Steve Singer [Fri, 26 Jul 2013 17:51:54 +0000 (13:51 -0400)]
Merge branch 'bug306'
Conflicts:
doc/adminguide/slonyupgrade.sgml
Steve Singer [Fri, 26 Jul 2013 14:41:13 +0000 (10:41 -0400)]
release notes update
Steve Singer [Fri, 26 Jul 2013 14:34:37 +0000 (10:34 -0400)]
bug 301 - adding a missing PQclear to fix a memory leak
This PQclear was added by Jan to the patch he posted in bugzilla.
Free the result set when dealing with the error.
Steve Singer [Fri, 26 Jul 2013 14:23:55 +0000 (10:23 -0400)]
Merge branch 'bug301'
Steve Singer [Fri, 26 Jul 2013 14:23:48 +0000 (10:23 -0400)]
Merge branch 'bug299'
Conflicts:
doc/adminguide/slonyupgrade.sgml
Steve Singer [Thu, 25 Jul 2013 19:37:17 +0000 (15:37 -0400)]
add in a check/test to ensure that the backup node in a failover is a forwarding subscriber
If the backup node is not a forwarding subscriber to all of the sets from the origin
then the failover can't complete. It is best to check this early on and warn the user
before the failover actually starts.
Steve Singer [Fri, 19 Jul 2013 22:22:38 +0000 (18:22 -0400)]
add documentation that mentions that the filenames of installable files have changed
Steve Singer [Fri, 26 Jul 2013 12:50:51 +0000 (08:50 -0400)]
The table definitions we create on a 2.1 -> 2.2 upgrade should match the updated DDL in slony1_base.sql
This is to fix bug 306
Steve Singer [Thu, 25 Jul 2013 19:37:17 +0000 (15:37 -0400)]
add in a check/test to ensure that the backup node in a failover is a forwarding subscriber
If the backup node is not a forwarding subscriber to all of the sets from the origin
then the failover can't complete. It is best to check this early on and warn the user
before the failover actually starts.
Steve Singer [Thu, 25 Jul 2013 19:30:06 +0000 (15:30 -0400)]
additional dstring_free() to fix an additional memory leak in sync_helper
Christopher Browne [Tue, 23 Jul 2013 20:02:36 +0000 (16:02 -0400)]
Lots of fixes to anachronisms in FAQ
Christopher Browne [Tue, 23 Jul 2013 20:02:02 +0000 (16:02 -0400)]
Fix typo
Christopher Browne [Mon, 22 Jul 2013 21:29:34 +0000 (17:29 -0400)]
More textual changes
Christopher Browne [Fri, 19 Jul 2013 22:27:02 +0000 (18:27 -0400)]
More fixes to anachonisms
Steve Singer [Fri, 19 Jul 2013 22:22:38 +0000 (18:22 -0400)]
add documentation that mentions that the filenames of installable files have changed
Christopher Browne [Fri, 19 Jul 2013 22:14:26 +0000 (18:14 -0400)]
Shift some anachronisms to the Obsolete section
Christopher Browne [Fri, 19 Jul 2013 19:55:23 +0000 (15:55 -0400)]
Tweak to release notes
Steve Singer [Fri, 19 Jul 2013 19:33:10 +0000 (15:33 -0400)]
documentation sgml fixes
Steve Singer [Fri, 19 Jul 2013 19:31:20 +0000 (15:31 -0400)]
updating minor upgrade instructions to reflect that slony server files are now versioned
and both versions can co-exist in the $libdir and $sharedir at the same time
Steve Singer [Fri, 19 Jul 2013 19:18:26 +0000 (15:18 -0400)]
documentation changes to reflect incompatibilities between slony 2.1.x and 2.2.0
Steve Singer [Fri, 19 Jul 2013 19:38:04 +0000 (15:38 -0400)]
array_length() doesn't exist on PG 8.3 use array_upper() instead
Steve Singer [Fri, 19 Jul 2013 19:36:16 +0000 (15:36 -0400)]
memory leak fixes in sync_helper
dstring_terminate does not call free() to deallocate memory dstring_free() does
Steve Singer [Fri, 19 Jul 2013 19:33:10 +0000 (15:33 -0400)]
documentation sgml fixes
Steve Singer [Fri, 19 Jul 2013 19:31:20 +0000 (15:31 -0400)]
updating minor upgrade instructions to reflect that slony server files are now versioned
and both versions can co-exist in the $libdir and $sharedir at the same time
Steve Singer [Fri, 19 Jul 2013 19:18:26 +0000 (15:18 -0400)]
documentation changes to reflect incompatibilities between slony 2.1.x and 2.2.0
Steve Singer [Thu, 18 Jul 2013 18:03:08 +0000 (14:03 -0400)]
don't process a set if sl_setsync has the initial :1:1: snapshot id
this snapshot id is reserved for setting up the value indicates that
an even (ie the ACCEPT_SET) is still pending.
Also be explicit about the origin this remote worker is querying
sl_setsync for.
Steve Singer [Fri, 12 Jul 2013 14:38:23 +0000 (10:38 -0400)]
bug 299 - put the WAKEUP event at the front of the event queue instead of the back.
This will prevent the remoteWorker thread for an old-origin from pulling
the set that was just moved if a SYNC event is already queued
Steve Singer [Wed, 10 Jul 2013 17:27:21 +0000 (13:27 -0400)]
updating release notes, bug 296,297,298
Steve Singer [Wed, 10 Jul 2013 17:23:56 +0000 (13:23 -0400)]
Merge branch 'bug298'
Steve Singer [Wed, 10 Jul 2013 17:23:51 +0000 (13:23 -0400)]
Merge branch 'bug297'
Jan Wieck [Wed, 10 Jul 2013 16:22:02 +0000 (12:22 -0400)]
Small fixes for failover code.
Steve Singer [Fri, 5 Jul 2013 02:02:38 +0000 (22:02 -0400)]
removing stray quote
Steve Singer [Fri, 5 Jul 2013 02:00:57 +0000 (22:00 -0400)]
more documentation items
Steve Singer [Fri, 5 Jul 2013 02:00:12 +0000 (22:00 -0400)]
add DDL_SCRIPT translation to the logshipping_toinsert
Steve Singer [Thu, 3 Nov 2011 21:04:51 +0000 (17:04 -0400)]
Initial version of a perl script to convert the new
COPY protocol style log shipping files to pure
SQL ones (INSERT,UPDATE,DELETE,TRUNCATE..)
Steve Singer [Thu, 4 Jul 2013 20:40:50 +0000 (16:40 -0400)]
documentation updates for log shipping
Make the slony_logshipper daemon have a proper page in the reference section
like slon and slonik. Clean up the rest of the logshipping instructions to be
more of a set of instructions versus question/answer format.
Steve Singer [Thu, 4 Jul 2013 17:19:24 +0000 (13:19 -0400)]
bug 297 - fix test_slony_state-dbi.pl so that it works with PG 9.2
Also fix the warning emails for sl_confirm and sl_event checks so that they
actually treat problem conditions as problems.
Steve Singer [Thu, 4 Jul 2013 15:47:45 +0000 (11:47 -0400)]
Add a test to LogShipping.js that tests DDL on logshipping nodes
Steve Singer [Wed, 3 Jul 2013 20:51:53 +0000 (16:51 -0400)]
bug 298 - process DDL_SCRIPT events from the log shipper SQL files
Steve Singer [Wed, 3 Jul 2013 20:50:42 +0000 (16:50 -0400)]
improve the error case detection of slony_logshipper
This commit will give a better error message when the log shipper connects to
a databse without the sl_archive_tracking tables.
Steve Singer [Wed, 3 Jul 2013 20:50:16 +0000 (16:50 -0400)]
Documentation fixes for EXECUTE SCRIPT
Steve Singer [Tue, 2 Jul 2013 19:14:31 +0000 (15:14 -0400)]
ensure that the transactions are closed as part of the FAILOVER before doing a MOVE SET
There were some code paths that could leave a transaction open such that when we later
got to a second set or node the LOCK SET could fail.
Also remove an unused if condition
Steve Singer [Fri, 28 Jun 2013 17:14:24 +0000 (13:14 -0400)]
Save the return code for the next iteration
If the moveSet fails as part of a failover we continue processing
for the other nodes. This fix make sure that we remember any failures
and slonik will return a non-zero return code when finished
Steve Singer [Thu, 27 Jun 2013 20:34:54 +0000 (16:34 -0400)]
fixing unit test, issue a sync to make sure the nodes have
caught up before doing the compareDb
Steve Singer [Thu, 27 Jun 2013 19:29:34 +0000 (15:29 -0400)]
documentation updates (FAILOVER)
Steve Singer [Thu, 27 Jun 2013 17:56:02 +0000 (13:56 -0400)]
removing unused variables introduced in the previous commits
Steve Singer [Thu, 27 Jun 2013 16:58:22 +0000 (12:58 -0400)]
DROP NODE will now pass all nodes being dropped to the stored functions and event
This allows slony to detect if a provider that is being dropped
only provides the subscription to some other node that is also
being dropped.
Also update the unit tests to reflect the new method
of FAILOVER involving provider-only nodes
Steve Singer [Wed, 26 Jun 2013 15:31:19 +0000 (11:31 -0400)]
bug 296 - part 1 FAILOVER fixes
This change allows a non-origin node to be specified on the FAILOVER command
so that subscriptions for these nodes are not redirected to the backup node.
An array of all the failed nodes (in addition to the one being failed over at this
stage of the process) is now passed to many of the failover stored functions
This commit also adds the SiteFailover unit test that tests multiple nodes in a site
failing over when the cluster does not have a complete path network
Steve Singer [Thu, 20 Jun 2013 21:22:54 +0000 (17:22 -0400)]
adding documentation note about 8.3 having different table names for pgbench
Steve Singer [Thu, 20 Jun 2013 20:06:58 +0000 (16:06 -0400)]
removing documentation note that applies to PG version prior to 8.0
This version of slony requires at least 8.3
Steve Singer [Thu, 20 Jun 2013 19:54:52 +0000 (15:54 -0400)]
more consistent documentation formatting
Steve Singer [Thu, 20 Jun 2013 19:52:48 +0000 (15:52 -0400)]
bug 294 - fixing documentation error - we just added set 2
Jan Wieck [Fri, 7 Jun 2013 20:06:31 +0000 (16:06 -0400)]
Merge branch 'master' of ssh://git.postgresql.org/slony1-engine
Jan Wieck [Fri, 7 Jun 2013 20:04:13 +0000 (16:04 -0400)]
Fix for bug291.
The apply trigger messed up the query if a row
was updated to all the same values in case a
table has a multi-column primary key.
Jan Wieck [Fri, 7 Jun 2013 19:32:44 +0000 (15:32 -0400)]
Add a test case to trigger bug 291.
Christopher Browne [Fri, 7 Jun 2013 17:32:47 +0000 (13:32 -0400)]
Varlena material has shifted to new location
Steve Singer [Fri, 7 Jun 2013 11:53:43 +0000 (07:53 -0400)]
only include malloc.h on WIN32 - FREEBSD (and others?) have replaced malloc.h with stdlib.h
Steve Singer [Thu, 6 Jun 2013 21:17:10 +0000 (17:17 -0400)]
updating the name of the version for the functions in the win32 export file
Steve Singer [Thu, 6 Jun 2013 21:03:25 +0000 (17:03 -0400)]
add PG VERSION based switch for HAS_LOOKUPEXPLICITNAMESPACE_2 to the MSVC config file
Steve Singer [Thu, 6 Jun 2013 20:47:52 +0000 (16:47 -0400)]
include malloc.h - alloca() is defined here on win32
Steve Singer [Thu, 6 Jun 2013 20:10:11 +0000 (16:10 -0400)]
updating versions for 2.2.0 beta 4
Steve Singer [Thu, 6 Jun 2013 19:09:24 +0000 (15:09 -0400)]
Merge branch 'execute-script-fixes'
Conflicts:
RELEASE
Steve Singer [Thu, 6 Jun 2013 19:05:33 +0000 (15:05 -0400)]
release notes for bug 252
Jan Wieck [Wed, 7 Dec 2011 15:56:42 +0000 (10:56 -0500)]
Fix typo in '<event pending>' conninfo string.
Jan Wieck [Tue, 6 Dec 2011 15:39:35 +0000 (10:39 -0500)]
Change the temporary conninfo on clone prepare to '<event pending>'
for consistency.
Steve Singer [Thu, 6 Jun 2013 18:11:38 +0000 (14:11 -0400)]
release notes update
Steve Singer [Thu, 6 Jun 2013 15:26:30 +0000 (11:26 -0400)]
fixing merge error - string replacements in SQL
Steve Singer [Thu, 6 Jun 2013 15:20:58 +0000 (11:20 -0400)]
Merge branch 'master' into execute-script-fixes
Conflicts:
src/slonik/slonik.c
Steve Singer [Thu, 6 Jun 2013 14:55:01 +0000 (10:55 -0400)]
updating windows compile docs to mention PGVER
Steve Singer [Thu, 6 Jun 2013 14:47:01 +0000 (10:47 -0400)]
Squashed commit of the following:
commit
0d1c7738dc21f2c9e398b790d101f0024dab0a54
Author: Steve Singer <ssinger@ca.afilias.info>
Date: Thu Jun 6 10:45:34 2013 -0400
win32 makefile fixes for the PGVER variable
commit
9f2eed91d129304a0e67534c977394cce04492cf
Merge:
c9a4687 86eac22
Author: Steve Singer <ssinger@ca.afilias.info>
Date: Thu Jun 6 09:42:31 2013 -0400
Merge remote-tracking branch 'origin/master' into pgcommon_22
commit
c9a4687f8fa28cf93a417c0b0597b8c6bf65cea7
Author: Steve Singer <ssinger@ca.afilias.info>
Date: Sun Jun 2 21:39:23 2013 -0400
Link with pgcommon on win32 when building with PG 9.3+
commit
abd5addee71a2fa55a634673730c2d1d18dfe8e9
Author: Steve Singer <ssinger@ca.afilias.info>
Date: Sun Jun 2 21:31:37 2013 -0400
expose HAVE_PGCOMMON to the slonik makefile so it can link in pgcommon if required
Conflicts:
configure.ac
commit
9217d72cdf740e7cb9ca7b540c3a7df334f81e0e
Author: Steve Singer <ssinger@ca.afilias.info>
Date: Tue May 28 10:55:40 2013 -0400
link with pgcommon in addition to pgport on PG 9.3+
PG 9.3 introduces a new library called libpgcommon we need to link
with this when we link with pgport.
Steve Singer [Fri, 31 May 2013 22:22:43 +0000 (18:22 -0400)]
uping version/beta number
Steve Singer [Fri, 31 May 2013 22:18:39 +0000 (18:18 -0400)]
updating release notes bug 290
Steve Singer [Fri, 31 May 2013 22:17:41 +0000 (18:17 -0400)]
versioning all symbols in the shared library.
Slony 2.2 allows multiple versions of the shared library to be exist
on the filesystem($libdir) at once. This means they can be loaded
into the same backend at once, but we need to ensure that the
exported symbols are unique
Steve Singer [Wed, 29 May 2013 21:13:09 +0000 (17:13 -0400)]
Merge branch 'master' into execute-script-fixes
Jan Wieck [Tue, 28 May 2013 19:05:23 +0000 (15:05 -0400)]
Fix the slonik script scanner to accept two single quotes
inside a single quoted string.
Jan Wieck [Tue, 28 May 2013 17:49:54 +0000 (13:49 -0400)]
Merge branch 'master' into execute-script-fixes
Steve Singer [Mon, 27 May 2013 17:35:19 +0000 (13:35 -0400)]
updating the release notes
Jan Wieck [Sun, 26 May 2013 14:54:40 +0000 (10:54 -0400)]
When upgrading from an earlier version of 2.2 to a newer version of 2.2
the old shared library (that now has the version number in its name) might
not still be present. This change avoids having to call C functions as
part of the slonik queries done during a UPDATE FUNCTIONS script.
Jan Wieck [Mon, 27 May 2013 14:45:56 +0000 (10:45 -0400)]
Fixing several compiler warnings (unused and
possibly used uninitialized things).
Jan Wieck [Fri, 24 May 2013 19:04:32 +0000 (15:04 -0400)]
Add SQL='string' option to EXECUTE SCRIPT.
Jan Wieck [Wed, 22 May 2013 15:55:16 +0000 (11:55 -0400)]
Sorting tokens and keywords.
Jan Wieck [Wed, 22 May 2013 15:32:07 +0000 (11:32 -0400)]
Fixing the magic @STRING@ replacement so that a long line with
such a string crossing a 256 character boundary doesn't fail.
Closing FILE* after reading script files.
Jan Wieck [Thu, 16 May 2013 21:26:38 +0000 (17:26 -0400)]
Merge remote branch 'origin/master'
Jan Wieck [Thu, 16 May 2013 17:37:35 +0000 (13:37 -0400)]
Make log_truncate() SECURITY DEFINER.
postgres [Thu, 16 May 2013 17:37:35 +0000 (13:37 -0400)]
Make log_truncate() SECURITY DEFINER.
Steve Singer [Thu, 28 Mar 2013 15:43:12 +0000 (11:43 -0400)]
Updating release notes for bug 286
Steve Singer [Mon, 18 Mar 2013 20:48:43 +0000 (16:48 -0400)]
Bug 286 PG 9.3 now takes two arguments on LookupExplicitNamespace
Add a configure check to determine if we have the 1 or 2 argument
version of LookupExplicitNamespace and pass the proper number
of arguments
Jan Wieck [Fri, 15 Feb 2013 16:08:07 +0000 (11:08 -0500)]
Merge branch 'master' of ssh://git.postgresql.org/slony1-engine
Jan Wieck [Fri, 15 Feb 2013 14:53:59 +0000 (09:53 -0500)]
Add a regression test that ensures DDL is purging the apply cache.
Steve Singer [Thu, 14 Feb 2013 19:50:40 +0000 (14:50 -0500)]
updates to docs and win32 makefile for MS-Windows.
(cherry picked from commit
aee311a97a2b11ab4b555943ae0e49fb21c449ff)
Conflicts:
doc/adminguide/installation.sgml
Jan Wieck [Wed, 13 Feb 2013 23:13:24 +0000 (18:13 -0500)]
fmgr_info() must be called in the applyCacheContext.
Subsequent calls using the FmgrInfo structure may add more
sub elements to the info. This will be allocated in the same
memory context, the fmgr_info() call was made in. That
context must exist at least long enough as the FmgrInfo
call itself exists.
Steve Singer [Fri, 1 Feb 2013 15:43:27 +0000 (10:43 -0500)]
fixing bug number
Steve Singer [Thu, 31 Jan 2013 22:12:57 +0000 (17:12 -0500)]
updating to release 2.2.0 B3
Steve Singer [Thu, 31 Jan 2013 22:10:30 +0000 (17:10 -0500)]
Bug 285 fix move set race condition.
During a move set it is possible on subscriber node, that isn't the
old or new origin, for remoteWorkerThread_$oldorigin to start
processing a SYNC event before the ACCEPT_SET is processed
by the remoteWorkerThread_$neworigin. If the ACCEPT_SET
transaction then commits before the update sl_setsync
query is executed by the remoteWorkertThread_$oldorigin
the update will update a row for the new origin when
it intends to update the old origin. This messes up
sl_setsync.
This change will add a condition to the WHERE clause so
a remoteWorkerThread won't UPDATE rows for a different
origin
Steve Singer [Mon, 28 Jan 2013 20:28:31 +0000 (15:28 -0500)]
updating documentation to list more recent postgresql versions
Christopher Browne [Thu, 24 Jan 2013 20:19:15 +0000 (15:19 -0500)]
Add release note material on bugs 273, 282
Jan Wieck [Wed, 23 Jan 2013 15:49:47 +0000 (10:49 -0500)]
Merge branch 'bug282'
Jan Wieck [Wed, 23 Jan 2013 15:49:17 +0000 (10:49 -0500)]
Merge branch 'bug273'
Christopher Browne [Thu, 17 Jan 2013 23:00:28 +0000 (18:00 -0500)]
Bug #284 - fixes to tutorial
Jan Wieck [Fri, 14 Dec 2012 20:01:31 +0000 (15:01 -0500)]
Fix for bug273
Do not force the event provider to be part of the providers
unless we don't find any provider at all otherwise.