VergeIO · VergeOS 26.2

VergeOS 26.2:
Complete the VMware Exit

  • Initial and final VMware transfers run 120% faster
  • Microsoft Failover Clustering on VergeFS, no storage array
  • Remote sites on one server with two network connections
The Infrastructure Operating System
verge.io
01 / 14
What Makes Us Different

Every Function. Unified Infrastructure.

Hypervisor Storage Networking Backup Disaster Recovery API API API API API VergeOS Infrastructure OS Unified infrastructure Hypervisor Storage Networking Backup DR
Five products, five upgrades, five vendors. One platform, one unified infrastructure.
  • VergeOS runs compute, storage, networking, and data protection as one unified infrastructure.
  • One install turns standard x86 servers into a complete infrastructure, including the DR site.
  • One upgrade moves the whole environment forward, so the maintenance window shrinks to hours.
  • An operating system, not a stack, means less overhead and a smaller footprint than assembled products, so the same physical servers hold more VMs.
verge.io
02 / 14
Where Teams Start

Four Ways Customers Put It to Work

Most environments arrive through one of these four doors and then expand into the others. VergeOS 26.2 strengthens three of them.

01
Exit VMware
Virtual machines import from the running VMware environment and come up on VergeOS with their storage and networking intact. Applications stay as they are.
26.2 · 120% faster transfers
02
Fight Data Center Inflation
Stretch the hardware you already own. VergeOS mixes server generations in one cluster, runs on refurbished nodes, and uses less RAM and storage per workload, so the next purchase order moves out a year or more.
26.2 · no array for clusters
03
Run Every Site the Same Way
From one place. Every location runs the same VergeOS and takes the same upgrade, automated with Terraform, Ansible, and Packer. A remote site becomes part of the environment, not its own project.
26.2 · one-server sites
04
Reduce Token Costs
Models run on GPUs inside your own tenant, on the same operating system that runs production. The per-token bill goes away, and prompts and data stay on your network.
verge.io
03 / 14
What It Looks Like

Mixed Hardware. One Platform.

VDC 01 Finance Dept · Enterprise Isolated tenant VM VM VM VM VM VM VDC 02 Acme Corp · CSP Customer Isolated tenant VM VM VM VM VM VM VDC 03 Dev/Test Lab · Enterprise Isolated tenant VM VM VM VM VM VM VERGEOS Dell HPE Cisco UCS Supermicro
Three isolated tenants, one operating system, four server vendors in the same cluster.
  • Step 1. Install VergeOS on your hardware. Dell, HPE, Cisco UCS, and Supermicro run side by side in the same cluster, including older generations and refurbished nodes.
  • Step 2. Migrate your VMs to the new infrastructure. They come across with their data and network settings intact.
  • Step 3. Use the capabilities VergeOS adds, starting with virtual data centers that give each department or customer its own isolated compute, storage, network, and protection.
verge.io
04 / 14
New in VergeOS 26.2

Three Changes That Complete the VMware Exit

The bulk of the VMs move quickly. The clustered applications and the small remote sites keep a second hypervisor and a VMware renewal alive. VergeOS 26.2 goes after each of those holdouts.

01
120%faster
Faster VMware Migration
Initial and final transfers run 120% faster than on VergeOS 26.1, so cutover gets even shorter. Tags, categories, and a unique serial number come across with each VM.
Exit VMware
02
SCSI-3PR
Mission-Critical Windows Workloads
Microsoft Failover Clustering runs on VergeFS with shared disks and SCSI-3 Persistent Reservations. Clustered SQL Server moves without a dedicated storage array.
Exit VMware · Fight Inflation
03
1server per site
Single-Server Distributed Data Centers
A remote site needs one server with two network connections and no core network, so retail, branch, and plant locations come along in the exit.
Run Every Site the Same Way
verge.io
05 / 14
Exit VMware · New in 26.2

Complete the VMware Exit. 120% Faster Transfers.

Production on three VMware hosts vMotion the VMs off host 3 Reimage host 3 with VergeOS Full transfer of host 2’s VMs, 120% faster in 26.2 Test the copies on VergeOS Final transfer: changed blocks only, 120% faster Cutover: the VMs now run on VergeOS Reimage host 2 with VergeOS Spread the VMs across both VergeOS hosts Full transfer of host 1’s VMs, tags and categories included Test the copies on VergeOS Final transfer: changed blocks only, 120% faster Cutover: the VMs now run on VergeOS Reimage host 1 with VergeOS One VergeOS cluster on the same three servers VMware ESXiServer 1 VergeOSServer 1 VMware ESXiServer 2 VergeOSServer 2 VMware ESXiServer 3 VergeOSServer 3 VM 1 VM 2 VM 3 VM 4 VM 5 VM 6 VM 7 VM 8 VM 4 VM 5 VM 6 VM 8 VM 1 VM 2 VM 3 VM 7 VERGEOS · ONE CLUSTER, SAME HARDWARE
  1. 1Start. Three VMware hosts running production.
  2. 2Free host 3. vMotion its VMs to the other two and reimage it with VergeOS.
  3. 3Move host 2’s VMs. Full transfer, test, final transfer of changed blocks, cut over.FASTER
  4. 4Convert host 2. Reimage it and spread the VMs across both VergeOS hosts.
  5. 5Move host 1’s VMs. Tags, categories, and a unique serial come across with each VM.NEW
  6. 6Finish. Reimage host 1 and spread the VMs across all three.
What 26.2 changes
2.2x transfer speed

Asynchronous I/O on NBD transfers speeds the initial and the final transfer. Changed block tracking still limits the final transfer to changed blocks.

VergeIO lab testing over NBD, VergeOS 26.2 vs. 26.1.
verge.io
06 / 14
Faster VMware Migration · Recorded Demo

Complete the VMware Exit. Watch the Transfer.

▶What to watch
  • Same workflow. VMware service, backup job, Import VMs. Nothing new to learn.
  • What changed underneath. 26.2 adds asynchronous I/O buffers to NBD transfers and faster VM list retrieval.
  • Prove it on your own VMs. Upgrade from 26.1.6 or later and time a pilot VM before and after.
Narrated demo · 0:38 · unmute for audio
verge.io
07 / 14
VMware Migration Testing

Prove It Runs, Then Complete the VMware Exit

Moving the data only proves VergeOS can receive it. Snapshots and virtual data centers prove each workload runs before it goes into production.

01
<1sec each
Snapshots for Aggressive Testing
VergeOS keeps the original VMware image and offers a snapshot when the VM is created. Snapshots are independent, so any one can become the new source.
02
1VDC per workload
Isolate Each Workload
Each migrated workload goes into its own virtual data center, with CPU and GPU limits that protect the workloads already certified in production.
03
5–10copies
Test in Parallel
Snapshot the VDC and hand a copy to each tester. Same IP addresses, no conflicts. Promote the best configuration and delete the rest.
04
24/7testing
AI on the Testing Team
A local LLM runs on your GPUs in its own VDC and works through test cases around the clock. Application data stays in your environment.
From tested to production. The certified VDC becomes production, with its VMs, network settings, and configuration. Nothing is rebuilt before go-live. Read the blog →
verge.io
08 / 14
VMware Migration Testing · Recorded Demo

AI-Tested Migration. Ready Before Cutover.

▶What to watch
  • A golden VDC. The migrated application keeps its known-good state as a snapshot, and each team gets its own copy.
  • A fresh copy for every test. The latest changed blocks from VMware land in a new, isolated VDC, so tests always see what runs today.
  • A local AI model on the team. It checks everything, explains what failed, and checks again after the fix: 41 of 41 checks passed.
Narrated demo · 5:49 · unmute for audio
verge.io
09 / 14
Mission-Critical Windows Workloads · New in 26.2

Failover Clustering on VergeFS. No Array.

On VMware, the cluster’s shared disk sits on an array The array stays for one cluster, and so does VMware VergeOS 26.2 creates the shared disk inside VergeFS Both cluster nodes attach to the same disk SCSI-3 Persistent Reservations: node A owns the disk Node A fails and node B takes the reservation The cluster runs on the same storage as every other VM WINDOWS FAILOVER CLUSTER Storage arrayDedicatedShared LUN · cluster diskIts own refresh cycle and support contract VergeFSServer 1Server 2Server 3Shared disk (NAA / WWN identity)Other VM data, drives inside the servers Multi-writer SCSI-3 PR SQL node AWindows cluster node SQL node BWindows cluster node PR Dedicated arraySecond hypervisorVMware renewal SQL Server failover instancesClustered file serversCluster Shared Volumes VERGEOS 26.2 · CLUSTERS ON VERGEFS · NO ARRAY
  1. 1The holdout. A Windows cluster needs a shared disk, so a dedicated array, and VMware, stay in place.
  2. 2Shared disk in VergeFS. Mark a virtual disk as shared. It presents NAA and WWN identity.
  3. 3Attach to every node. One multi-writer disk attaches to each node in the guest cluster.
  4. 4Reservations arbitrate. SCSI-3 Persistent Reservations decide ownership, so failover works the way Windows expects.
  5. 5Retire the array. SQL Server, clustered file servers, and Cluster Shared Volumes run on the same storage as every other VM.
Scope: guest clusters. The cluster software inside the guests coordinates writes, as it does on any platform.
verge.io
10 / 14
Mission-Critical Windows Workloads · Recorded Demo

Failover Clustering on VergeOS Shared Disks

▶What to watch
  • Two-node Windows failover cluster. SQL Server and a clustered file server on VergeFS shared disks, with no external SAN.
  • Power cut to the active node. SQL Server is back on the surviving node in 31.4 seconds; the file share in 25.3 seconds.
  • Node returns. The failed node powers back on and rejoins the cluster.
Narrated demo · 1:49 · unmute for audio
verge.io
11 / 14
Distributed Data Centers · New in 26.2

Every Site, Even Single Server Locations

The main data center runs a VergeOS cluster Retail, branch, and plant sites run one server each Two network connections carry external and workload traffic No core network at the site Every site runs the same VergeOS and takes the same upgrade Any number of sites, run the same way The core and every site in one management interface One interface, one view, every site on VergeOS 26.2 VergeOSMain data centerServer 1Server 2Server 3Server 4Production cluster Same VergeOS · same upgrade RETAIL STOREVergeOS server NIC 1NIC 2 No core network BRANCH OFFICEVergeOS server NIC 1NIC 2 No core network PLANT FLOORVergeOS server NIC 1NIC 2 No core network and any number of sites VERGEOS · SINGLE MANAGEMENT INTERFACEOverviewSitesWorkloadsStorageNetworksUpdatesAll sites4 locations · 7 servers · one platformVergeOS 26.2Main data center4 serversOnlineVergeOS 26.2Retail store1 serverOnlineVergeOS 26.2Branch office1 serverOnlineVergeOS 26.2Plant floor1 serverOnlineVergeOS 26.2 ONE SERVER · TWO CONNECTIONS · NO CORE NETWORK ONE INTERFACE · EVERY SITE · SAME VERGEOS
  1. 1The core. The main data center runs a VergeOS cluster.
  2. 2One server per site. Retail, branch, and plant locations run VergeOS on a single server.
  3. 3Two connections. Both ports carry external and workload traffic.
  4. 4No core network. The site runs without one.
  5. 5Same platform everywhere. Every site runs the same VergeOS and takes the same upgrade, so those locations come along in the VMware exit.
  6. 6Single management interface. The main data center and every site come together in one interface.
Plays automatically · Replay to restart
verge.io
12 / 14
Distributed Data Centers · Recorded Demo

One Server at the Edge. Recovered at the Main Site.

▶What to watch
  • A retail store on one server. The store runs its applications on a single VergeOS 26.2 node.
  • Protected by site sync. The store replicates back to the main data center.
  • Recovered at the main site. The store comes back up in a virtual data center at the core.
Narrated demo · 2:35 · unmute for audio
verge.io
13 / 14
VergeOS 26.2 · Thank You

Complete the VMware Exit on Hardware You Already Own

VergeOS 26.2 brings the clustered applications and the single-server sites across with the rest of the estate. Bring the workload you think has to stay on VMware.

Request a
VergeOS Demo
verge.io/request-a-demo
Schedule a Technical
Virtual Whiteboard Session
meetings.hubspot.com/gcrump/technical-whiteboard-session
Every 26.2 capability in detail · verge.io/vergeos26
14 / 14
Arrow keys or Previous and Next to advance