Редагувати

Поділитися через


sys.dm_os_schedulers (Transact-SQL)

Applies to: SQL Server Azure SQL Database Azure SQL Managed Instance Azure Synapse Analytics Analytics Platform System (PDW)

Returns one row per scheduler in SQL Server where each scheduler is mapped to an individual processor. Use this view to monitor the condition of a scheduler or to identify runaway tasks. For more information about schedulers, see the Thread and Task Architecture Guide.

Note

To call this from Azure Synapse Analytics or Analytics Platform System (PDW), use the name sys.dm_pdw_nodes_os_schedulers. This syntax is not supported by serverless SQL pool in Azure Synapse Analytics.

Column name Data type Description
scheduler_address varbinary(8) Memory address of the scheduler. Is not nullable.
parent_node_id int ID of the node that the scheduler belongs to, also known as the parent node. This represents a nonuniform memory access (NUMA) node. Is not nullable.
scheduler_id int ID of the scheduler. All schedulers that are used to run regular queries have ID numbers less than 1048576. Those schedulers that have IDs greater than or equal to 1048576 are used internally by SQL Server, such as the dedicated administrator connection scheduler. Is not nullable.
cpu_id smallint CPU ID assigned to the scheduler.

Is not nullable.

Note: 255 does not indicate no affinity as it did in SQL Server 2005 (9.x). See sys.dm_os_threads (Transact-SQL) for additional affinity information.
status nvarchar(60) Indicates the status of the scheduler. Can be one of the following values:

- HIDDEN ONLINE
- HIDDEN OFFLINE
- VISIBLE ONLINE
- VISIBLE OFFLINE
- VISIBLE ONLINE (DAC)
- HOT_ADDED

Is not nullable.

HIDDEN schedulers are used to process requests that are internal to the Database Engine. VISIBLE schedulers are used to process user requests.

OFFLINE schedulers map to processors that are offline in the affinity mask and are, therefore, not being used to process any requests. ONLINE schedulers map to processors that are online in the affinity mask and are available to process threads.

DAC indicates the scheduler is running under a dedicated administrator connection.

HOT ADDED indicates the schedulers were added in response to a hot add CPU event.
is_online bit If SQL Server is configured to use only some of the available processors on the server, this configuration can mean that some schedulers are mapped to processors that are not in the affinity mask. If that is the case, this column returns 0. This value means that the scheduler is not being used to process queries or batches.

Is not nullable.
is_idle bit 1 = Scheduler is idle. No workers are currently running. Is not nullable.
preemptive_switches_count int Number of times that workers on this scheduler have switched to the preemptive mode.

To execute code that is outside SQL Server (for example, extended stored procedures and distributed queries), a thread has to execute outside the control of the non-preemptive scheduler. To do this, a worker switches to preemptive mode.
context_switches_count int Number of context switches that have occurred on this scheduler. Is not nullable.

To allow for other workers to run, the current running worker has to relinquish control of the scheduler or switch context.

Note: If a worker yields the scheduler and puts itself into the runnable queue and then finds no other workers, the worker will select itself. In this case, the context_switches_count is not updated, but the yield_count is updated.
idle_switches_count int Number of times the scheduler has been waiting for an event while idle. This column is similar to context_switches_count. Is not nullable.
current_tasks_count int Number of current tasks that are associated with this scheduler. This count includes the following:

- Tasks that are waiting for a worker to execute them.
- Tasks that are currently waiting or running (in SUSPENDED or RUNNABLE state).

When a task is completed, this count is decremented. Is not nullable.
runnable_tasks_count int Number of workers, with tasks assigned to them, that are waiting to be scheduled on the runnable queue. Is not nullable.
current_workers_count int Number of workers that are associated with this scheduler. This count includes workers that are not assigned any task. Is not nullable.
active_workers_count int Number of workers that are active. An active worker is never preemptive, must have an associated task, and is either running, runnable, or suspended. Is not nullable.
work_queue_count bigint Number of tasks in the pending queue. These tasks are waiting for a worker to pick them up. Is not nullable.
pending_disk_io_count int Number of pending I/Os that are waiting to be completed. Each scheduler has a list of pending I/Os that are checked to determine whether they have been completed every time there is a context switch. The count is incremented when the request is inserted. This count is decremented when the request is completed. This number does not indicate the state of the I/Os. Is not nullable.
load_factor int Internal value that indicates the perceived load on this scheduler. This value is used to determine whether a new task should be put on this scheduler or another scheduler. This value is useful for debugging purposes when it appears that schedulers are not evenly loaded. The routing decision is made based on the load on the scheduler. SQL Server also uses a load factor of nodes and schedulers to help determine the best location to acquire resources. When a task is enqueued, the load factor is increased. When a task is completed, the load factor is decreased. Using the load factors helps SQL Server OS balance the work load better. Is not nullable.
yield_count int Internal value that is used to indicate progress on this scheduler. This value is used by the Scheduler Monitor to determine whether a worker on the scheduler is not yielding to other workers on time. This value does not indicate that the worker or task transitioned to a new worker. Is not nullable.
last_timer_activity bigint In CPU ticks, the last time that the scheduler timer queue was checked by the scheduler. Is not nullable.
failed_to_create_worker bit Set to 1 if a new worker could not be created on this scheduler. This generally occurs because of memory constraints. Is nullable.
active_worker_address varbinary(8) Memory address of the worker that is currently active. Is nullable. For more information, see sys.dm_os_workers (Transact-SQL).
memory_object_address varbinary(8) Memory address of the scheduler memory object. Not NULLABLE.
task_memory_object_address varbinary(8) Memory address of the task memory object. Is not nullable. For more information, see sys.dm_os_memory_objects (Transact-SQL).
quantum_length_us bigint Identified for informational purposes only. Not supported. Future compatibility is not guaranteed. Exposes the scheduler quantum used by SQLOS.
total_cpu_usage_ms bigint Applies to: SQL Server 2016 (13.x) and later

Total CPU consumed by this scheduler as reported by non-preemptive workers. Is not nullable.
total_cpu_idle_capped_ms bigint Identified for informational purposes only. Not supported. Future compatibility is not guaranteed. Indicates throttling based on Service Level Objective, will always be 0 for non-Azure versions of SQL Server. Is nullable.
total_scheduler_delay_ms bigint Applies to: SQL Server 2016 (13.x) and later

The time between one worker switching out and another one switching in. Can be caused by preemptive workers delaying the scheduling of the next non-preemptive worker, or due to the OS scheduling threads from other processes. Is not nullable.
ideal_workers_limit int Applies to: SQL Server 2019 (15.x) and later

How many workers should ideally be on the scheduler. If the current workers exceed the limit due to imbalanced task load, once they become idle they will be trimmed. Is not nullable.
pdw_node_id int Applies to: Azure Synapse Analytics, Analytics Platform System (PDW)

The identifier for the node that this distribution is on.

Permissions

On SQL Server and SQL Managed Instance, requires VIEW SERVER STATE permission.

On SQL Database Basic, S0, and S1 service objectives, and for databases in elastic pools, the server admin account, the Microsoft Entra admin account, or membership in the ##MS_ServerStateReader## server role is required. On all other SQL Database service objectives, either the VIEW DATABASE STATE permission on the database, or membership in the ##MS_ServerStateReader## server role is required.

Permissions for SQL Server 2022 and later

Requires VIEW SERVER PERFORMANCE STATE permission on the server.

Examples

A. Monitoring hidden and nonhidden schedulers

The following query outputs the state of workers and tasks in SQL Server across all schedulers. This query was executed on a computer system that has the following:

  • Two processors (CPUs)

  • Two (NUMA) nodes

  • One CPU per NUMA node

  • Affinity mask set to 0x03.

SELECT  
    scheduler_id,  
    cpu_id,  
    parent_node_id,  
    current_tasks_count,  
    runnable_tasks_count,  
    current_workers_count,  
    active_workers_count,  
    work_queue_count  
  FROM sys.dm_os_schedulers;  

Here's the result set.

scheduler_id cpu_id parent_node_id current_tasks_count  
------------ ------ -------------- -------------------  
0            1      0              9                    
257          255    0              1                    
1            0      1              10                   
258          255    1              1                    
255          255    32             2                    
  
runnable_tasks_count current_workers_count  
-------------------- ---------------------  
0                    11                     
0                    1                      
0                    18                     
0                    1                      
0                    3                      
  
active_workers_count work_queue_count  
-------------------- --------------------  
6                    0  
1                    0  
8                    0  
1                    0  
1                    0  

The output provides the following information:

  • There are five schedulers. Two schedulers have an ID value < 1048576. Schedulers with ID >= 1048576 are known as hidden schedulers. Scheduler 255 represents the dedicated administrator connection (DAC). There is one DAC scheduler per instance. Resource monitors that coordinate memory pressure use scheduler 257 and scheduler 258, one per NUMA node

  • There are 23 active tasks in the output. These tasks include user requests in addition to resource management tasks that have been started by SQL Server. Examples of SQL Server tasks are RESOURCE MONITOR (one per NUMA node), LAZY WRITER (one per NUMA node), LOCK MONITOR, CHECKPOINT, and LOG WRITER.

  • NUMA node 0 is mapped to CPU 1 and NUMA node 1 is mapped to CPU 0. SQL Server typically starts on a NUMA node other than node 0.

  • With runnable_tasks_count returning 0, there are no actively running tasks. However, active sessions may exist.

  • Scheduler 255 representing DAC has 3 workers associated with it. These workers are allocated at SQL Server startup and do not change. These workers are used to process DAC queries only. The two tasks on this scheduler represent a connection manager and an idle worker.

  • active_workers_count represents all workers that have associated tasks and are running under non-preemptive mode. Some tasks, such as network listeners, run under preemptive scheduling.

  • Hidden schedulers do not process typical user requests. The DAC scheduler is the exception. This DAC scheduler has one thread to process requests.

B. Monitoring nonhidden schedulers in a busy system

The following query shows the state of heavily loaded nonhidden schedulers, where more requests exist than can be handled by available workers. In this example, 256 workers are assigned tasks. Some tasks are waiting for an assignment to a worker. Lower runnable count implies that multiple tasks are waiting for a resource.

Note

You can find the state of workers by querying sys.dm_os_workers. For more information, see sys.dm_os_workers (Transact-SQL).

Here is the query:

SELECT  
    scheduler_id,  
    cpu_id,  
    current_tasks_count,  
    runnable_tasks_count,  
    current_workers_count,  
    active_workers_count,  
    work_queue_count  
  FROM sys.dm_os_schedulers  
  WHERE scheduler_id < 255;  

Here's the result set.

scheduler_id current_tasks_count runnable_tasks_count  
------------ ------------------- --------------------  
0            144                 0                     
1            147                 1                     
  
current_workers_count active_workers_count work_queue_count  
--------------------- -------------------- --------------------  
128                   125                  16  
128                   126                  19  

By comparison, the following result shows multiple runnable tasks where no task is waiting to obtain a worker. The work_queue_count is 0 for both schedulers.

scheduler_id current_tasks_count runnable_tasks_count  
------------ ------------------- --------------------  
0            107                 98                    
1            110                 100                   
  
current_workers_count active_workers_count work_queue_count  
--------------------- -------------------- --------------------  
128                   104                  0  
128                   108                  0  

See also

SQL Server Operating System Related Dynamic Management Views (Transact-SQL)