Skip to content

Commit

Permalink
Add docs for reindex data stream REST endpoints (#120653) (#121121)
Browse files Browse the repository at this point in the history
Add documentation for new REST endpoints related to data stream upgrade. 
Endpoints:
- /_migration/reindex
- /_migration/reindex/{index}/_status
- /_migration/reindex/{index}/_cancel
- /_create_from/{source}/{dest}
  • Loading branch information
parkertimmins authored Jan 29, 2025
1 parent ff1893c commit ba7d38a
Show file tree
Hide file tree
Showing 9 changed files with 733 additions and 4 deletions.
142 changes: 142 additions & 0 deletions docs/reference/migration/apis/create-index-from-source.asciidoc
Original file line number Diff line number Diff line change
@@ -0,0 +1,142 @@
[[indices-create-index-from-source]]
=== Create index from source API
++++
<titleabbrev>Create index from source</titleabbrev>
++++

.New API reference
[sidebar]
--
For the most up-to-date API details, refer to {api-es}/group/endpoint-indices[Index APIs].
--

[[indices-create-index-from-source-api-request]]
==== {api-request-title}

`PUT /_create_from/<source>/<dest>`

`POST/_create_from/<source>/<dest>`

[[indices-create-index-from-source-api-prereqs]]
==== {api-prereq-title}

* If the {es} {security-features} are enabled, you must have the `manage`
<<privileges-list-indices,index privilege>> for the index.

[[indices-create-index-from-source-api-desc]]
==== {api-description-title}
This api allows you to add a new index to an {es} cluster, using an existing source index as a basis for the new index.
The settings and mappings from the source index will copied over to the destination index. You can also provide
override settings and mappings which will be combined with the source settings and mappings when creating the
destination index.

[[indices-create-index-from-source-api-path-params]]
==== {api-path-parms-title}

`<source>`::
(Required, string) Name of the existing source index which will be used as a basis.

`<dest>`::
(Required, string) Name of the destination index which will be created.


[role="child_attributes"]
[[indices-create-index-from-source-api-request-body]]
==== {api-request-body-title}

`settings_override`::
(Optional, <<index-modules-settings,index setting object>>) Settings which override the source settings.

`mappings_override`::
(Optional, <<mapping,mapping object>>) Mappings which override the source mappings.

`remove_index_blocks`::
(Optional, boolean) Filter out any index blocks from the source index when creating the destination index.
Defaults to `true`.

[[indices-create-index-from-source-api-example]]
==== {api-examples-title}

Start by creating a source index that we'll copy using this API.

[source,console]
--------------------------------------------------
PUT /my-index
{
"settings": {
"index": {
"number_of_shards": 3,
"blocks.write": true
}
},
"mappings": {
"properties": {
"field1": { "type": "text" }
}
}
}
--------------------------------------------------
// TESTSETUP

Now we create a destination index from the source index. This new index will have the same mappings and settings
as the source index.

[source,console]
--------------------------------------------------
POST _create_from/my-index/my-new-index
--------------------------------------------------


Alternatively, we could override some of the source's settings and mappings. This will use the source settings
and mappings as a basis and combine these with the overrides to create the destination settings and mappings.

[source,console]
--------------------------------------------------
POST _create_from/my-index/my-new-index
{
"settings_override": {
"index": {
"number_of_shards": 5
}
},
"mappings_override": {
"properties": {
"field2": { "type": "boolean" }
}
}
}
--------------------------------------------------

Since the destination index is empty, we very likely will want to write into the index after creation.
This would not be possible if the source index contains an <<index-block-settings,index write block>> which is copied over to the destination index.
One way to handle this is to remove the index write block using a settings override. For example, the following
settings override removes all index blocks.


[source,console]
--------------------------------------------------
POST _create_from/my-index/my-new-index
{
"settings_override": {
"index": {
"blocks.write": null,
"blocks.read": null,
"blocks.read_only": null,
"blocks.read_only_allow_delete": null,
"blocks.metadata": null
}
}
}
--------------------------------------------------

Since this is a common scenario, index blocks are actually removed by default. This is controlled with the parameter
`remove_index_blocks`, which defaults to `true`. If we want the destination index to contains the index blocks from
the source index, we can do the following:

[source,console]
--------------------------------------------------
POST _create_from/my-index/my-new-index
{
"remove_index_blocks": false
}
--------------------------------------------------
64 changes: 64 additions & 0 deletions docs/reference/migration/apis/data-stream-reindex-cancel.asciidoc
Original file line number Diff line number Diff line change
@@ -0,0 +1,64 @@
[role="xpack"]
[[data-stream-reindex-cancel-api]]
=== Reindex data stream cancel API
++++
<titleabbrev>Reindex data stream cancel</titleabbrev>
++++

.New API reference
[sidebar]
--
For the most up-to-date API details, refer to {api-es}/group/endpoint-migration[Migration APIs].
--

include::{es-ref-dir}/migration/apis/shared-migration-apis-tip.asciidoc[]

Cancels a running data stream reindex task which was started by the <<data-stream-reindex-api, data stream reindex API>>.
Any backing indices that have already been reindexed and swapped into the data stream will remain in the data stream.
Only backing indices which are currently being reindexed, or pending backing indices which are still waiting to be reindexed, will be cancelled.
Once a data stream reindex task is cancelled it will no longer be accessible through the
<<data-stream-reindex-status-api,status API>>. If a reindex task is not currently running
this API will return `resource_not_found_exception`.


///////////////////////////////////////////////////////////
[source,console]
------------------------------------------------------
POST _migration/reindex
{
"source": {
"index": "my-data-stream"
},
"mode": "upgrade"
}
------------------------------------------------------
// TESTSETUP
// TEST[setup:my_data_stream]
///////////////////////////////////////////////////////////


[source,console]
----
POST _migration/reindex/my-data-stream/_cancel
----
// TEST[teardown:data_stream_cleanup]

[[data-stream-reindex-cancel-request]]
==== {api-request-title}

`GET /_migration/reindex/<data-stream>/_cancel`


[[data-stream-reindex-cancel-prereqs]]
==== {api-prereq-title}

* If the {es} {security-features} are enabled, you must have the `manage`
<<privileges-list-indices,index privilege>> for the data stream.

[[data-stream-reindex-cancel-path-params]]
==== {api-path-parms-title}

`<data-stream>`::
(Required, string)
Name of data stream to cancel reindexing.

157 changes: 157 additions & 0 deletions docs/reference/migration/apis/data-stream-reindex-status.asciidoc
Original file line number Diff line number Diff line change
@@ -0,0 +1,157 @@
[role="xpack"]
[[data-stream-reindex-status-api]]
=== Reindex data stream status API
++++
<titleabbrev>Reindex data stream status</titleabbrev>
++++

.New API reference
[sidebar]
--
For the most up-to-date API details, refer to {api-es}/group/endpoint-migration[Migration APIs].
--

include::{es-ref-dir}/migration/apis/shared-migration-apis-tip.asciidoc[]

Obtains the current status of a reindex task for the requested data stream. This status is
available while the reindex task is running and for 24 hours after completion of the task,
whether it succeeds or fails. If the task is cancelled, the status is no longer available.
If the task fails, the exception will be listed within the status.

///////////////////////////////////////////////////////////
[source,console]
------------------------------------------------------
POST _migration/reindex
{
"source": {
"index": "my-data-stream"
},
"mode": "upgrade"
}
------------------------------------------------------
// TESTSETUP
// TEST[setup:my_data_stream]
[source,console]
------------------------------------------------------
POST /_migration/reindex/my-data-stream/_cancel
DELETE _data_stream/my-data-stream
DELETE _index_template/my-data-stream-template
------------------------------------------------------
// TEARDOWN
///////////////////////////////////////////////////////////


[[data-stream-reindex-status-api-request]]
==== {api-request-title}

`GET /_migration/reindex/<data-stream>/_status`


[[data-stream-reindex-status-prereqs]]
==== {api-prereq-title}

* If the {es} {security-features} are enabled, you must have the `manage`
<<privileges-list-indices,index privilege>> for the data stream.

[[data-stream-reindex-status-path-params]]
==== {api-path-parms-title}

`<data-stream>`::
(Required, string)
Name of data stream to get status for. The reindex task for the
data stream should be currently running or have been completed in the last 24 hours.


[role="child_attributes"]
[[data-stream-reindex-status-response-body]]
==== {api-response-body-title}

`start_time`::
(Optional, <<time-units,time value>>) The time when the reindex task started.

`start_time_millis`::
(integer) The time when the reindex task started, in milliseconds since the epoch.

`complete`::
(boolean) `false` if the reindex task is still running, and `true` if the task has completed with success or failure.

`total_indices_in_data_stream`::
(integer) The total number of backing indices in the data stream, including the write index.

`total_indices_requiring_upgrade`::
(integer) The number of backing indices that need to be upgraded. These will consist of the indices which have an
older version and are not read-only.

`successes`::
(integer) The number of backing indices which have already been successfully upgraded.

`in_progress`::
(array of objects) Information on the backing indices which are currently being reindexed.
+
.Properties of objects in `in_progress`
[%collapsible%open]
=====
`index`::
(string) The name of the source backing index.
`total_doc_count`::
(integer) The number of documents in the source backing index.
`reindexed_doc_count`::
(integer) The number of documents which have already been added to the destination backing index.
=====

`pending`::
(integer) The number of backing indices which still need to be upgraded and have not yet been started.

`errors`::
(array of objects) Information on any errors which have occurred.
+
.Properties of objects in `errors`
[%collapsible%open]
=====
`index`::
(string) The name of a backing index which has had an error during reindex.
`message`::
(string) Description of the error.
=====

`exceptions`::
(Optional, string)
Exception message for a reindex failure if the failure could not be tied to a particular index.


[[data-stream-reindex-status-example]]
==== {api-examples-title}

[source,console]
----
GET _migration/reindex/my-data-stream/_status
----

The following is a typical response:
[source,console-result]
----
{
"start_time_millis": 1737676174349,
"complete": false,
"total_indices_in_data_stream": 4,
"total_indices_requiring_upgrade": 3,
"successes": 1,
"in_progress": [
{
"index": ".ds-my-data-stream-2025.01.23-000002",
"total_doc_count": 10000000,
"reindexed_doc_count": 1000
}
],
"pending": 1,
"errors": []
}
----
// TEST[skip:cannot easily clean up reindex task between tests]

For a more in-depth example showing the usage of this API along with the <<data-stream-reindex-api,reindex>> and <<data-stream-reindex-cancel-api,cancel>> APIs,
see this <<reindex-data-stream-api-example,example>>.
Loading

0 comments on commit ba7d38a

Please sign in to comment.