Thread: pgsql: Fix race condition between hot standby and restoring a full-page
pgsql: Fix race condition between hot standby and restoring a full-page
From
Heikki Linnakangas
Date:
Fix race condition between hot standby and restoring a full-page image. There was a window in RestoreBackupBlock where a page would be zeroed out, but not yet locked. If a backend pinned and locked the page in that window, it saw the zeroed page instead of the old page or new page contents, which could lead to missing rows in a result set, or errors. To fix, replace RBM_ZERO with RBM_ZERO_AND_LOCK, which atomically pins, zeroes, and locks the page, if it's not in the buffer cache already. In stable branches, the old RBM_ZERO constant is renamed to RBM_DO_NOT_USE, to avoid breaking any 3rd party extensions that might use RBM_ZERO. More importantly, this avoids renumbering the other enum values, which would cause even bigger confusion in extensions that use ReadBufferExtended, but haven't been recompiled. Backpatch to all supported versions; this has been racy since hot standby was introduced. Branch ------ REL9_2_STABLE Details ------- http://git.postgresql.org/pg/commitdiff/7eab804c22d40e1c0936c88334a8861d43d8e863 Modified Files -------------- src/backend/access/hash/hashpage.c | 13 +++++----- src/backend/access/heap/heapam.c | 3 +-- src/backend/access/transam/xlog.c | 6 +---- src/backend/access/transam/xlogutils.c | 17 +++++++++---- src/backend/storage/buffer/bufmgr.c | 41 ++++++++++++++++++++++++++++---- src/include/storage/bufmgr.h | 11 ++++++--- 6 files changed, 66 insertions(+), 25 deletions(-)