vSAN witness recovery

A vSAN witness failure becomes a data-availability incident when another fault removes object quorum.

In two-node and stretched vSAN designs, witness components provide votes rather than normal data capacity. A witness outage alone may be tolerated, but a simultaneous site, host, disk-group or network fault can make objects inaccessible.

Understand the symptom

vSAN availability depends on a majority of votes for each object, not simply on visible data components.

WHY

What is happening

An object can have readable data components yet remain inaccessible without quorum. Improper witness replacement can introduce a new identity while objects still reference the old witness components. Force provisioning may create reduced-redundancy objects and is not a recovery substitute for restoring the original vote set.

What you may see

Symptoms that change the recovery path.

01

Objects show inaccessible

The object lacks enough active votes for quorum.

02

VM cannot power on

vSAN cannot create or access swap and namespace objects.

03

Witness is partitioned or down

The tie-breaking fault domain is absent.

04

Second failure follows witness loss

A host, site, cache device or disk group removes additional votes.

Protect the source

Actions that can make the situation worse.

01

Improper replacement changes witness identity

Existing objects may remain tied to inaccessible witness components.

02

Force provisioning reduces protection

New objects may be created without the intended redundancy.

03

Policy changes can trigger resynchronization

Large writes can alter the failure state before evidence is understood.

Do this now

Safe steps before evaluation.

01

Stop placement and policy changes

Pause evacuation, resync-heavy work and new VM creation.

02

Capture object health

Record inaccessible UUIDs, component layouts, votes, policies and fault domains.

03

Preserve logs and identities

Keep vCenter, ESXi and witness logs plus cluster and host UUIDs.

04

Stabilize networking and power

Avoid rebooting multiple fault domains while quorum is unresolved.

NEPFOSYS process

How a data-first evaluation proceeds.

01

Quorum reconstruction

NEPFOSYS maps object votes, witness identity, sites, hosts and disk groups.

02

Least-change recovery plan

Network restoration, original witness recovery and supported replacement paths are prioritized.

03

Object validation

VM namespaces, virtual disks and application data are checked after access returns.

A clear next step

Preserve the original device before another repair or recovery attempt.

Start a case with the exact symptoms, incident timeline, prior attempts, encryption details, and the files that matter most. NEPFOSYS will route the case to the appropriate DFW intake or nationwide mail-in path.

Start with a case review

Questions, answered

What customers ask before starting.

Does a vSAN witness store normal VM data?

Its primary role is voting metadata, but those votes can be essential to object quorum.

Why are data components present but inaccessible?

vSAN requires a majority of object votes to protect consistency, not merely a readable component.

Should I enable force provisioning?

Not as a first recovery step. It can create reduced-availability objects and does not restore missing quorum to existing data.

DFW local · Nationwide mail-in

Protect the device. Start with the facts.

Start my recovery case