Fixes for single-page hash index vacuum.
authorRobert Haas <rhaas@postgresql.org>
Mon, 20 Mar 2017 19:49:09 +0000 (15:49 -0400)
committerRobert Haas <rhaas@postgresql.org>
Mon, 20 Mar 2017 19:49:09 +0000 (15:49 -0400)
commit953477ca3526e28f9aeeb41d23b16eed0084c7d2
treead1b7373e82044025156843cbc9c80965d44d7f1
parentbc18126a6bcb85b51dc082c3ef4417dc016ebd9c
Fixes for single-page hash index vacuum.

Clear LH_PAGE_HAS_DEAD_TUPLES during replay, similar to what gets done
for btree.  Update hashdesc.c for xl_hash_vacuum_one_page.

Oversights in commit 6977b8b7f4dfb40896ff5e2175cad7fdbda862eb spotted
by Amit Kapila.  Patch by Ashutosh Sharma.

Bump WAL version.  The original patch to make hash indexes write-ahead
logged probably should have done this, and the single page vacuuming
patch probably should have done it again, but better late than never.

Discussion: http://postgr.es/m/CAA4eK1Kd=mJ9xreovcsh0qMiAj-QqCphHVQ_Lfau1DR9oVjASQ@mail.gmail.com
src/backend/access/hash/hash.c
src/backend/access/hash/hash_xlog.c
src/backend/access/hash/hashinsert.c
src/backend/access/rmgrdesc/hashdesc.c
src/include/access/hash_xlog.h
src/include/access/xlog_internal.h