Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Skip to content

Commit 306ff0a

Browse files
committed
Further 9.6 release note improvements.
Call out the major enhancements in this release as identified by pgsql-advocacy discussion, and rearrange some of the entries to make those items more prominent. Other minor improvements per advice from Vitaly Burovoy, Masahiko Sawada, Peter Geoghegan, and Andres Freund.
1 parent 98f158e commit 306ff0a

File tree

1 file changed

+125
-109
lines changed

1 file changed

+125
-109
lines changed

doc/src/sgml/release-9.6.sgml

Lines changed: 125 additions & 109 deletions
Original file line numberDiff line numberDiff line change
@@ -17,13 +17,46 @@
1717
Major enhancements in <productname>PostgreSQL</> 9.6 include:
1818
</para>
1919

20-
<!-- This list duplicates items below, but without authors or details-->
20+
<!-- Items in this list summarize one or more items below -->
2121

2222
<itemizedlist>
2323

2424
<listitem>
2525
<para>
26-
To be filled in, but see below for the long version
26+
Parallel sequential scans, joins and aggregates
27+
</para>
28+
</listitem>
29+
30+
<listitem>
31+
<para>
32+
Elimination of repetitive scanning of old data by autovacuum
33+
</para>
34+
</listitem>
35+
36+
<listitem>
37+
<para>
38+
Synchronous replication now allows multiple standby servers for
39+
increased reliability
40+
</para>
41+
</listitem>
42+
43+
<listitem>
44+
<para>
45+
Full-text search for phrases
46+
</para>
47+
</listitem>
48+
49+
<listitem>
50+
<para>
51+
Support for remote joins, sorts, and updates
52+
in <filename>postgres_fdw</>
53+
</para>
54+
</listitem>
55+
56+
<listitem>
57+
<para>
58+
Substantial performance improvements, especially in the area of
59+
improving scalability on many-CPU servers
2760
</para>
2861
</listitem>
2962

@@ -315,6 +348,7 @@
315348
2016-04-24 [8f91d87d4] Fix documentation & config inconsistencies around 428b1d
316349
2016-04-26 [72a98a639] Don't open formally non-existent segments in _mdfd_getse
317350
2016-05-04 [a71248708] Fix transient mdsync() errors of truncated relations due
351+
2016-02-16 [7975c5e0a] Allow the WAL writer to flush WAL at a reduced rate.
318352
-->
319353
<para>
320354
Where feasible, trigger kernel writeback after a configurable number
@@ -549,16 +583,6 @@
549583

550584
<listitem>
551585
<!--
552-
2016-02-16 [7975c5e0a] Allow the WAL writer to flush WAL at a reduced rate.
553-
-->
554-
<para>
555-
Reduce the WAL writer's flush rate to reduce storage bottlenecks
556-
(Andres Freund)
557-
</para>
558-
</listitem>
559-
560-
<listitem>
561-
<!--
562586
2016-04-08 [719c84c1b] Extend relations multiple blocks at a time to improve sc
563587
-->
564588
<para>
@@ -567,32 +591,26 @@
567591
</para>
568592

569593
<para>
570-
This reduces kernel traffic and improves scalability when multiple
571-
processes are inserting into the same relation.
594+
This improves scalability by decreasing contention.
572595
</para>
573596
</listitem>
574597

575598
<listitem>
576599
<!--
577600
2016-04-08 [071180377] Use quicksort, not replacement selection, for external s
601+
2016-03-17 [0011c0091] Improve memory management for external sorts.
578602
-->
579603
<para>
580604
Improve sorting performance by using quicksort, not replacement
581605
selection, within steps of an external sort (Peter Geoghegan)
582606
</para>
583607

584608
<para>
585-
This behavior can be adjusted via the new configuration
586-
parameter <xref linkend="guc-replacement-sort-tuples">.
587-
</para>
588-
</listitem>
589-
590-
<listitem>
591-
<!--
592-
2016-03-17 [0011c0091] Improve memory management for external sorts.
593-
-->
594-
<para>
595-
Improve memory management for external sorts (Peter Geoghegan)
609+
The new approach makes better use of CPU cache for typical cache
610+
sizes and data volumes. Where necessary, the behavior can be
611+
adjusted via the new configuration
612+
parameter <xref linkend="guc-replacement-sort-tuples">,
613+
which see for further details.
596614
</para>
597615
</listitem>
598616

@@ -610,29 +628,31 @@
610628
<listitem>
611629
<!--
612630
2015-11-06 [a76ef15d9] Add sort support routine for the UUID data type.
631+
2016-02-03 [b47b4dbf6] Extend sortsupport for text to more opclasses.
632+
2016-02-17 [f1f5ec1ef] Reuse abbreviated keys in ordered [set] aggregates.
613633
-->
614634
<para>
615-
Speed up sorting of <type>uuid</>-type columns (Peter Geoghegan)
635+
Speed up sorting of <type>uuid</>, <type>bytea</>,
636+
and <type>char(n)</> fields by using <quote>abbreviated</> keys
637+
(Peter Geoghegan)
616638
</para>
617-
</listitem>
618639

619-
<listitem>
620-
<!--
621-
2015-12-16 [b648b7034] Speed up CREATE INDEX CONCURRENTLY's TID sort.
622-
-->
623640
<para>
624-
Speed up <command>CREATE INDEX CONCURRENTLY</> by treating TIDs
625-
as 64-bit integers during the sort phase (Peter Geoghegan)
641+
Support for abbreviated keys has also been added to the non-default
642+
operator classes <literal>text_pattern_ops</>,
643+
<literal>varchar_pattern_ops</>, and <literal>bpchar_pattern_ops</>.
644+
Processing of ordered-set aggregates can also now exploit
645+
abbreviated keys.
626646
</para>
627647
</listitem>
628648

629649
<listitem>
630650
<!--
631-
2016-02-17 [f1f5ec1ef] Reuse abbreviated keys in ordered [set] aggregates.
651+
2015-12-16 [b648b7034] Speed up CREATE INDEX CONCURRENTLY's TID sort.
632652
-->
633653
<para>
634-
Improve performance of ordered-set aggregates by reusing
635-
abbreviated-key information (Peter Geoghegan)
654+
Speed up <command>CREATE INDEX CONCURRENTLY</> by treating TIDs
655+
as 64-bit integers during the sort phase (Peter Geoghegan)
636656
</para>
637657
</listitem>
638658

@@ -1179,6 +1199,41 @@ XXX this is pending backpatch, may need to remove
11791199

11801200
<listitem>
11811201
<!--
1202+
2016-04-06 [989be0810] Support multiple synchronous standby servers.
1203+
2016-04-27 [4c804fbdf] Clean up parsing of synchronous_standby_names GUC variab
1204+
2016-04-30 [17d5db352] Remove warning about num_sync being too large in synchro
1205+
-->
1206+
<para>
1207+
Support synchronous replication with multiple synchronous standby
1208+
servers, not just one (Masahiko Sawada, Beena Emerson, Michael
1209+
Paquier, Fujii Masao, Kyotaro Horiguchi)
1210+
</para>
1211+
1212+
<para>
1213+
The number of standby servers that must acknowledge a commit before
1214+
it's considered done is now configurable.
1215+
</para>
1216+
</listitem>
1217+
1218+
<listitem>
1219+
<!--
1220+
2016-03-29 [314cbfc5d] Add new replication mode synchronous_commit = 'remote_ap
1221+
-->
1222+
<para>
1223+
Add new setting <literal>remote_apply</> for configuration parameter
1224+
<xref linkend="guc-synchronous-commit"> (Thomas Munro)
1225+
</para>
1226+
1227+
<para>
1228+
In this mode, the master waits for the transaction to be applied on
1229+
the standby server, not just written to disk. That means that you
1230+
can count on a transaction started on the standby to see all commits
1231+
previously acknowledged by the master.
1232+
</para>
1233+
</listitem>
1234+
1235+
<listitem>
1236+
<!--
11821237
2015-08-11 [6fcd88511] Allow pg_create_physical_replication_slot() to reserve W
11831238
2015-09-06 [c314ead5b] Add ability to reserve WAL upon slot creation via replic
11841239
-->
@@ -1222,36 +1277,6 @@ XXX this is pending backpatch, may need to remove
12221277
</para>
12231278
</listitem>
12241279

1225-
<listitem>
1226-
<!--
1227-
2016-04-06 [989be0810] Support multiple synchronous standby servers.
1228-
2016-04-27 [4c804fbdf] Clean up parsing of synchronous_standby_names GUC variab
1229-
2016-04-30 [17d5db352] Remove warning about num_sync being too large in synchro
1230-
-->
1231-
<para>
1232-
Support synchronous replication with multiple synchronous standby
1233-
servers, not just one (Sawada Masahiko, Beena Emerson, Michael
1234-
Paquier, Fujii Masao, Kyotaro Horiguchi)
1235-
</para>
1236-
</listitem>
1237-
1238-
<listitem>
1239-
<!--
1240-
2016-03-29 [314cbfc5d] Add new replication mode synchronous_commit = 'remote_ap
1241-
-->
1242-
<para>
1243-
Add new setting <literal>remote_apply</> for configuration parameter
1244-
<xref linkend="guc-synchronous-commit"> (Thomas Munro)
1245-
</para>
1246-
1247-
<para>
1248-
In this mode, the master waits for the transaction to be applied on
1249-
the standby server, not just written to disk. That means that you
1250-
can count on a transaction started on the standby to see all commits
1251-
previously acknowledged by the master.
1252-
</para>
1253-
</listitem>
1254-
12551280
</itemizedlist>
12561281

12571282
</sect3>
@@ -1379,7 +1404,7 @@ XXX this is pending backpatch, may need to remove
13791404
<para>
13801405
Add options to <command>ALTER OPERATOR</command> to change the
13811406
selectivity functions associated with an existing operator
1382-
(Uriy Zhuravlev)
1407+
(Yury Zhuravlev)
13831408
</para>
13841409
</listitem>
13851410

@@ -1578,7 +1603,7 @@ XXX this is pending backpatch, may need to remove
15781603
<para>
15791604
Allow omitting one or both boundaries in an array slice specifier,
15801605
for example <literal>array_col[3:]</>
1581-
(YUriy Zhuravlev)
1606+
(Yury Zhuravlev)
15821607
</para>
15831608

15841609
<para>
@@ -1637,7 +1662,7 @@ XXX this is pending backpatch, may need to remove
16371662
2016-04-07 [bb140506d] Phrase full text search.
16381663
-->
16391664
<para>
1640-
Improve full text search to support searching for phrases, that is,
1665+
Improve full-text search to support searching for phrases, that is,
16411666
lexemes appearing adjacent to each other in a specific order, or with
16421667
at most a specified distance between them
16431668
(Teodor Sigaev, Oleg Bartunov, Dmitry Ivanov)
@@ -1844,7 +1869,7 @@ XXX this is pending backpatch, may need to remove
18441869
</para>
18451870

18461871
<para>
1847-
These are <function>delete()</>, <function>filter()</>,
1872+
These are <function>ts_delete()</>, <function>ts_filter()</>,
18481873
<function>unnest()</>, <function>tsvector_to_array()</>,
18491874
<function>array_to_tsvector()</>, and a variant
18501875
of <function>setweight()</> that sets the weight only for specified
@@ -1944,7 +1969,7 @@ XXX this is pending backpatch, may need to remove
19441969
-->
19451970
<para>
19461971
Add <function>pg_size_bytes()</> function to convert human-readable
1947-
size strings to numbers (Pavel Stehule, Vitaly Burovoy)
1972+
size strings to numbers (Pavel Stehule, Vitaly Burovoy, Dean Rasheed)
19481973
</para>
19491974

19501975
<para>
@@ -2459,8 +2484,9 @@ This commit is also listed under libpq and PL/pgSQL
24592484
2016-05-05 [7a622b273] Rename pgbench min/max to least/greatest, and fix handli
24602485
-->
24612486
<para>
2462-
Support some <link linkend="pgbench-builtin-functions">built-in
2463-
functions</> in expressions in backslash commands (Fabien Coelho)
2487+
Support floating-point arithmetic, as well as
2488+
some <link linkend="pgbench-builtin-functions">built-in functions</>,
2489+
in expressions in backslash commands (Fabien Coelho)
24642490
</para>
24652491
</listitem>
24662492

@@ -2756,16 +2782,6 @@ This commit is also listed under libpq and PL/pgSQL
27562782

27572783
<listitem>
27582784
<!--
2759-
2016-02-03 [b47b4dbf6] Extend sortsupport for text to more opclasses.
2760-
-->
2761-
<para>
2762-
Generalize the SortSupport for <type>text</> so that it can be
2763-
used by other types such as <type>bpchar</> (Peter Geoghegan)
2764-
</para>
2765-
</listitem>
2766-
2767-
<listitem>
2768-
<!--
27692785
2016-02-04 [c1772ad92] Change the way that LWLocks for extensions are allocated
27702786
-->
27712787
<para>
@@ -3038,7 +3054,7 @@ This commit is also listed under libpq and PL/pgSQL
30383054
-->
30393055
<para>
30403056
Add selectivity estimation functions for <filename>contrib/intarray</>
3041-
operators, to improve plans for queries using those operators (Uriy
3057+
operators, to improve plans for queries using those operators (Yury
30423058
Zhuravlev, Alexander Korotkov)
30433059
</para>
30443060
</listitem>
@@ -3189,31 +3205,6 @@ This commit is also listed under libpq and PL/pgSQL
31893205

31903206
<listitem>
31913207
<!--
3192-
2016-01-28 [96198d94c] Avoid multiple foreign server connections when all use s
3193-
-->
3194-
<para>
3195-
Use a single foreign-server connection for local user IDs that all
3196-
map to the same remote user (Ashutosh Bapat)
3197-
</para>
3198-
</listitem>
3199-
3200-
<listitem>
3201-
<!--
3202-
2016-02-03 [dc203dc3a] postgres_fdw: Allow fetch_size to be set per-table or pe
3203-
-->
3204-
<para>
3205-
Allow the fetch size to be set as a server or table option
3206-
(Corey Huinker)
3207-
</para>
3208-
3209-
<para>
3210-
Formerly, <filename>postgres_fdw</> always fetched 100 rows at a time
3211-
from remote queries; now that behavior is adjustable.
3212-
</para>
3213-
</listitem>
3214-
3215-
<listitem>
3216-
<!--
32173208
2015-11-03 [f18c944b6] postgres_fdw: Add ORDER BY to some remote SQL queries.
32183209
2015-12-22 [ccd8f9792] postgres_fdw: Consider requesting sorted data so we can
32193210
-->
@@ -3253,6 +3244,31 @@ This commit is also listed under libpq and PL/pgSQL
32533244

32543245
<listitem>
32553246
<!--
3247+
2016-02-03 [dc203dc3a] postgres_fdw: Allow fetch_size to be set per-table or pe
3248+
-->
3249+
<para>
3250+
Allow the fetch size to be set as a server or table option
3251+
(Corey Huinker)
3252+
</para>
3253+
3254+
<para>
3255+
Formerly, <filename>postgres_fdw</> always fetched 100 rows at a time
3256+
from remote queries; now that behavior is adjustable.
3257+
</para>
3258+
</listitem>
3259+
3260+
<listitem>
3261+
<!--
3262+
2016-01-28 [96198d94c] Avoid multiple foreign server connections when all use s
3263+
-->
3264+
<para>
3265+
Use a single foreign-server connection for local user IDs that all
3266+
map to the same remote user (Ashutosh Bapat)
3267+
</para>
3268+
</listitem>
3269+
3270+
<listitem>
3271+
<!--
32563272
2016-04-21 [f039eaac7] Allow queries submitted by postgres_fdw to be canceled.
32573273
-->
32583274
<para>

0 commit comments

Comments
 (0)