SQL Server Always On recovery

A suspended SQL Server availability database needs its log chain and replica state preserved before failover or removal.

An availability database can become suspended after an I/O error, a failed redo operation, transaction-log damage, resource pressure or a connectivity interruption. The visible replica state is only part of the incident; the primary and secondary log positions must still be reconciled.

Understand the symptom

Availability Groups move log records, then harden and redo them on each secondary.

WHY

What is happening

SYNCHRONIZED, SYNCHRONIZING, NOT SYNCHRONIZING and SUSPENDED describe different points in that pipeline. Recovery must establish the last hardened and redone log sequence numbers, the health of every database file and log, quorum state, failover history and whether the primary itself is trustworthy. Resuming data movement is appropriate only after the reason for suspension is understood.

What you may see

Symptoms that change the recovery path.

01

Data movement is suspended

The dashboard or dynamic management views show SUSPENDED and the send or redo queues stop progressing.

02

Secondary is not synchronizing

Log blocks are not arriving, hardening or redoing at the expected rate.

03

Redo or I/O errors appear

The secondary cannot apply a log record because of storage, page or file damage.

04

Failover readiness is unclear

Replica roles changed, quorum moved or the primary is unavailable before synchronization was confirmed.

Protect the source

Actions that can make the situation worse.

01

Forced failover can lose transactions

An asynchronous or unsynchronized secondary may not contain the latest committed log records.

02

Re-seeding overwrites evidence

A fresh seed can replace the secondary state that shows where the log stream failed.

03

Resume can repeat damaging I/O

If the underlying volume is unstable, redo can continue stressing or changing it.

Do this now

Safe steps before evaluation.

01

Pause topology changes

Do not remove replicas, force failover, resume repeatedly or initialize new seeds.

02

Capture replica state

Record roles, synchronization health, send and redo queues, LSNs, cluster events and the first SQL error.

03

Preserve every recovery source

Keep database files, transaction logs, log backups, full and differential backups, keys and both replica states.

04

Stabilize the storage layer

Snapshot or image affected volumes and investigate I/O, filesystem, SAN or RAID events before repair.

NEPFOSYS process

How a data-first evaluation proceeds.

01

Timeline and replica mapping

NEPFOSYS correlates failover events, LSN positions, backup chains and the health of each copy.

02

Protected recovery testing

Resume, restore, page-level extraction or controlled reseeding options are evaluated on protected copies.

03

Database validation

Priority tables, consistency and the selected recovery point are verified before production return planning.

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.

Can I safely click Resume Data Movement?

Only after confirming why movement stopped and that the secondary storage and database files are healthy. Repeated resume attempts are not a diagnosis.

Does SUSPENDED mean the primary is corrupt?

Not necessarily. The fault can be on one secondary, in transport, during redo or in shared infrastructure, but every copy should be preserved until the scope is known.

Can a forced failover lose data?

Yes. If the target replica is not synchronized, transactions not yet hardened there may be absent after failover.

DFW local · Nationwide mail-in

Protect the device. Start with the facts.

Start my recovery case