On Thu, Dec 11, 2014 at 08:35:43PM -0500, Peter Eisentraut wrote:
> On 12/11/14 11:44 AM, Kevin Grittner wrote:
> > We want to finish with VACUUM FREEZE without the FULL, unless we
> > don't care about missing visibility maps and free space maps.
I have create the attached initdb patch to update this.
> Why would we care, and if we do, why does VACUUM FULL remove them?
>
> You can also run plain VACUUM after FULL to put the maps back.
>
> But the documentation is apparently missing details about this.
It is a long-standing TODO item I tried to fix, but couldn't:
Allow VACUUM FULL and CLUSTER to update the visibility map
index-only scans : abnormal heap fetches after VACUUM FULL
http://www.postgresql.org/message-id/[email protected]
--
Bruce Momjian <[email protected]> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ Everyone has their own god. +
diff --git a/src/bin/initdb/initdb.c b/src/bin/initdb/initdb.c
new file mode 100644
index 18614e7..6a87467
*** a/src/bin/initdb/initdb.c
--- b/src/bin/initdb/initdb.c
*************** vacuum_db(void)
*** 2334,2340 ****
PG_CMD_OPEN;
! PG_CMD_PUTS("ANALYZE;\nVACUUM FULL;\nVACUUM FREEZE;\n");
PG_CMD_CLOSE;
--- 2334,2341 ----
PG_CMD_OPEN;
! /* Run analyze before VACUUM so the statistics are frozen. */
! PG_CMD_PUTS("ANALYZE;\nVACUUM FREEZE;\n");
PG_CMD_CLOSE;
--
Sent via pgsql-hackers mailing list ([email protected])
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
