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

Commit e427ea7

Browse files
tglsfdctglsfdc
committed
Release notes for 9.4.2, 9.3.7, 9.2.11, 9.1.16, 9.0.20.
1 parent d817b7a commit e427ea7

File tree

5 files changed

+3061
-2
lines changed

5 files changed

+3061
-2
lines changed

‎doc/src/sgml/release-9.0.sgml

+380
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,386 @@
11
<!-- doc/src/sgml/release-9.0.sgml -->
22
<!-- See header comment in release.sgml about typical markup -->
33

4+
<sect1 id="release-9-0-20">
5+
<title>Release 9.0.20</title>
6+
7+
<note>
8+
<title>Release Date</title>
9+
<simpara>2015-05-21</simpara>
10+
</note>
11+
12+
<para>
13+
This release contains a variety of fixes from 9.0.19.
14+
For information about new features in the 9.0 major release, see
15+
<xref linkend="release-9-0">.
16+
</para>
17+
18+
<para>
19+
The <productname>PostgreSQL</> community will stop releasing updates
20+
for the 9.0.X release series in September 2015.
21+
Users are encouraged to update to a newer release branch soon.
22+
</para>
23+
24+
<sect2>
25+
<title>Migration to Version 9.0.20</title>
26+
27+
<para>
28+
A dump/restore is not required for those running 9.0.X.
29+
</para>
30+
31+
<para>
32+
However, if you are upgrading from a version earlier than 9.0.18,
33+
see <xref linkend="release-9-0-18">.
34+
</para>
35+
36+
</sect2>
37+
38+
<sect2>
39+
<title>Changes</title>
40+
41+
<itemizedlist>
42+
43+
<listitem>
44+
<para>
45+
Fix incorrect checking of deferred exclusion constraints after a HOT
46+
update (Tom Lane)
47+
</para>
48+
49+
<para>
50+
If a new row that potentially violates a deferred exclusion constraint
51+
is HOT-updated (that is, no indexed columns change and the row can be
52+
stored back onto the same table page) later in the same transaction,
53+
the exclusion constraint would be reported as violated when the check
54+
finally occurred, even if the row(s) the new row originally conflicted
55+
with had been deleted.
56+
</para>
57+
</listitem>
58+
59+
<listitem>
60+
<para>
61+
Prevent improper reordering of antijoins (NOT EXISTS joins) versus
62+
other outer joins (Tom Lane)
63+
</para>
64+
65+
<para>
66+
This oversight in the planner has been observed to cause <quote>could
67+
not find RelOptInfo for given relids</> errors, but it seems possible
68+
that sometimes an incorrect query plan might get past that consistency
69+
check and result in silently-wrong query output.
70+
</para>
71+
</listitem>
72+
73+
<listitem>
74+
<para>
75+
Fix incorrect matching of subexpressions in outer-join plan nodes
76+
(Tom Lane)
77+
</para>
78+
79+
<para>
80+
Previously, if textually identical non-strict subexpressions were used
81+
both above and below an outer join, the planner might try to re-use
82+
the value computed below the join, which would be incorrect because the
83+
executor would force the value to NULL in case of an unmatched outer row.
84+
</para>
85+
</listitem>
86+
87+
<listitem>
88+
<para>
89+
Fix GEQO planner to cope with failure of its join order heuristic
90+
(Tom Lane)
91+
</para>
92+
93+
<para>
94+
This oversight has been seen to lead to <quote>failed to join all
95+
relations together</> errors in queries involving <literal>LATERAL</>,
96+
and that might happen in other cases as well.
97+
</para>
98+
</listitem>
99+
100+
<listitem>
101+
<para>
102+
Fix possible deadlock at startup
103+
when <literal>max_prepared_transactions</> is too small
104+
(Heikki Linnakangas)
105+
</para>
106+
</listitem>
107+
108+
<listitem>
109+
<para>
110+
Don't archive useless preallocated WAL files after a timeline switch
111+
(Heikki Linnakangas)
112+
</para>
113+
</listitem>
114+
115+
<listitem>
116+
<para>
117+
Recursively <function>fsync()</> the data directory after a crash
118+
(Abhijit Menon-Sen, Robert Haas)
119+
</para>
120+
121+
<para>
122+
This ensures consistency if another crash occurs shortly later. (The
123+
second crash would have to be a system-level crash, not just a database
124+
crash, for there to be a problem.)
125+
</para>
126+
</listitem>
127+
128+
<listitem>
129+
<para>
130+
Fix autovacuum launcher's possible failure to shut down, if an error
131+
occurs after it receives SIGTERM (&Aacute;lvaro Herrera)
132+
</para>
133+
</listitem>
134+
135+
<listitem>
136+
<para>
137+
Cope with unexpected signals in <function>LockBufferForCleanup()</>
138+
(Andres Freund)
139+
</para>
140+
141+
<para>
142+
This oversight could result in spurious errors about <quote>multiple
143+
backends attempting to wait for pincount 1</>.
144+
</para>
145+
</listitem>
146+
147+
<listitem>
148+
<para>
149+
Avoid waiting for WAL flush or synchronous replication during commit of
150+
a transaction that was read-only so far as the user is concerned
151+
(Andres Freund)
152+
</para>
153+
154+
<para>
155+
Previously, a delay could occur at commit in transactions that had
156+
written WAL due to HOT page pruning, leading to undesirable effects
157+
such as sessions getting stuck at startup if all synchronous replicas
158+
are down. Sessions have also been observed to get stuck in catchup
159+
interrupt processing when using synchronous replication; this will fix
160+
that problem as well.
161+
</para>
162+
</listitem>
163+
164+
<listitem>
165+
<para>
166+
Fix crash when manipulating hash indexes on temporary tables
167+
(Heikki Linnakangas)
168+
</para>
169+
</listitem>
170+
171+
<listitem>
172+
<para>
173+
Fix possible failure during hash index bucket split, if other processes
174+
are modifying the index concurrently (Tom Lane)
175+
</para>
176+
</listitem>
177+
178+
<listitem>
179+
<para>
180+
Check for interrupts while analyzing index expressions (Jeff Janes)
181+
</para>
182+
183+
<para>
184+
<command>ANALYZE</> executes index expressions many times; if there are
185+
slow functions in such an expression, it's desirable to be able to
186+
cancel the <command>ANALYZE</> before that loop finishes.
187+
</para>
188+
</listitem>
189+
190+
<listitem>
191+
<para>
192+
Add the name of the target server to object description strings for
193+
foreign-server user mappings (&Aacute;lvaro Herrera)
194+
</para>
195+
</listitem>
196+
197+
<listitem>
198+
<para>
199+
Recommend setting <literal>include_realm</> to 1 when using
200+
Kerberos/GSSAPI/SSPI authentication (Stephen Frost)
201+
</para>
202+
203+
<para>
204+
Without this, identically-named users from different realms cannot be
205+
distinguished. For the moment this is only a documentation change, but
206+
it will become the default setting in <productname>PostgreSQL</> 9.5.
207+
</para>
208+
</listitem>
209+
210+
<listitem>
211+
<para>
212+
Remove code for matching IPv4 <filename>pg_hba.conf</> entries to
213+
IPv4-in-IPv6 addresses (Tom Lane)
214+
</para>
215+
216+
<para>
217+
This hack was added in 2003 in response to a report that some Linux
218+
kernels of the time would report IPv4 connections as having
219+
IPv4-in-IPv6 addresses. However, the logic was accidentally broken in
220+
9.0. The lack of any field complaints since then shows that it's not
221+
needed anymore. Now we have reports that the broken code causes
222+
crashes on some systems, so let's just remove it rather than fix it.
223+
(Had we chosen to fix it, that would make for a subtle and potentially
224+
security-sensitive change in the effective meaning of
225+
IPv4 <filename>pg_hba.conf</> entries, which does not seem like a good
226+
thing to do in minor releases.)
227+
</para>
228+
</listitem>
229+
230+
<listitem>
231+
<para>
232+
While shutting down service on Windows, periodically send status
233+
updates to the Service Control Manager to prevent it from killing the
234+
service too soon; and ensure that <application>pg_ctl</> will wait for
235+
shutdown (Krystian Bigaj)
236+
</para>
237+
</listitem>
238+
239+
<listitem>
240+
<para>
241+
Reduce risk of network deadlock when using <application>libpq</>'s
242+
non-blocking mode (Heikki Linnakangas)
243+
</para>
244+
245+
<para>
246+
When sending large volumes of data, it's important to drain the input
247+
buffer every so often, in case the server has sent enough response data
248+
to cause it to block on output. (A typical scenario is that the server
249+
is sending a stream of NOTICE messages during <literal>COPY FROM
250+
STDIN</>.) This worked properly in the normal blocking mode, but not
251+
so much in non-blocking mode. We've modified <application>libpq</>
252+
to opportunistically drain input when it can, but a full defense
253+
against this problem requires application cooperation: the application
254+
should watch for socket read-ready as well as write-ready conditions,
255+
and be sure to call <function>PQconsumeInput()</> upon read-ready.
256+
</para>
257+
</listitem>
258+
259+
<listitem>
260+
<para>
261+
Fix array handling in <application>ecpg</> (Michael Meskes)
262+
</para>
263+
</listitem>
264+
265+
<listitem>
266+
<para>
267+
Fix <application>psql</> to sanely handle URIs and conninfo strings as
268+
the first parameter to <command>\connect</>
269+
(David Fetter, Andrew Dunstan, &Aacute;lvaro Herrera)
270+
</para>
271+
272+
<para>
273+
This syntax has been accepted (but undocumented) for a long time, but
274+
previously some parameters might be taken from the old connection
275+
instead of the given string, which was agreed to be undesirable.
276+
</para>
277+
</listitem>
278+
279+
<listitem>
280+
<para>
281+
Suppress incorrect complaints from <application>psql</> on some
282+
platforms that it failed to write <filename>~/.psql_history</> at exit
283+
(Tom Lane)
284+
</para>
285+
286+
<para>
287+
This misbehavior was caused by a workaround for a bug in very old
288+
(pre-2006) versions of <application>libedit</>. We fixed it by
289+
removing the workaround, which will cause a similar failure to appear
290+
for anyone still using such versions of <application>libedit</>.
291+
Recommendation: upgrade that library, or use <application>libreadline</>.
292+
</para>
293+
</listitem>
294+
295+
<listitem>
296+
<para>
297+
Fix <application>pg_dump</>'s rule for deciding which casts are
298+
system-provided casts that should not be dumped (Tom Lane)
299+
</para>
300+
</listitem>
301+
302+
<listitem>
303+
<para>
304+
Fix dumping of views that are just <literal>VALUES(...)</> but have
305+
column aliases (Tom Lane)
306+
</para>
307+
</listitem>
308+
309+
<listitem>
310+
<para>
311+
In <application>pg_upgrade</>, force timeline 1 in the new cluster
312+
(Bruce Momjian)
313+
</para>
314+
315+
<para>
316+
This change prevents upgrade failures caused by bogus complaints about
317+
missing WAL history files.
318+
</para>
319+
</listitem>
320+
321+
<listitem>
322+
<para>
323+
In <application>pg_upgrade</>, check for improperly non-connectable
324+
databases before proceeding
325+
(Bruce Momjian)
326+
</para>
327+
</listitem>
328+
329+
<listitem>
330+
<para>
331+
In <application>pg_upgrade</>, quote directory paths
332+
properly in the generated <literal>delete_old_cluster</> script
333+
(Bruce Momjian)
334+
</para>
335+
</listitem>
336+
337+
<listitem>
338+
<para>
339+
In <application>pg_upgrade</>, preserve database-level freezing info
340+
properly
341+
(Bruce Momjian)
342+
</para>
343+
344+
<para>
345+
This oversight could cause missing-clog-file errors for tables within
346+
the <literal>postgres</> and <literal>template1</> databases.
347+
</para>
348+
</listitem>
349+
350+
<listitem>
351+
<para>
352+
Run <application>pg_upgrade</> and <application>pg_resetxlog</> with
353+
restricted privileges on Windows, so that they don't fail when run by
354+
an administrator (Muhammad Asif Naeem)
355+
</para>
356+
</listitem>
357+
358+
<listitem>
359+
<para>
360+
Fix slow sorting algorithm in <filename>contrib/intarray</> (Tom Lane)
361+
</para>
362+
</listitem>
363+
364+
<listitem>
365+
<para>
366+
Fix compile failure on Sparc V8 machines (Rob Rowan)
367+
</para>
368+
</listitem>
369+
370+
<listitem>
371+
<para>
372+
Update time zone data files to <application>tzdata</> release 2015d
373+
for DST law changes in Egypt, Mongolia, and Palestine, plus historical
374+
changes in Canada and Chile. Also adopt revised zone abbreviations for
375+
the America/Adak zone (HST/HDT not HAST/HADT).
376+
</para>
377+
</listitem>
378+
379+
</itemizedlist>
380+
381+
</sect2>
382+
</sect1>
383+
4384
<sect1 id="release-9-0-19">
5385
<title>Release 9.0.19</title>
6386

0 commit comments

Comments
 (0)