
/stratiserv/stratiservmanagement path where applicable. - Kubernetes and container services — Managed through StratiSYSTEM on TCP 3199. When the dedicated container-management interface is enabled, it is available on TCP 3197.
- Lustre and other supported parallel/high-performance data services — Activated and managed through StratiSYSTEM on TCP 3199; workload-facing service ports depend on the deployed service design.
- Application Delivery / VDI services — Managed through StratiSYSTEM on TCP 3199. The dedicated VDI administration interface is available on TCP 12443 when enabled.
- Monitoring and observability services — Core health is visible through StratiSYSTEM on TCP 3199. The dedicated monitoring interface is available on TCP 30443 when enabled.
- Global scheduling and automation — StratiSYSTEM automation is integrated into the platform; the dedicated Global Scheduler interface is available on TCP 3013 when enabled.
- OpenStack and private-cloud services — Activated from StratiSYSTEM management on TCP 3199; OpenStack service endpoints are defined by the deployed cloud configuration.
- Supported AI/GPU infrastructure and accelerator services — Provisioning and operational visibility are accessed through StratiSYSTEM on TCP 3199; workload-specific application ports remain service dependent.
- Advanced networking services — Supported SR-IOV, RDMA, DPDK, VPP, OVN, and related capabilities are configured through the applicable StratiSYSTEM management workflows on TCP 3199 and host-level networking tools where required.
For storage services, confirm service-specific VIPs, DNS names, networks, pool names, and protection settings before activation. Enable HA variants only when node count, addressing, and quorum design support them. Validate each production service before enabling additional protocols.
MANAGEMENT CONSOLES
StratiSYSTEM presents management functions according to the services installed on the node or cluster. The primary StratiSYSTEM administration console is available from any participating node at:
https://<node-ip-or-hostname>:3199
The main dashboard provides cluster status and access to installed management modules. The navigation presented in the interface reflects the services enabled on that system.
For StratiSTOR and the storage portion of HyperSERV, the dedicated storage-administration interface is available directly on TCP 3198 when enabled:
https://<node-ip-or-hostname>:3198
For StratiSERV virtual-machine administration, use the StratiSERV module within the primary TCP 3199 console. A direct path may be available at:
https://<node-ip-or-hostname>:3199/stratiserv/stratiserv


BOOTSTRAP ADMINISTRATIVE ACCOUNTS
Initial administrative accounts may include:
- StratiSTOR / HyperSERV storage administration:
stratistor-admin - StratiSERV administration:
stratiserv-admin
Use the bootstrap credentials supplied with the deployment or defined in current_deployment.json.
SECURITY REQUIREMENT: Change all bootstrap and default credentials immediately after first successful access and before exposing management interfaces to broader production networks. Store replacement credentials in the customer-approved secret-management system.
MANAGEMENT PORT REFERENCE
The following table consolidates the StratiSYSTEM management interfaces referenced throughout this guide. Not every interface is present on every deployment; availability depends on the product role and enabled services.
| Interface | TCP Port | Typical Endpoint / Purpose |
| Primary StratiSYSTEM Administration | 3199 | https://<node>:3199 — Main StratiSTOR / StratiSERV / HyperSERV management console. |
| StratiSTOR Storage Administration | 3198 | https://<node>:3198 — Direct storage-management interface when enabled. |
| StratiSERV VM Administration | 3199 | https://<node>:3199/stratiserv/stratiserv — VM-management module within the primary console. |
| Container Administration | 3197 | https://<node>:3197 — Dedicated container-management interface when enabled. |
| Monitoring | 30443 | https://<node>:30443 — Dedicated monitoring and observability interface when enabled. |
| Global Scheduler | 3013 | https://<node>:3013 — Dedicated scheduling/automation interface when enabled. |
| VDI Administration | 12443 | https://<node>:12443 — Dedicated Application Delivery / VDI administration interface when enabled. |
NOTE: These are management-plane interfaces. Workload-facing protocol ports such as SMB, NFS, iSCSI, NVMe-oF, S3, Kubernetes, OpenStack, and application-specific services are governed by their respective service configurations and network/security policies.
POST-DEPLOYMENT VALIDATION
Complete the following acceptance checks before placing the environment into production:
- Every intended node appears in the cluster with the expected hostname and management address.
- Cluster-private communication succeeds between all nodes using the intended addresses, interfaces, and MTU.
- Time synchronization is healthy and node clocks are aligned.
- Licensing is valid on every node and matches the node Machine ID.
- No preparation or provisioning phase remains incomplete.
- For StratiSTOR/HyperSERV, the expected storage-device count is online and storage health is stable.
- For StratiSERV/HyperSERV, VM and/or Kubernetes/container services are active as designed.
- Configured VIPs are unique, reachable, and hosted by the expected service.
- Management UI access works through the approved management path.
- Installation media has been unmounted and normal boot order has been restored.
POST-DEPLOYMENT SECURITY HARDENING
- Rotate all bootstrap passwords and verify that default credentials no longer work.
- Restrict StratiSYSTEM management ports to approved administrative networks and users. This includes TCP 3199 for the primary console and, when enabled, TCP 3198 for direct StratiSTOR administration, 3197 for container management, 30443 for monitoring, 3013 for the Global Scheduler, and 12443 for VDI administration.
- Configure certificate trust and replace default or self-signed certificates when required by customer policy.
- Confirm Secure Boot, disk encryption, accelerator-driver support, and passthrough options match the approved security design.
- Verify alert recipients, audit/log retention, DNS, and time synchronization.
- Remove temporary installation files, copied license material, and unsecured credential notes from administrative systems.
- Document recovery and break-glass credentials in the approved secret-management platform.
OPERATIONAL HANDOFF
Record the final production configuration so the deployed system can be operated, expanded, and supported without reconstructing installation details later.
- As-built record: node names, serial numbers, management/front-end/private addresses, interfaces, MTUs, VIPs, and service endpoints.
- Storage map: OS disks, data devices, cache devices, device groups, RAID policy, and service/pool mapping.
- Configuration archive: validated deployment JSON and SDCache/device-group definitions, with secrets removed or protected.
- Access record: administrative ownership, secret location, certificate status, and approved management networks.
- Health evidence: screenshots or command output demonstrating node, cluster, storage, service, and license health.
- Support record: installed StratiSYSTEM release, relevant license identifiers, partner/customer contacts, and escalation path.
TROUBLESHOOTING CHECKPOINTS
Node Unreachable After Installation
Use the local or BMC console and verify address configuration, routing, NIC link state, switch VLAN assignment, and IP uniqueness:
ip -4 address
ip route
ip link
Deployment JSON Validation Failure
Restore the approved backup or correct the JSON syntax, then validate it before redistribution:
python3 -m json.tool <path>/current_deployment.json >/dev/null
Deployment File Checksums Differ
Select one approved master deployment file, redistribute it, and compare checksums again:
sha256sum <path>/current_deployment.json
License Rejected
Confirm that the license matches the node Machine ID and the intended StratiSYSTEM release. Obtain a corrected node-specific license if necessary.
Preparation Does Not Complete
Do not proceed to provisioning. Review the preparation log, resolve the first relevant error, and rerun the supported preparation workflow:
ls -l /.prepcomplete
less /.preplog
tail -n 200 /.preplog
Peer Node Fails Provisioning Phase 1
Do not start Phase 2 on Node 1. Verify networking, DNS/hostname resolution, time synchronization, credentials, and deployment-file consistency, then repeat the supported peer-node workflow after correcting the issue.
CREATING A BOOTABLE USB DRIVE
The StratiSYSTEM OS ISO can be written to a USB device for systems without remote virtual-media support. Rufus or an equivalent ISO-writing utility can be used.
For the currently documented Rufus workflow, use:
- Target System: BIOS or UEFI
- Partition Scheme: MBR
- Quick Format: Enabled
- Cluster Size: 16 KB (default)

After the write completes, safely eject the USB device. The server may require a temporary boot-device selection or boot-order change before starting from USB.
NETWORKING CONSIDERATIONS
Production StratiSYSTEM environments should use redundant front-end and cluster-private connectivity wherever the hardware and switching environment permit it. Supported bonding designs can include active-backup, load-balancing modes, and IEEE 802.3ad/LACP.
For LACP environments, a Layer 3+4 transmit hash policy is recommended when supported by the network design because it distributes multiple socket flows more effectively across member links. Example:
nmcli connection modify <BOND_NAME> bond.options "xmit_hash_policy=layer3+4,miimon=100,mode=802.3ad,lacp_rate=fast,downdelay=0,updelay=0"
Validate switch-side LACP configuration, MTU, VLAN membership, link speed, and redundancy before placing the cluster into production. For high-performance workloads, also review NUMA placement, PCIe bandwidth, NIC queues, RSS, RDMA configuration, and advanced StratiSYSTEM networking capabilities such as SR-IOV, DPDK, VPP, and hardware offload where they are part of the approved architecture.