pg.ddx.io pgsql-committers@postgresql.org mailing list archivehelp / color / mirror / Atom feed
pgsql: Consider parallel awareness when removing single-child Appends 4+ messages / 1 participants [nested] [flat]
* pgsql: Consider parallel awareness when removing single-child Appends @ 2022-01-25 08:10 David Rowley <drowley@postgresql.org> 0 siblings, 0 replies; 4+ messages in thread From: David Rowley @ 2022-01-25 08:10 UTC (permalink / raw) To: pgsql-committers@lists.postgresql.org Consider parallel awareness when removing single-child Appends 8edd0e794 added some code to remove Append and MergeAppend nodes when they contained a single child node. As it turned out, this was unsafe to do when the Append/MergeAppend was parallel_aware and the child node was not. Removing the Append/MergeAppend, in this case, could lead to the child plan being called multiple times by parallel workers when it was unsafe to do so. Here we fix this by just not removing the Append/MergeAppend when the parallel_aware flag of the parent and child node don't match. Reported-by: Yura Sokolov Bug: #17335 Discussion: https://postgr.es/m/b59605fecb20ba9ea94e70ab60098c237c870628.camel%40postgrespro.ru Backpatch-through: 12, where 8edd0e794 was first introduced Branch ------ master Details ------- https://git.postgresql.org/pg/commitdiff/f9a74c1498d25098606fc5c4621216d38a1cc57d Modified Files -------------- src/backend/optimizer/plan/setrefs.c | 24 ++++++++++++++++++++---- 1 file changed, 20 insertions(+), 4 deletions(-) ^ permalink raw reply [nested|flat] 4+ messages in thread
* pgsql: Consider parallel awareness when removing single-child Appends @ 2022-01-25 08:14 David Rowley <drowley@postgresql.org> 0 siblings, 0 replies; 4+ messages in thread From: David Rowley @ 2022-01-25 08:14 UTC (permalink / raw) To: pgsql-committers@lists.postgresql.org Consider parallel awareness when removing single-child Appends 8edd0e794 added some code to remove Append and MergeAppend nodes when they contained a single child node. As it turned out, this was unsafe to do when the Append/MergeAppend was parallel_aware and the child node was not. Removing the Append/MergeAppend, in this case, could lead to the child plan being called multiple times by parallel workers when it was unsafe to do so. Here we fix this by just not removing the Append/MergeAppend when the parallel_aware flag of the parent and child node don't match. Reported-by: Yura Sokolov Bug: #17335 Discussion: https://postgr.es/m/b59605fecb20ba9ea94e70ab60098c237c870628.camel%40postgrespro.ru Backpatch-through: 12, where 8edd0e794 was first introduced Branch ------ REL_14_STABLE Details ------- https://git.postgresql.org/pg/commitdiff/357ff66153b4791d32c45724d8a0d3c91d89c937 Modified Files -------------- src/backend/optimizer/plan/setrefs.c | 24 ++++++++++++++++++++---- 1 file changed, 20 insertions(+), 4 deletions(-) ^ permalink raw reply [nested|flat] 4+ messages in thread
* pgsql: Consider parallel awareness when removing single-child Appends @ 2022-01-25 08:15 David Rowley <drowley@postgresql.org> 0 siblings, 0 replies; 4+ messages in thread From: David Rowley @ 2022-01-25 08:15 UTC (permalink / raw) To: pgsql-committers@lists.postgresql.org Consider parallel awareness when removing single-child Appends 8edd0e794 added some code to remove Append and MergeAppend nodes when they contained a single child node. As it turned out, this was unsafe to do when the Append/MergeAppend was parallel_aware and the child node was not. Removing the Append/MergeAppend, in this case, could lead to the child plan being called multiple times by parallel workers when it was unsafe to do so. Here we fix this by just not removing the Append/MergeAppend when the parallel_aware flag of the parent and child node don't match. Reported-by: Yura Sokolov Bug: #17335 Discussion: https://postgr.es/m/b59605fecb20ba9ea94e70ab60098c237c870628.camel%40postgrespro.ru Backpatch-through: 12, where 8edd0e794 was first introduced Branch ------ REL_13_STABLE Details ------- https://git.postgresql.org/pg/commitdiff/f8807e7742e5ce058b6fe05d797852e082e6d2c8 Modified Files -------------- src/backend/optimizer/plan/setrefs.c | 24 ++++++++++++++++++++---- 1 file changed, 20 insertions(+), 4 deletions(-) ^ permalink raw reply [nested|flat] 4+ messages in thread
* pgsql: Consider parallel awareness when removing single-child Appends @ 2022-01-25 08:16 David Rowley <drowley@postgresql.org> 0 siblings, 0 replies; 4+ messages in thread From: David Rowley @ 2022-01-25 08:16 UTC (permalink / raw) To: pgsql-committers@lists.postgresql.org Consider parallel awareness when removing single-child Appends 8edd0e794 added some code to remove Append and MergeAppend nodes when they contained a single child node. As it turned out, this was unsafe to do when the Append/MergeAppend was parallel_aware and the child node was not. Removing the Append/MergeAppend, in this case, could lead to the child plan being called multiple times by parallel workers when it was unsafe to do so. Here we fix this by just not removing the Append/MergeAppend when the parallel_aware flag of the parent and child node don't match. Reported-by: Yura Sokolov Bug: #17335 Discussion: https://postgr.es/m/b59605fecb20ba9ea94e70ab60098c237c870628.camel%40postgrespro.ru Backpatch-through: 12, where 8edd0e794 was first introduced Branch ------ REL_12_STABLE Details ------- https://git.postgresql.org/pg/commitdiff/2ccd8fb29d74f91b291be20df43f9c0194642ae8 Modified Files -------------- src/backend/optimizer/plan/setrefs.c | 24 ++++++++++++++++++++---- 1 file changed, 20 insertions(+), 4 deletions(-) ^ permalink raw reply [nested|flat] 4+ messages in thread
end of thread, other threads:[~2022-01-25 08:16 UTC | newest] Thread overview: 4+ messages (download: mbox mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 2022-01-25 08:10 pgsql: Consider parallel awareness when removing single-child Appends David Rowley <drowley@postgresql.org> 2022-01-25 08:14 pgsql: Consider parallel awareness when removing single-child Appends David Rowley <drowley@postgresql.org> 2022-01-25 08:15 pgsql: Consider parallel awareness when removing single-child Appends David Rowley <drowley@postgresql.org> 2022-01-25 08:16 pgsql: Consider parallel awareness when removing single-child Appends David Rowley <drowley@postgresql.org>
This inbox is served by DDX for PostgreSQL; see mirroring instructions for how to clone and mirror all data and code used for this inbox