Allow incremental sorts for windowing functions
This expands on the work done in d2d8a229b and allows incremental sort
to be considered during create_window_paths().
Author: David Rowley
Reviewed-by: Daniel Gustafsson, Tomas Vondra
Discussion:
https://postgr.es/m/CAApHDvoOHobiA2x13NtWnWLcTXYj9dd
Doc: fix misstatement in v13 release notes.
Parallel vacuuming isn't restricted to b-tree indexes.
Noted by Peter Eisentraut.
Discussion:
https://postgr.es/m/f1c43223-3987-a23f-2063-18fd0aa4f...@2ndquadrant.com
Branch
--
REL_13_STABLE
Details
---
https://git.postgresql.org/pg/commitdif
Doc: improve release notes' info about FROM UNPACKAGED feature removal.
Per gripe from Jonathan Katz.
Discussion:
https://postgr.es/m/e0a4d177-d003-8ebb-5296-5a445472b...@postgresql.org
Branch
--
REL_13_STABLE
Details
---
https://git.postgresql.org/pg/commitdiff/d42c6176446440b185fcb95
Tag refs/tags/REL_13_RC1 was created.
Report resource usage at the end of recovery
Reporting this has been rather useful in some recent recovery speedup
work. It also seems like something that will be useful to the average DBA
too.
Author: David Rowley
Reviewed-by: Thomas Munro
Discussion:
https://postgr.es/m/CAApHDvqYVORiZxq2xPvP6
Fix use-after-free bug with event triggers in an extension script
ALTER TABLE commands in an extension script are added to an event
trigger command list; but starting with commit b5810de3f4 they do so in
a memory context that's too short-lived, so when execution ends and time
comes to use the entr
Fix use-after-free bug with event triggers in an extension script
ALTER TABLE commands in an extension script are added to an event
trigger command list; but starting with commit b5810de3f4 they do so in
a memory context that's too short-lived, so when execution ends and time
comes to use the entr
Optimize compactify_tuples function
This function could often be seen in profiles of vacuum and could often
be a significant bottleneck during recovery. The problem was that a qsort
was performed in order to sort an array of item pointers in reverse offset
order so that we could use that to safely
Fix initialization of RelationSyncEntry for streaming transactions.
In commit 464824323e, for each RelationSyncEntry we maintained the list
of xids (streamed_txns) for which we have already sent the schema. This
helps us to track when to send the schema to the downstream node for
replication of st
HashAgg: release write buffers sooner by rewinding tape.
This was an oversight. The purpose of 7fdd919ae7 was to avoid keeping
tape buffers around unnecessisarily, but HashAgg didn't rewind early
enough.
Reviewed-by: Peter Geoghegan
Discussion:
https://postgr.es/m/1fb1151c2cddf8747d14e0532da283c
HashAgg: release write buffers sooner by rewinding tape.
This was an oversight. The purpose of 7fdd919ae7 was to avoid keeping
tape buffers around unnecessisarily, but HashAgg didn't rewind early
enough.
Reviewed-by: Peter Geoghegan
Discussion:
https://postgr.es/m/1fb1151c2cddf8747d14e0532da283c
Change LogicalTapeSetBlocks() to use nBlocksWritten.
Previously, it was based on nBlocksAllocated to account for tapes with
open write buffers that may not have made it to the BufFile yet.
That was unnecessary, because callers do not need to get the number of
blocks while a tape has an open write
Change LogicalTapeSetBlocks() to use nBlocksWritten.
Previously, it was based on nBlocksAllocated to account for tapes with
open write buffers that may not have made it to the BufFile yet.
That was unnecessary, because callers do not need to get the number of
blocks while a tape has an open write
13 matches
Mail list logo
