doc: Mention BRIN indexes support parallel builds
authorTomas Vondra <[email protected]>
Mon, 16 Dec 2024 18:08:52 +0000 (19:08 +0100)
committerTomas Vondra <[email protected]>
Mon, 16 Dec 2024 18:20:22 +0000 (19:20 +0100)
Two places in the documentation suggest B-tree is the only index access
method allowing parallel builds. Commit b4375717 added parallel builds
for BRIN too, but failed to update the docs. So fix that, and back
to 17, where parallel BRIN builds were introduced.

Author: Egor Rogov
Back-through: 17
Discussion: https://postgr.es/m/114e2d5d-125e-07d8-94aa-5ad175fb7443@postgrespro.ru

doc/src/sgml/config.sgml
doc/src/sgml/ref/create_index.sgml

index c782fc99a1bc452e01043766e5efef6123a9a7d2..4b4d305d26d9a23e7632bc4d9539c7fcdeac2815 100644 (file)
@@ -2810,7 +2810,7 @@ include_dir 'conf.d'
          Sets the maximum number of parallel workers that can be
          started by a single utility command.  Currently, the parallel
          utility commands that support the use of parallel workers are
-         <command>CREATE INDEX</command> only when building a B-tree index,
+         <command>CREATE INDEX</command> when building a B-tree or BRIN index,
          and <command>VACUUM</command> without <literal>FULL</literal>
          option.  Parallel workers are taken from the pool of processes
          established by <xref linkend="guc-max-worker-processes"/>, limited
index 621bc0e253ceb7b051a8324b8eac5b60cf35564f..208389e80060b33aa2f830fbb101b1d40fddd4bf 100644 (file)
@@ -808,7 +808,7 @@ Indexes:
    leveraging multiple CPUs in order to process the table rows faster.
    This feature is known as <firstterm>parallel index
    build</firstterm>.  For index methods that support building indexes
-   in parallel (currently, only B-tree),
+   in parallel (currently, B-tree and BRIN),
    <varname>maintenance_work_mem</varname> specifies the maximum
    amount of memory that can be used by each index build operation as
    a whole, regardless of how many worker processes were started.