Tom Lane 2095206de1 Adjust btree index build to not use shared buffers, thereby avoiding the ... locking conflict against concurrent CHECKPOINT that was discussed a few weeks ago. Also, if not using WAL archiving (which is always true ATM but won't be if PITR makes it into this release), there's no need to WAL-log the index build process; it's sufficient to force-fsync the completed index before commit. This seems to gain about a factor of 2 in my tests, which is consistent with writing half as much data. I did not try it with WAL on a separate drive though --- probably the gain would be a lot less in that scenario.		2004-06-02 17:28:18 +00:00
..
common	Use the new List API function names throughout the backend, and disable the	2004-05-30 23:40:41 +00:00
gist	Cleanup vectors of GISTENTRY and eliminate problem with 64-bit strict-aligned	2004-03-30 15:45:33 +00:00
hash	Centralize implementation of delay code by creating a pg_usleep()	2004-02-10 03:42:45 +00:00
heap	Get rid of the former rather baroque mechanism for propagating the values	2004-05-27 17:12:57 +00:00
index	Tweak indexscan and seqscan code to arrange that steps from one page to	2004-04-21 18:24:26 +00:00
nbtree	Adjust btree index build to not use shared buffers, thereby avoiding the	2004-06-02 17:28:18 +00:00
rtree	Centralize implementation of delay code by creating a pg_usleep()	2004-02-10 03:42:45 +00:00
transam	Per previous discussions, get rid of use of sync(2) in favor of	2004-05-31 03:48:10 +00:00
Makefile	$Header: -> $PostgreSQL Changes ...	2003-11-29 19:52:15 +00:00