projects / users / rhaas / postgres.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 6ad8ac6) | patch
Don't bother to lock bufmgr partitions in pg_buffercache.
authorHeikki Linnakangas <heikki.linnakangas@iki.fi>
Thu, 29 Sep 2016 10:16:30 +0000 (13:16 +0300)
committerHeikki Linnakangas <heikki.linnakangas@iki.fi>
Thu, 29 Sep 2016 10:16:30 +0000 (13:16 +0300)
commit6e654546fb61f62cc982d0c8f62241b3b30e7ef8
tree8c94b730bf7e132b05bfe0411227bc25cf0cd062tree
parent6ad8ac6026287e3ccbc4d606b6ab6116ccc0eec8commit | diff
Don't bother to lock bufmgr partitions in pg_buffercache.

That makes the view a lot less disruptive to use on a production system.
Without the locks, you don't get a consistent snapshot across all buffers,
but that's OK. It wasn't a very useful guarantee in practice.

Ivan Kartyshov, reviewed by Tomas Vondra and Robert Haas.

Discusssion: <f9d6cab2-73a7-7a84-55a8-07dcb8516ae5@postgrespro.ru>
contrib/pg_buffercache/pg_buffercache_pages.c diff | blob | blame | history
Robert Haas's development tree.