VMware vSAN / Storage virtualization

vSAN Storage Virtualization

Plan and deploy vSAN, diagnose and repair faults, and expand or migrate clusters, handing capacity, policies, failure domains and business validation to the operations team.

  • Planning and deployment / Compatibility and policies
  • Fault diagnosis / Repair and remediation
  • Expansion and migration / Operations handover

A resource pool is not simply the sum of raw disks. Protection overhead, rebuild headroom and independent backups need separate review.

How objects enter the resource pool

Follow one object across nodes through writes, reads, repair and rebalancing

This time-compressed example uses OSA RAID1 / FTT1: four hosts contribute local disks to a distributed datastore consumed by VMs. Failure states illustrate a sequence, not live cluster telemetry.

Operating example / OSA RAID1 · FTT1VM writes object A

The animation shows a VM writing object A, synchronous copies on H1 and H2, and a witness component on H4 participating in quorum. After H2 goes offline, H1 serves reads; policy is repaired on H3 when conditions permit. H5 later contributes capacity and triggers rebalancing.

VM / workloadObject AWrite request · Business VM
vSphere / vSAN distributed layerShared datastoreDistributed datastore for VMs · Policy-based object placementOSA example: SSD cache + SSD/HDD capacity
Witness / Logical componentWitness component on H4 (logical view)OSA RAID1 / FTT1 · No object payload
H1Online
Replica A1OSA disk group · SSD cache + SSD/HDD capacity · Local reads
H2Synchronization / Offline example
Replica A2OSA disk group · SSD cache + SSD/HDD capacity · Reads unavailable after failure
H3Available for repair
Replica recovery locationOSA disk group · SSD cache + SSD/HDD capacity · Copy when conditions permit
H4Online
Hosts witness / Reserved spaceOSA disk group · SSD cache + SSD/HDD capacity · Reserved headroom
Expansion / RebalancingH5 joins the resource poolAdd capacity and move objects by policy
Static relationships and directionsTime-compressed data blocksFailure / Repair states

The illustrative cycle lasts about 12 seconds. Pause holds the current frame; replay returns to writing object A. vSAN policy availability is not independent backup. Confirm design against versions, certified hardware, capacity and failure domains.

VMware vSAN storage virtualization architecture and implementation workflow
User-supplied conceptual architecture. SSD/HDD labels retain generic legacy terminology; verify the selected OSA or ESA architecture, version and certified equipment. The image is not a customer project or VMware-certified configuration.

Fault diagnosis and repair

When vSAN fails, Yuqi Intelligent helps diagnose, repair and validate recovery

Yuqi Intelligent has five VMware engineers handling cluster takeover, diagnosis, repair, migration and expansion. Services cover impact assessment, root-cause investigation, repair implementation and business validation for existing clusters.

A typical sequence is symptom → investigation scope → treatment. For inaccessible objects or disk faults, preserve evidence, review replicas and backups, and confirm the plan before acting.

SymptomInvestigation scopeTreatment direction
Cluster health alerts / Offline hostsReview vSAN Health, host status, failure domains, logs and recent changes.Preserve evidence, assess impact and available replicas, then arrange an authorized maintenance window.
Disk, disk-group or firmware/driver faultsCheck disk health, disk groups, controllers, firmware, drivers and supported version combinations.Distinguish hardware, disk-group and compatibility faults, then repair or replace components under the approved plan.
Storage-network latency, packet loss or partitionCheck vSAN, vMotion and management paths, switch ports, MTU, links and the event timeline.Resolve unstable links or partitions first, then observe object state and resynchronization.
Insufficient capacity / Resynchronization faultsReview protection policies, object layout, free space, resynchronization queues and maintenance headroom.Assess expansion, migration and capacity release to restore the resources required for resynchronization.
Inaccessible objects / VM read-write errorsCheck object consistency, component state, host maintenance, disk faults, network partitions and relevant logs.Preserve evidence and review replicas/backups first; coordinate vendor support where necessary before repair.

Repair workflow

Assess, preserve evidence, diagnose, repair with authorization and validate business
  1. 01Intake and impact assessmentIdentify affected clusters, hosts, objects and VMs.
  2. 02Logs / State preservation / Backup reviewPreserve evidence and verify available replicas, backups and recovery conditions.
  3. 03Root cause and repair planSeparate hardware, network, capacity, policy and version factors, documenting known risks.
  4. 04Repair in an authorized windowReplace, repair or rebuild under the approved plan and change authorization.
  5. 05Business validation / HandoverValidate health, object consistency and business reads/writes, then hand over recommendations.
Repair deliverables

Deliver repair records, root causes, known risks, health checks, object-consistency and business read/write validation, and follow-up recommendations.

Why it matters / Suitable scenarios

Turn local disks into a policy-managed resource pool

As VMs grow, shared-storage expansion becomes constrained or ownership of nodes, disks, networking and maintenance is fragmented, vSAN can combine supported local devices into a virtualized storage view. Suitability depends on hardware, versions, networking, capacity and maintenance windows, not node count alone.

A

Consolidate server-local disks

Bring supported local disks into a distributed datastore, reducing dependence on one host's storage.

B

Expand virtualization clusters

New nodes contribute compute and capacity, after checking policy, networking, failure domains and rebuild headroom.

C

Observe maintenance and failures

Include object state, node health, synchronization and repair in shared inspection and handover records.

From isolated drives to shared storage

vSAN illustration showing isolated hosts and disks consolidated into shared storage
Conceptual consolidation diagram showing the relationship between hosts, networking and capacity.Open full-size diagram
Inventory current resources
Inventory hosts, local disks, workloads and growth before deciding what equipment can be retained.
Define cluster boundaries
Plan storage traffic, protection policies and failure domains, then define independent backup and recovery objectives.
Validate workload fit
Use compatibility, networking and usable capacity to confirm which workloads fit the cluster.

Core capabilities

Design capacity, policies and failure boundaries together

Resource poolHosts, disks, controllers and support boundaries
Storage policiesFTT, RAID/protection methods and object placement
Failure domainsNodes, racks, switches and maintenance impact
Capacity accountingProtection overhead, rebuild, maintenance and growth headroom
Expansion and rebalancingNew-node capacity and object movement
Operations handoverAlerts, inspections, changes and responsibilities

Capacity is not the sum of raw disks, and availability does not replace backups. Review policies and device types separately for OSA and ESA.

Check cluster capacity and workload

Illustrative cluster-management interface showing hosts, VMs, resource pools and utilization
Illustrative interface with sample data. Validate against measured capacity and policy records from the actual cluster.Open full-size diagram
Calculate usable capacity
Account for protection overhead and plan for rebuilds, maintenance, snapshots and growth.
Measure the resource baseline
Review CPU, memory, storage utilization and latency at business peaks, then document expansion thresholds.
Monitor policy compliance
Use the actual cluster to check object policies, headroom and resynchronization, then define alert and change procedures.

How Yuqi Intelligent implements

Turn compatibility and failure boundaries into an executable checklist

Start with assessment, combining ReadyNode/hardware compatibility, networks, disks, policies, migration windows and failure exercises in one delivery path instead of adding capacity and ownership records after deployment.

01ReadyNode and versions

Check supported combinations of hosts, controllers, disks, firmware, drivers and vSAN versions.

0210 / 25GbE design

Choose speed, latency, redundancy and traffic separation by workload, version, ReadyNode certification and network conditions. Do not treat 10GbE as a universal ESA prerequisite.

03Disks, policies and capacity

Distinguish OSA disk groups from ESA single-tier NVMe pools, calculating protection overhead, rebuild space and growth headroom.

04Migration and failure exercises

Pilot and retain independent backups before validating maintenance, offline nodes, rebuilding, alerts, business access and rollback windows.

Typical topology

Separate node, dual-switch and network responsibilities

The topology illustrates relationships, not port, VLAN, MTU, link-aggregation or version-support requirements. Define management, vMotion and vSAN network isolation and redundancy for each project.

SW-AUplinks / Redundancy
SW-BUplinks / Redundancy
Management networkvMotion networkvSAN network
3 nodesRAID1 / FTT1 exampleVerify witness and failure-domain requirements
4 nodesRebuild in remaining spaceA node failure does not mean immediate restoration of all redundancy
Multi-node expansionCompute and capacity grow togetherObserve rebalancing and maintenance headroom after joining

vSAN cluster and policy relationships

Illustrative vSAN topology showing virtual machines, policies, hosts, networking and shared storage
Conceptual cluster topology. Device roles, placement and failure domains depend on the chosen architecture and version.Open full-size diagram
Hosts and datastore
Record cluster membership, storage devices and network paths, and validate device roles for the actual OSA or ESA design.
Policies and failure domains
Select policies around protection requirements and check that replicas or erasure-coded components can be placed across the required failure domains.
Test failure and recovery
Test host outages, maintenance, rebuilding and recovery, and retain traceable results.

Delivery and operations handover

Deliver a cluster the operations team can maintain

01Cluster and configuration baseline

Host, disk, network, version, license and management-access records.

02Policy and capacity planning

Protection policies, failure domains, capacity accounting, reserves and expansion criteria.

03Performance and failure validation

Representative workloads, reads/writes, links, offline nodes, rebuilds, alerts and rollback records.

04Architecture and operations documentation

Topology, inspections, changes, replacement, backup boundaries, owners and recommendations.

Final deliveryA deployed vSAN cluster and shared datastore, together with policies, test records and operations documentation.

Where the project also involves VMware Server Virtualization compute resources or Data Backup for independent recovery, documentation separates dependencies and responsibilities.

Operations monitoring and handover

Illustrative cluster health view showing hosts, capacity, alarms and recent tasks
Illustrative interface with sample data. Use project monitoring records for actual health, capacity, alarms and changes.Open full-size diagram
Maintain a state inventory
Hand over node, network, disk and policy configurations, along with metrics, alert severity and named owners.
Log incident response
Record symptoms, actions and verification for resynchronization, capacity alerts and disk or host incidents.
Document change procedures
Document expansion, replacement, upgrades, maintenance and rollback conditions for ongoing operations.

View the original vSAN technical reference

Frequently asked questions

Clarify constraints before selecting the architecture

Does vSAN simply add up all local disk capacity?

No. Deduct protection, replication or erasure coding, object layout, metadata, rebuilding and maintenance reserves. Verify usable capacity against architecture, policies, devices and growth assumptions.

Do OSA and ESA use the same disk structure?

No. OSA commonly uses disk groups with SSD cache and SSD/HDD capacity tiers. ESA uses supported single-tier NVMe pools; do not carry dedicated cache and HDD tiers into ESA. Verify versions, certified hardware and project requirements.

What can three-node or four-node clusters support?

Three-node RAID1/FTT1 is an example, subject to version-specific node, witness, failure-domain and policy requirements. Four-node RAID1 may rebuild in remaining space after a node fails; this is not immediate restoration of all redundancy.

Can vSAN replace backups?

No. Policies and cluster availability are not independent backups or recovery. VM restart normally involves capabilities such as vSphere HA; backups, retention and recovery exercises still need separate design.

Can you take over a vSAN cluster deployed by another team?

Start with a takeover assessment of versions, hardware, firmware/drivers, disk groups, networking, policies, capacity, logs and backups. Define repair, migration or expansion boundaries before changing production.

What is checked first during vSAN fault diagnosis?

Start with symptoms and impact, then check cluster health, offline hosts, disks/groups, firmware/drivers, network latency/loss/partitions, capacity, resynchronization, inaccessible objects and VM read/write errors. Preserve evidence and review copies/backups before agreeing treatment.

Under what conditions can repair or rebuilding proceed?

Confirm the authorized window, remaining capacity, quorum/policies, available replicas and business-validation path before selecting repair, replacement or rebuild. Actions depend on versions and risk assessment; lossless or immediate recovery is not guaranteed.

Next step / Storage assessment

Start with existing hosts, disks and business windows.

Share nodes, disks, networking, VM priorities and migration timing to assess OSA/ESA boundaries, policies and validation scope.

Talk to a technical adviser

SCOPE & HANDOVER

VMware vSAN: scoping and handover

For vSAN deployment, provide ESXi and vCenter versions, nodes, disks, networking, licences and workloads. Confirm architecture, hardware compatibility, failure domains and usable capacity before comparing implementation, migration and tests; raw disk totals do not equal usable space.

Planning and prerequisites

What should be confirmed for vmware vsan?

Usable capacity depends on policies, redundancy, metadata, rebuild allowance and growth. Also check versions, hardware compatibility, networking and licences.

Verification and handover

What should be checked at handover?

Review health, object policies, networking and resynchronization. Test agreed host or drive failures and document capacity, policies and recovery observations.