Microsoft.Batch (stable:2024-07-01.20.0)

2025/08/27 • 5 updated methods

Pool_Resize (updated)
Description :  You can only resize a Pool when its allocation state is steady. If the Pool is already resizing, the request fails with status code 409. When you resize a Pool, the Pool's allocation state changes from steady to resizing. You cannot resize Pools which are configured for automatic scaling. If you try to do this, the Batch service returns an error 409. If you resize a Pool downwards, the Batch service chooses which Compute Nodes to remove. To remove specific Compute Nodes, use the Pool remove Compute Nodes API instead.
Reference :  Link ¶

⚶ Changes

{
  "#id": "Pool_Resize",
  "$parameters": {
    "poolResizeParameter": {
      "$properties": [
        {
          "#name": "nodeDeallocationOption",
          "Enum": {
            "new": [
              [
                "requeue",
                "Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Remove Compute Nodes as soon as Tasks have been terminated."
              ],
              [
                "terminate",
                "Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Remove Compute Nodes as soon as Tasks have been terminated."
              ],
              [
                "taskcompletion",
                "Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Remove Compute Nodes when all Tasks have completed."
              ],
              [
                "retaineddata",
                "Deprecated, we encourage you to upload task data to Azure Storage in your task and use `taskcompletion` instead. Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Remove Compute Nodes when all Task retention periods have expired."
              ]
            ],
            "old": [
              [
                "requeue",
                "Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Remove Compute Nodes as soon as Tasks have been terminated."
              ],
              [
                "terminate",
                "Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Remove Compute Nodes as soon as Tasks have been terminated."
              ],
              [
                "taskcompletion",
                "Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Remove Compute Nodes when all Tasks have completed."
              ],
              [
                "retaineddata",
                "Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Remove Compute Nodes when all Task retention periods have expired."
              ]
            ]
          }
        }
      ]
    }
  }
}

⚼ Request

POST:  /pools/{poolId}/resize
{
poolId:
{
Description: The ID of the Pool to resize. ,
Required: True ,
}
string ,
poolResizeParameter:
{
Description: The parameters for the request. ,
Required: True ,
}
{
targetDedicatedNodes:
{
Format: int32 ,
Required: False ,
}
integer ,
targetLowPriorityNodes:
{
Format: int32 ,
Required: False ,
}
integer ,
resizeTimeout:
{
Description: The default value is 15 minutes. The minimum value is 5 minutes. If you specify a value less than 5 minutes, the Batch service returns an error; if you are calling the REST API directly, the HTTP status code is 400 (Bad Request). ,
Format: duration ,
Required: False ,
}
string ,
nodeDeallocationOption:
{
Description: The default value is requeue. ,
Enum:
{
requeue: Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Remove Compute Nodes as soon as Tasks have been terminated. ,
terminate: Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Remove Compute Nodes as soon as Tasks have been terminated. ,
taskcompletion: Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Remove Compute Nodes when all Tasks have completed. ,
retaineddata: Deprecated, we encourage you to upload task data to Azure Storage in your task and use `taskcompletion` instead. Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Remove Compute Nodes when all Task retention periods have expired. ,
}
,
Required: False ,
}
enum ,
}
,
timeout:
{
Description: The maximum time that the server can spend processing the request, in seconds. The default is 30 seconds. If the value is larger than 30, the default will be used instead. ,
Format: int32 ,
Required: False ,
}
integer ,
client-request-id:
{
Description: The caller-generated request identity, in the form of a GUID with no decoration such as curly braces, e.g. 9C4D50EE-2D56-4CD3-8152-34347DC9F2B0. ,
Format: uuid ,
Required: False ,
}
string ,
return-client-request-id:
{
Description: Whether the server should return the client-request-id in the response. ,
Required: False ,
}
boolean ,
ocp-date:
{
Description: The time the request was issued. Client libraries typically set this to the current system clock time; set it explicitly if you are calling the REST API directly. ,
Format: date-time-rfc1123 ,
Required: False ,
}
string ,
If-Match:
{
Description: An ETag value associated with the version of the resource known to the client. The operation will be performed only if the resource's current ETag on the service exactly matches the value specified by the client. ,
Required: False ,
}
string ,
If-None-Match:
{
Description: An ETag value associated with the version of the resource known to the client. The operation will be performed only if the resource's current ETag on the service does not match the value specified by the client. ,
Required: False ,
}
string ,
If-Modified-Since:
{
Description: A timestamp indicating the last modified time of the resource known to the client. The operation will be performed only if the resource on the service has been modified since the specified time. ,
Format: date-time-rfc1123 ,
Required: False ,
}
string ,
If-Unmodified-Since:
{
Description: A timestamp indicating the last modified time of the resource known to the client. The operation will be performed only if the resource on the service has not been modified since the specified time. ,
Format: date-time-rfc1123 ,
Required: False ,
}
string ,
api-version:
{
Description: Client API Version. ,
Required: True ,
}
string ,
}

⚐ Response (202)

{
client-request-id:
{
Description: The client-request-id provided by the client during the request. This will be returned only if the return-client-request-id parameter was set to true. ,
}
string ,
request-id:
{
Description: A unique identifier for the request that was made to the Batch service. If a request is consistently failing and you have verified that the request is properly formulated, you may use this value to report the error to Microsoft. In your report, include the value of this request ID, the approximate time that the request was made, the Batch Account against which the request was made, and the region that Account resides in. ,
}
string ,
etag:
{
Description: The ETag HTTP response header. This is an opaque string. You can use it to detect whether the resource has changed between requests. In particular, you can pass the ETag to one of the If-Modified-Since, If-Unmodified-Since, If-Match or If-None-Match headers. ,
}
string ,
last-modified:
{
Description: The time at which the resource was last modified. ,
}
string ,
dataserviceid:
{
Description: The OData ID of the resource to which the request applied. ,
}
string ,
}

⚐ Response (default)

{
code:
{
Required: False ,
}
string ,
message:
{
Required: False ,
}
{
lang:
{
Required: False ,
}
string ,
value:
{
Required: False ,
}
string ,
}
,
values:
{
Required: False ,
}
[
{
key:
{
Required: False ,
}
string ,
value:
{
Required: False ,
}
string ,
}
,
]
,
}
Pool_RemoveNodes (updated)
Description :  This operation can only run when the allocation state of the Pool is steady. When this operation runs, the allocation state changes from steady to resizing. Each request may remove up to 100 nodes.
Reference :  Link ¶

⚶ Changes

{
  "#id": "Pool_RemoveNodes",
  "$parameters": {
    "nodeRemoveParameter": {
      "$properties": [
        {
          "#name": "nodeDeallocationOption",
          "Enum": {
            "new": [
              [
                "requeue",
                "Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Remove Compute Nodes as soon as Tasks have been terminated."
              ],
              [
                "terminate",
                "Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Remove Compute Nodes as soon as Tasks have been terminated."
              ],
              [
                "taskcompletion",
                "Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Remove Compute Nodes when all Tasks have completed."
              ],
              [
                "retaineddata",
                "Deprecated, we encourage you to upload task data to Azure Storage in your task and use `taskcompletion` instead. Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Remove Compute Nodes when all Task retention periods have expired."
              ]
            ],
            "old": [
              [
                "requeue",
                "Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Remove Compute Nodes as soon as Tasks have been terminated."
              ],
              [
                "terminate",
                "Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Remove Compute Nodes as soon as Tasks have been terminated."
              ],
              [
                "taskcompletion",
                "Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Remove Compute Nodes when all Tasks have completed."
              ],
              [
                "retaineddata",
                "Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Remove Compute Nodes when all Task retention periods have expired."
              ]
            ]
          }
        }
      ]
    }
  }
}

⚼ Request

POST:  /pools/{poolId}/removenodes
{
poolId:
{
Description: The ID of the Pool from which you want to remove Compute Nodes. ,
Required: True ,
}
string ,
nodeRemoveParameter:
{
Description: The parameters for the request. ,
Required: True ,
}
{
nodeList:
{
Description: A maximum of 100 nodes may be removed per request. ,
Required: True ,
}
[
string ,
]
,
resizeTimeout:
{
Description: The default value is 15 minutes. The minimum value is 5 minutes. If you specify a value less than 5 minutes, the Batch service returns an error; if you are calling the REST API directly, the HTTP status code is 400 (Bad Request). ,
Format: duration ,
Required: False ,
}
string ,
nodeDeallocationOption:
{
Description: The default value is requeue. ,
Enum:
{
requeue: Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Remove Compute Nodes as soon as Tasks have been terminated. ,
terminate: Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Remove Compute Nodes as soon as Tasks have been terminated. ,
taskcompletion: Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Remove Compute Nodes when all Tasks have completed. ,
retaineddata: Deprecated, we encourage you to upload task data to Azure Storage in your task and use `taskcompletion` instead. Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Remove Compute Nodes when all Task retention periods have expired. ,
}
,
Required: False ,
}
enum ,
}
,
timeout:
{
Description: The maximum time that the server can spend processing the request, in seconds. The default is 30 seconds. If the value is larger than 30, the default will be used instead. ,
Format: int32 ,
Required: False ,
}
integer ,
client-request-id:
{
Description: The caller-generated request identity, in the form of a GUID with no decoration such as curly braces, e.g. 9C4D50EE-2D56-4CD3-8152-34347DC9F2B0. ,
Format: uuid ,
Required: False ,
}
string ,
return-client-request-id:
{
Description: Whether the server should return the client-request-id in the response. ,
Required: False ,
}
boolean ,
ocp-date:
{
Description: The time the request was issued. Client libraries typically set this to the current system clock time; set it explicitly if you are calling the REST API directly. ,
Format: date-time-rfc1123 ,
Required: False ,
}
string ,
If-Match:
{
Description: An ETag value associated with the version of the resource known to the client. The operation will be performed only if the resource's current ETag on the service exactly matches the value specified by the client. ,
Required: False ,
}
string ,
If-None-Match:
{
Description: An ETag value associated with the version of the resource known to the client. The operation will be performed only if the resource's current ETag on the service does not match the value specified by the client. ,
Required: False ,
}
string ,
If-Modified-Since:
{
Description: A timestamp indicating the last modified time of the resource known to the client. The operation will be performed only if the resource on the service has been modified since the specified time. ,
Format: date-time-rfc1123 ,
Required: False ,
}
string ,
If-Unmodified-Since:
{
Description: A timestamp indicating the last modified time of the resource known to the client. The operation will be performed only if the resource on the service has not been modified since the specified time. ,
Format: date-time-rfc1123 ,
Required: False ,
}
string ,
api-version:
{
Description: Client API Version. ,
Required: True ,
}
string ,
}

⚐ Response (202)

{
client-request-id:
{
Description: The client-request-id provided by the client during the request. This will be returned only if the return-client-request-id parameter was set to true. ,
}
string ,
request-id:
{
Description: A unique identifier for the request that was made to the Batch service. If a request is consistently failing and you have verified that the request is properly formulated, you may use this value to report the error to Microsoft. In your report, include the value of this request ID, the approximate time that the request was made, the Batch Account against which the request was made, and the region that Account resides in. ,
}
string ,
etag:
{
Description: The ETag HTTP response header. This is an opaque string. You can use it to detect whether the resource has changed between requests. In particular, you can pass the ETag to one of the If-Modified-Since, If-Unmodified-Since, If-Match or If-None-Match headers. ,
}
string ,
last-modified:
{
Description: The time at which the resource was last modified. ,
}
string ,
dataserviceid:
{
Description: The OData ID of the resource to which the request applied. ,
}
string ,
}

⚐ Response (default)

{
code:
{
Required: False ,
}
string ,
message:
{
Required: False ,
}
{
lang:
{
Required: False ,
}
string ,
value:
{
Required: False ,
}
string ,
}
,
values:
{
Required: False ,
}
[
{
key:
{
Required: False ,
}
string ,
value:
{
Required: False ,
}
string ,
}
,
]
,
}
ComputeNode_Reboot (updated)
Description :  You can restart a Compute Node only if it is in an idle or running state.
Reference :  Link ¶

⚶ Changes

{
  "#id": "ComputeNode_Reboot",
  "$parameters": {
    "nodeRebootParameter": {
      "$properties": [
        {
          "#name": "nodeRebootOption",
          "Enum": {
            "new": [
              [
                "requeue",
                "Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Restart the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "terminate",
                "Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Restart the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "taskcompletion",
                "Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Restart the Compute Node when all Tasks have completed."
              ],
              [
                "retaineddata",
                "Deprecated, we encourage you to upload task data to Azure Storage in your task and use `taskcompletion` instead. Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Restart the Compute Node when all Task retention periods have expired."
              ]
            ],
            "old": [
              [
                "requeue",
                "Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Restart the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "terminate",
                "Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Restart the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "taskcompletion",
                "Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Restart the Compute Node when all Tasks have completed."
              ],
              [
                "retaineddata",
                "Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Restart the Compute Node when all Task retention periods have expired."
              ]
            ]
          }
        }
      ]
    }
  }
}

⚼ Request

POST:  /pools/{poolId}/nodes/{nodeId}/reboot
{
poolId:
{
Description: The ID of the Pool that contains the Compute Node. ,
Required: True ,
}
string ,
nodeId:
{
Description: The ID of the Compute Node that you want to restart. ,
Required: True ,
}
string ,
nodeRebootParameter:
{
Description: The parameters for the request. ,
Required: False ,
}
{
nodeRebootOption:
{
Description: The default value is requeue. ,
Enum:
{
requeue: Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Restart the Compute Node as soon as Tasks have been terminated. ,
terminate: Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Restart the Compute Node as soon as Tasks have been terminated. ,
taskcompletion: Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Restart the Compute Node when all Tasks have completed. ,
retaineddata: Deprecated, we encourage you to upload task data to Azure Storage in your task and use `taskcompletion` instead. Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Restart the Compute Node when all Task retention periods have expired. ,
}
,
Required: False ,
}
enum ,
}
,
timeout:
{
Description: The maximum time that the server can spend processing the request, in seconds. The default is 30 seconds. If the value is larger than 30, the default will be used instead. ,
Format: int32 ,
Required: False ,
}
integer ,
client-request-id:
{
Description: The caller-generated request identity, in the form of a GUID with no decoration such as curly braces, e.g. 9C4D50EE-2D56-4CD3-8152-34347DC9F2B0. ,
Format: uuid ,
Required: False ,
}
string ,
return-client-request-id:
{
Description: Whether the server should return the client-request-id in the response. ,
Required: False ,
}
boolean ,
ocp-date:
{
Description: The time the request was issued. Client libraries typically set this to the current system clock time; set it explicitly if you are calling the REST API directly. ,
Format: date-time-rfc1123 ,
Required: False ,
}
string ,
api-version:
{
Description: Client API Version. ,
Required: True ,
}
string ,
}

⚐ Response (202)

{
client-request-id:
{
Description: The client-request-id provided by the client during the request. This will be returned only if the return-client-request-id parameter was set to true. ,
}
string ,
request-id:
{
Description: A unique identifier for the request that was made to the Batch service. If a request is consistently failing and you have verified that the request is properly formulated, you may use this value to report the error to Microsoft. In your report, include the value of this request ID, the approximate time that the request was made, the Batch Account against which the request was made, and the region that Account resides in. ,
}
string ,
etag:
{
Description: The ETag HTTP response header. This is an opaque string. You can use it to detect whether the resource has changed between requests. In particular, you can pass the ETag to one of the If-Modified-Since, If-Unmodified-Since, If-Match or If-None-Match headers. ,
}
string ,
last-modified:
{
Description: The time at which the resource was last modified. ,
}
string ,
dataserviceid:
{
Description: The OData ID of the resource to which the request applied. ,
}
string ,
}

⚐ Response (default)

{
code:
{
Required: False ,
}
string ,
message:
{
Required: False ,
}
{
lang:
{
Required: False ,
}
string ,
value:
{
Required: False ,
}
string ,
}
,
values:
{
Required: False ,
}
[
{
key:
{
Required: False ,
}
string ,
value:
{
Required: False ,
}
string ,
}
,
]
,
}
ComputeNode_Reimage (updated)
Description :  You can reinstall the operating system on a Compute Node only if it is in an idle or running state.
Reference :  Link ¶

⚶ Changes

{
  "#id": "ComputeNode_Reimage",
  "$parameters": {
    "nodeReimageParameter": {
      "$properties": [
        {
          "#name": "nodeReimageOption",
          "Enum": {
            "new": [
              [
                "requeue",
                "Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Reimage the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "terminate",
                "Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Reimage the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "taskcompletion",
                "Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Reimage the Compute Node when all Tasks have completed."
              ],
              [
                "retaineddata",
                "Deprecated, we encourage you to upload task data to Azure Storage in your task and use `taskcompletion` instead. Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Reimage the Compute Node when all Task retention periods have expired."
              ]
            ],
            "old": [
              [
                "requeue",
                "Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Reimage the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "terminate",
                "Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Reimage the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "taskcompletion",
                "Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Reimage the Compute Node when all Tasks have completed."
              ],
              [
                "retaineddata",
                "Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Reimage the Compute Node when all Task retention periods have expired."
              ]
            ]
          }
        }
      ]
    }
  }
}

⚼ Request

POST:  /pools/{poolId}/nodes/{nodeId}/reimage
{
poolId:
{
Description: The ID of the Pool that contains the Compute Node. ,
Required: True ,
}
string ,
nodeId:
{
Description: The ID of the Compute Node that you want to restart. ,
Required: True ,
}
string ,
nodeReimageParameter:
{
Description: The parameters for the request. ,
Required: False ,
}
{
nodeReimageOption:
{
Description: The default value is requeue. ,
Enum:
{
requeue: Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Reimage the Compute Node as soon as Tasks have been terminated. ,
terminate: Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Reimage the Compute Node as soon as Tasks have been terminated. ,
taskcompletion: Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Reimage the Compute Node when all Tasks have completed. ,
retaineddata: Deprecated, we encourage you to upload task data to Azure Storage in your task and use `taskcompletion` instead. Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Reimage the Compute Node when all Task retention periods have expired. ,
}
,
Required: False ,
}
enum ,
}
,
timeout:
{
Description: The maximum time that the server can spend processing the request, in seconds. The default is 30 seconds. If the value is larger than 30, the default will be used instead. ,
Format: int32 ,
Required: False ,
}
integer ,
client-request-id:
{
Description: The caller-generated request identity, in the form of a GUID with no decoration such as curly braces, e.g. 9C4D50EE-2D56-4CD3-8152-34347DC9F2B0. ,
Format: uuid ,
Required: False ,
}
string ,
return-client-request-id:
{
Description: Whether the server should return the client-request-id in the response. ,
Required: False ,
}
boolean ,
ocp-date:
{
Description: The time the request was issued. Client libraries typically set this to the current system clock time; set it explicitly if you are calling the REST API directly. ,
Format: date-time-rfc1123 ,
Required: False ,
}
string ,
api-version:
{
Description: Client API Version. ,
Required: True ,
}
string ,
}

⚐ Response (202)

{
client-request-id:
{
Description: The client-request-id provided by the client during the request. This will be returned only if the return-client-request-id parameter was set to true. ,
}
string ,
request-id:
{
Description: A unique identifier for the request that was made to the Batch service. If a request is consistently failing and you have verified that the request is properly formulated, you may use this value to report the error to Microsoft. In your report, include the value of this request ID, the approximate time that the request was made, the Batch Account against which the request was made, and the region that Account resides in. ,
}
string ,
etag:
{
Description: The ETag HTTP response header. This is an opaque string. You can use it to detect whether the resource has changed between requests. In particular, you can pass the ETag to one of the If-Modified-Since, If-Unmodified-Since, If-Match or If-None-Match headers. ,
}
string ,
last-modified:
{
Description: The time at which the resource was last modified. ,
}
string ,
dataserviceid:
{
Description: The OData ID of the resource to which the request applied. ,
}
string ,
}

⚐ Response (default)

{
code:
{
Required: False ,
}
string ,
message:
{
Required: False ,
}
{
lang:
{
Required: False ,
}
string ,
value:
{
Required: False ,
}
string ,
}
,
values:
{
Required: False ,
}
[
{
key:
{
Required: False ,
}
string ,
value:
{
Required: False ,
}
string ,
}
,
]
,
}
ComputeNode_Deallocate (updated)
Description :  You can deallocate a Compute Node only if it is in an idle or running state.
Reference :  Link ¶

⚶ Changes

{
  "#id": "ComputeNode_Deallocate",
  "$parameters": {
    "nodeDeallocateParameter": {
      "$properties": [
        {
          "#name": "nodeDeallocateOption",
          "Enum": {
            "new": [
              [
                "requeue",
                "Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Deallocate the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "terminate",
                "Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Deallocate the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "taskcompletion",
                "Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Deallocate the Compute Node when all Tasks have completed."
              ],
              [
                "retaineddata",
                "Deprecated, we encourage you to upload task data to Azure Storage in your task and use `taskcompletion` instead. Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Deallocate the Compute Node when all Task retention periods have expired."
              ]
            ],
            "old": [
              [
                "requeue",
                "Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Deallocate the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "terminate",
                "Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Deallocate the Compute Node as soon as Tasks have been terminated."
              ],
              [
                "taskcompletion",
                "Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Deallocate the Compute Node when all Tasks have completed."
              ],
              [
                "retaineddata",
                "Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Deallocate the Compute Node when all Task retention periods have expired."
              ]
            ]
          }
        }
      ]
    }
  }
}

⚼ Request

POST:  /pools/{poolId}/nodes/{nodeId}/deallocate
{
poolId:
{
Description: The ID of the Pool that contains the Compute Node. ,
Required: True ,
}
string ,
nodeId:
{
Description: The ID of the Compute Node that you want to deallocate. ,
Required: True ,
}
string ,
nodeDeallocateParameter:
{
Description: The parameters for the request. ,
Required: False ,
}
{
nodeDeallocateOption:
{
Description: The default value is requeue. ,
Enum:
{
requeue: Terminate running Task processes and requeue the Tasks. The Tasks will run again when a Compute Node is available. Deallocate the Compute Node as soon as Tasks have been terminated. ,
terminate: Terminate running Tasks. The Tasks will be completed with failureInfo indicating that they were terminated, and will not run again. Deallocate the Compute Node as soon as Tasks have been terminated. ,
taskcompletion: Allow currently running Tasks to complete. Schedule no new Tasks while waiting. Deallocate the Compute Node when all Tasks have completed. ,
retaineddata: Deprecated, we encourage you to upload task data to Azure Storage in your task and use `taskcompletion` instead. Allow currently running Tasks to complete, then wait for all Task data retention periods to expire. Schedule no new Tasks while waiting. Deallocate the Compute Node when all Task retention periods have expired. ,
}
,
Required: False ,
}
enum ,
}
,
timeout:
{
Description: The maximum time that the server can spend processing the request, in seconds. The default is 30 seconds. If the value is larger than 30, the default will be used instead. ,
Format: int32 ,
Required: False ,
}
integer ,
client-request-id:
{
Description: The caller-generated request identity, in the form of a GUID with no decoration such as curly braces, e.g. 9C4D50EE-2D56-4CD3-8152-34347DC9F2B0. ,
Format: uuid ,
Required: False ,
}
string ,
return-client-request-id:
{
Description: Whether the server should return the client-request-id in the response. ,
Required: False ,
}
boolean ,
ocp-date:
{
Description: The time the request was issued. Client libraries typically set this to the current system clock time; set it explicitly if you are calling the REST API directly. ,
Format: date-time-rfc1123 ,
Required: False ,
}
string ,
api-version:
{
Description: Client API Version. ,
Required: True ,
}
string ,
}

⚐ Response (202)

{
client-request-id:
{
Description: The client-request-id provided by the client during the request. This will be returned only if the return-client-request-id parameter was set to true. ,
}
string ,
request-id:
{
Description: A unique identifier for the request that was made to the Batch service. If a request is consistently failing and you have verified that the request is properly formulated, you may use this value to report the error to Microsoft. In your report, include the value of this request ID, the approximate time that the request was made, the Batch Account against which the request was made, and the region that Account resides in. ,
}
string ,
etag:
{
Description: The ETag HTTP response header. This is an opaque string. You can use it to detect whether the resource has changed between requests. In particular, you can pass the ETag to one of the If-Modified-Since, If-Unmodified-Since, If-Match or If-None-Match headers. ,
}
string ,
last-modified:
{
Description: The time at which the resource was last modified. ,
}
string ,
dataserviceid:
{
Description: The OData ID of the resource to which the request applied. ,
}
string ,
}

⚐ Response (default)

{
code:
{
Required: False ,
}
string ,
message:
{
Required: False ,
}
{
lang:
{
Required: False ,
}
string ,
value:
{
Required: False ,
}
string ,
}
,
values:
{
Required: False ,
}
[
{
key:
{
Required: False ,
}
string ,
value:
{
Required: False ,
}
string ,
}
,
]
,
}