संपादित करें

sys.dm_os_linux_vm_stats (Transact-SQL)

Applies to: SQL Server 2025 (17.x) on Linux

Returns Linux operating system-level virtual memory statistics, including metrics related to SQL Server and other processes running on the system, in SQL Server 2025 (17.x) Cumulative Update (CU) 1 and later versions.

Column name Data type Nullable Description
vm_metric_name nvarchar(256) No Virtual memory metric name.
count bigint No Corresponding statistic for that metric.

Permissions

Requires VIEW SERVER PERFORMANCE STATE permission on the server.

Remarks

sys.dm_os_linux_vm_stats provides system-wide memory observability to help you analyze memory pressure, page faults, reclaim activity, non-uniform memory access (NUMA) behavior, and to correlate SQL Server performance with overall OS memory health.

Each row represents a single virtual memory metric, typically sourced from Linux interfaces such as /proc/vmstat. Metric availability and meaning can vary by Linux distribution, kernel version, and configuration.

Use this DMV with other Linux-specific DMVs for holistic monitoring:

Usage scenarios

Common scenarios for using sys.dm_os_linux_vm_stats include:

  • Investigating out-of-memory (OOM) events on Linux hosts.
  • Correlating SQL Server memory symptoms with OS reclaim activity.
  • Understanding NUMA-related memory behavior on multinode systems.
  • Performing deep OS-level observability directly from Transact-SQL.

Examples

A. View all virtual memory statistics

The following query returns all available virtual memory metrics reported by the Linux kernel:

SELECT *
FROM sys.dm_os_linux_vm_stats;

B. Identify page fault activity

The following query highlights page fault-related metrics, which can help identify memory pressure or inefficient memory access patterns:

SELECT vm_metric_name,
       count
FROM sys.dm_os_linux_vm_stats
WHERE vm_metric_name IN ('pgfault', 'pgmajfault');

C. Monitor locality of NUMA memory

Returns NUMA-related virtual memory metrics to help understand memory locality across nodes:

SELECT vm_metric_name,
       count
FROM sys.dm_os_linux_vm_stats
WHERE vm_metric_name LIKE 'numa%';

D. Analyze activity for memory reclaim and compaction

The following query helps you diagnose memory reclaim behavior and compaction pressure on the system:

SELECT vm_metric_name,
       count
FROM sys.dm_os_linux_vm_stats
WHERE vm_metric_name LIKE 'pgsteal%'
      OR vm_metric_name LIKE 'pgscan%'
      OR vm_metric_name LIKE 'compact%';

Related content

  • Last updated on