DB STARTUP events in AG replica SQL error log

Question

HI Team,

we got an Availability group with 4 nodes (1 primary, 2 synch nodes, 1 asynch node).

in the secondary replica sql error logs I see frequent messages as below for various secondary databases.

Example:

Buffer Pool scan took 11 seconds: database ID 5, command 'DB STARTUP', operation 'FlushCache', scanned buffers 80175186, total iterated buffers 103125903, wait time 4 ms. See 'https://go.microsoft.com/fwlink/?linkid=2132602' for more information.

I wonder first of all why 'DB STARTUP' events are happening in secondary and what cause it. I understand the fact the buffer pool scan occur when DB Recovery and some of the similar events occur.

SQL Server 2019 :

Microsoft SQL Server 2019 (RTM-CU32-GDR) (KB5063757) - 15.0.4440.1 (X64) Jul 14 2025 12:59:40 Copyright (C) 2019 Microsoft Corporation Enterprise Edition: Core-based Licensing (64-bit) on Windows Server 2022 Standard 10.0 <X64> (Build 20348: )

Answer 1

Hi @subhash-DBA,

What you’re seeing in the error log is expected behavior in Availability Groups. Even on secondary replicas, SQL Server runs database recovery phases (redo/undo) to apply log records from the primary. During these transitions, the engine issues a DB STARTUP event and performs a buffer pool scan (FlushCache) to ensure consistency.

A few points to clarify:

Why on secondaries: Recovery is required on all replicas, not just the primary. That’s why you see DB STARTUP messages in the secondary logs.

Buffer pool scan: Large databases can make these scans take several seconds. The message is informational, not an error.

Frequency: If you notice these messages only after service restarts, patching, or role changes, that’s normal. If they occur very frequently, it may indicate replica instability (network interruptions, sync state changes).

Reference: Microsoft added more detailed buffer pool scan logging in recent builds, so these entries are more visible now.

https://learn.microsoft.com/en-us/troubleshoot/sql/database-engine/performance/buffer-pool-scan-runs-slowly-large-memory-machines

Unless you’re seeing repeated recovery cycles every few minutes, these messages don’t indicate a fault. They simply reflect the recovery and synchronization work SQL Server is doing on the secondary.

Thanks,

Lakshmi.