Steve Singer [Mon, 15 Nov 2010 20:50:05 +0000 (15:50 -0500)]
Fixing a merge issue from commit
535565a4cad1e742b8dbdcaeafba4e1d7daf9c38.
The change did not match what was in REL_2_0_STABLE and did not work as expected.
This patch has a history of merge issues.
Christopher Browne [Fri, 5 Nov 2010 16:03:45 +0000 (12:03 -0400)]
Add release notes for bug #162
Christopher Browne [Fri, 5 Nov 2010 15:57:45 +0000 (11:57 -0400)]
Bug #162: refactoring of typenameTypeId() requires using different
function signatures with different PG versions.
Notably, on 9.1, it's used as
typenameTypeId(NULL, txid_snapshot_typname)
Per refactoring commit:
http://git.postgresql.org/gitweb?p=postgresql.git;a=commit;h=
35670340f57d78d6ab023b1fb42fd81339f85d4c
We were already detecting 1/2/3 parm versions of this function, so
all that's necessary is to add #ifdef to choose between them in
src/backend/slony1_funcs.c
Steve Singer [Mon, 15 Nov 2010 15:59:59 +0000 (10:59 -0500)]
Bug 155.
Fix segfaults that sometimes result the postgresql connections going away.
Check for a closed file descriptor (-1) before calling FD_ISSET
(cherry picked from commit
8cd3d07e6af3752234e781ba6f7716840b8e6eb9)
Christopher Browne [Wed, 10 Nov 2010 19:47:39 +0000 (14:47 -0500)]
Add release notes for bug #156
Christopher Browne [Wed, 8 Sep 2010 21:47:53 +0000 (17:47 -0400)]
Bug #156: Add health check to slon
Steve Singer [Wed, 10 Nov 2010 16:12:24 +0000 (11:12 -0500)]
Adding section to security on passwords.
Describes how standard pg_hba.conf options exist and how you can
put passwords in the conninfo or use .pgpass
Christopher Browne [Tue, 9 Nov 2010 22:04:47 +0000 (17:04 -0500)]
Drop out marginally useful dropthings documentation
Christopher Browne [Tue, 9 Nov 2010 21:57:10 +0000 (16:57 -0500)]
Drop not-very-useful defineset document
Christopher Browne [Tue, 9 Nov 2010 21:53:55 +0000 (16:53 -0500)]
shift testbed, using slonik sections into appendix, as they're not totally to be deprecated.
Christopher Browne [Tue, 9 Nov 2010 21:10:17 +0000 (16:10 -0500)]
Dropped best practices section, as the important material has been
drawn into other sections.
Add triggers section back to "advanced topics."
Christopher Browne [Tue, 9 Nov 2010 21:09:55 +0000 (16:09 -0500)]
Drop best
Christopher Browne [Tue, 9 Nov 2010 17:44:28 +0000 (12:44 -0500)]
Remove material put into security section from the best practices section, some reformatting/tagging of security material
Christopher Browne [Tue, 9 Nov 2010 17:32:16 +0000 (12:32 -0500)]
Shifted TZ discussion out of best practices into prerequisites.sgml
Shifted superuser discussion into security.sgml, and reformatted somewhat.
Some other tagging fixes as consequence
Christopher Browne [Thu, 4 Nov 2010 20:43:51 +0000 (16:43 -0400)]
Bug 144 - make it clearer that "add partition/empty table" functions
need to run inside EXECUTE SCRIPT context
Steve Singer [Tue, 19 Oct 2010 20:03:08 +0000 (16:03 -0400)]
Moving locking from the "Left out" section.
It now shows up as an advanced chapter.
Steve Singer [Tue, 19 Oct 2010 19:53:42 +0000 (15:53 -0400)]
Documentation changes: Making the section on event cleanup more clear.
Conflicts:
doc/adminguide/events.sgml
Steve Singer [Tue, 19 Oct 2010 19:44:42 +0000 (15:44 -0400)]
Adding the locking page back into the 2.0 docs
Steve Singer [Tue, 19 Oct 2010 19:21:44 +0000 (15:21 -0400)]
Documentation patch. Mention that slony likes to take exclusive locks out on
sl_event and that this can conflict with a pg_dump.
Thanks Aleksey Tsalolikhin
(cherry picked from commit
5bf39d55b832ea04f7769a75cdfb4d1e7d61169a)
Steve Singer [Tue, 19 Oct 2010 17:02:19 +0000 (13:02 -0400)]
Documentation changes: Making the section on event cleanup more clear.
(cherry picked from commit
b39d1528480b298c90dfbe20579971c920bcbcf1)
Steve Singer [Fri, 8 Oct 2010 14:32:52 +0000 (10:32 -0400)]
Removing extranous '//' from the perl script.
Bug 158
(cherry picked from commit
66d2a80d3accde0d1686d83de7c5b36f8c14e4ac)
Steve Singer [Fri, 8 Oct 2010 13:48:38 +0000 (09:48 -0400)]
Adding "wait on" option to the example of a moveset.
(cherry picked from commit
9cef0f3621cbce3f1aa6abba5f06f750ba7db308)
Christopher Browne [Wed, 6 Oct 2010 16:34:20 +0000 (12:34 -0400)]
Add new documentation pages to gitignore
Christopher Browne [Wed, 6 Oct 2010 16:27:26 +0000 (12:27 -0400)]
Merge branch 'master' of ssh://git.postgresql.org/slony1-engine
Conflicts:
doc/adminguide/addthings.sgml
doc/adminguide/ddlchanges.sgml
doc/adminguide/performance.sgml
doc/adminguide/security.sgml
doc/adminguide/slonik_ref.sgml
doc/adminguide/slony.sgml
doc/adminguide/slonyupgrade.sgml
doc/adminguide/startingstopping.sgml
Christopher Browne [Wed, 6 Oct 2010 16:16:37 +0000 (12:16 -0400)]
Cleaning up errors when building documentation tree.
- Quite a few SGML tagging errors
- References to various portions of the documentation were broken
because some sections were being left out
- Some functions have changed arguments, hence references needed
to change
Steve Singer [Wed, 6 Oct 2010 15:30:02 +0000 (11:30 -0400)]
Merging/moving back in the new to 2.1 section on lowering access with security definers
Steve Singer [Wed, 6 Oct 2010 15:18:22 +0000 (11:18 -0400)]
Fixing documentation issues resulting from merging in changes from REL_2_0_STABLE
Steve Singer [Wed, 6 Oct 2010 14:59:25 +0000 (10:59 -0400)]
Fixing missing link issues in the documentation.
Either by re-adding a few sections in that probably should not have
been removed or by removing the links.
(cherry picked from commit
51949482d3cca3048a55ff9bb982e4a69c6bbb99)
Steve Singer [Wed, 6 Oct 2010 14:30:55 +0000 (10:30 -0400)]
Fixing up various tagging issues causing errors in the documentation build
(cherry picked from commit
4235c5373f2831e0646ce1b8aafc5f468b51330e)
Christopher Browne [Tue, 5 Oct 2010 22:05:05 +0000 (18:05 -0400)]
Merge branch 'master' of ssh://git.postgresql.org/slony1-engine
Conflicts:
doc/adminguide/addthings.sgml
doc/adminguide/firstdb.sgml
doc/adminguide/legal.sgml
Christopher Browne [Tue, 5 Oct 2010 22:00:41 +0000 (18:00 -0400)]
Walked through SGML files to add missing <indexterm> entries.
Christopher Browne [Tue, 5 Oct 2010 22:00:30 +0000 (18:00 -0400)]
Add bookindex label when generating blank bookindex.sgml so that spurious i****.html file is not generated
Steve Singer [Fri, 1 Oct 2010 18:21:49 +0000 (14:21 -0400)]
Rework of the adminguide.
1) Divided the guide into chapters
2) Rewrote various sections including the introduction
3) Added sections on events, security, performance
4) Removed the "best practices" section and put the content that
still applies elsewhere
Various other edits
Merge from 2_0_STABLE
Conflicts:
doc/adminguide/help.sgml
doc/adminguide/intro.sgml
Steve Singer [Fri, 24 Sep 2010 23:14:50 +0000 (19:14 -0400)]
Documentation updates to reflect auto vacuum changes.
Patch from Aleksey Tsalolikhin
(cherry picked from commit
11b4d3bac5ce5177fa17312db254aacc9486c230)
Christopher Browne [Mon, 20 Sep 2010 19:32:38 +0000 (15:32 -0400)]
Tagging fixes
Christopher Browne [Mon, 20 Sep 2010 19:16:55 +0000 (15:16 -0400)]
Fix tagging problems with admin guide docs.
Looks like some <para> tags got missed in previous patches.
Conflicts:
doc/adminguide/bestpractices.sgml
Steve Singer [Mon, 20 Sep 2010 13:35:40 +0000 (09:35 -0400)]
Documentation updates for 2.0.5.
Removing documentation that does not apply to 2.0.x
For older versions see the adminguide for that version.
(cherry picked from commit
f6e236dd80a9ef824a3760b73708d1cefd8d95c2)
Conflicts:
doc/adminguide/addthings.sgml
doc/adminguide/bestpractices.sgml
doc/adminguide/defineset.sgml
doc/adminguide/faq.sgml
doc/adminguide/slonik_ref.sgml
doc/adminguide/slony.sgml
doc/adminguide/slonyupgrade.sgml
doc/adminguide/subscribenodes.sgml
Steve Singer [Fri, 17 Sep 2010 14:49:43 +0000 (10:49 -0400)]
1. The update sl_subscribe statement was duplicated,
the exact same SQL was listed twice.
2. Do not delete the sl_subscribe row for the backup node.
We were finding that a slave could end up with data the
backup node did not have when it became a master because
the backup node was ignoring the sync because it didn't
think it had a provider.
This reverts parts of
d50fba024aac79020549973e98a4ea9d5bccdab5
(cherry picked from commit
7970189ec93faf3ee71cf49e97529980bbd219aa)
Steve Singer [Mon, 13 Sep 2010 20:46:23 +0000 (16:46 -0400)]
Fixes for win32 compiles.
-Try to include pgbindir in the path when we invoke the test program to check PQisthreadsafe so libpq.dll can
be found.
-slon_terminate_worker is not used on win32 (slon does not fork)
-updating expected results(parsestatement test) to match what is currently generated on win32
(cherry picked from commit
6c7c91252f437e9dd5c86b6f2a0f92840c135b6d)
Steve Singer [Mon, 13 Sep 2010 20:38:25 +0000 (16:38 -0400)]
Fix for bug154.
Call updateRelName from the ddlScript_complete_int function instead of ddlScript_complete
because we want this function to be called on both subscribers and the origin.
This issue addresses something like renaming a table via EXECUTE SCRIPT where
sl_table will continue to have the old name on subscribers.
(cherry picked from commit
dfd8816556017fb0a3316b799accf617907c9b62)
Steve Singer [Mon, 13 Sep 2010 20:37:27 +0000 (16:37 -0400)]
Replacing calls to signal with sigaction.
AIX + Solaris have an implementation of
signal that resets the signal handler after it is called once.
Keep calls to signal() on CYGWIN
(cherry picked from commit
7a6c8c0dced9fed8f806dabe8c7e11fbd8d8b18c)
Steve Singer [Mon, 13 Sep 2010 20:37:07 +0000 (16:37 -0400)]
Additional attempt at bug #40.
If libpq isn't already in the system shared library search
path directories then it might not be found when configure
tries to run the test program.
Build the path to libpq into the test executable via linker
flags. The linker flags for this vary based on platform.
The logic was based on logic in postgresql/src/makefiles
in the Postgresql source tree.
Also perform a runtime check to make sure the runtime
libpq is thread-safe.
(cherry picked from commit
2ee266a20be37964b6aa0df1554a0b5998fbba4a)
Steve Singer [Tue, 7 Sep 2010 17:11:23 +0000 (13:11 -0400)]
Bug 149.
Fix for when local listen fails to start properly.
It was observed that slon can wait for the local listen cond to be set where the local listener
thread stops on an error. In this case we want to signal the mutex but flag the local listener
as not having started. The main slon thread should then abort.
This patch had previously been reverted due to an improper merge. This
version should have been merged properly.
(cherry picked from commit
69b54a2625384010f4bee355fd49853a56d6cd6c)
Steve Singer [Tue, 7 Sep 2010 13:36:59 +0000 (09:36 -0400)]
Fix for Bug 121.
Issue the subscribe request against the origin not the provider.
We sometimes get situations where the subscribe request goes
and gets issues on the provider and reaches the origin before
it reaches the receiver.
The origin generates the ENABLE_SUBSCRIPTION as being originated
on the origin. The ENABLE_SUBSCRIPTION can then reach the
receiver before the SUBSCRIBE_SET reaches the receiver.
This causes some of these node -1 errors.
(cherry picked from commit
0b59bd96e8eebbd97778185b7b29dfa696b56927)
Christopher Browne [Mon, 30 Aug 2010 19:30:28 +0000 (15:30 -0400)]
Bug #40 - apply to HEAD
Christopher Browne [Fri, 27 Aug 2010 22:26:45 +0000 (18:26 -0400)]
Add more files ignorable by git because they are generated as parts of builds
Jan Wieck [Tue, 24 Aug 2010 21:17:00 +0000 (17:17 -0400)]
Bug #115
Cause the parser to generate an ERROR if a WAIT FOR EVENT
occurs inside of a TRY block.
In reality, the wait itself isn't the problem. It is the combination
of creating the event (sync, store whatever) and waiting for it
inside the same try block. Try blocks don't commit until they end,
so the event never gets committed and logically, the wait must wait
forever.
(cherry picked from commit
783eee0a78628e62ae081fbebd366f9d3fad5af8)
Steve Singer [Tue, 13 Jul 2010 13:02:54 +0000 (09:02 -0400)]
Fix for when local listen fails to start properly. It was observed that slon can wait for the local listen cond to be set where the local listener thread stops on an error. In this case we want to signal the mutex but flag the local listener as not having started. The main slon thread should then abort.
(cherry picked from commit
ba234cfe686f33fd10d2af1c557af66f8ec17b02)
Steve Singer [Fri, 27 Aug 2010 18:17:33 +0000 (14:17 -0400)]
Additional bug 132 fixes:
The worker will only restart when it can't get the nodelock if
the worker is in a 'restart' state because it previously
exited asking its watchdog to restart.
To faciliate this the watchdog will no longer exec when it needs
to restart its child but instead just call fork() again.
Also removing more slon_log instances from the signal handler
since they are no async safe
Additional fix for bug136:
-Do not use evcommand when calling reshape subscriber since this
function does not generate an event.
Steve Singer [Tue, 24 Aug 2010 13:28:14 +0000 (09:28 -0400)]
Only restart the slon during a reshape if sl_subscribe was
actually updated.
Bug 136 additional fixes
(cherry picked from commit
3c7b70e40bdb913c65e0c5499bc3ac25268da5d4)
Christopher Browne [Wed, 18 Aug 2010 16:22:00 +0000 (12:22 -0400)]
Change query looking for the need to "reshape subscription" to
return a count of nodes, rather than the list of nodes.
This is effectively a minor performance tweak.
Steve Singer [Thu, 5 Aug 2010 14:45:35 +0000 (10:45 -0400)]
If the nodelock can not be obtained then the worker will keep retrying instead
of exiting. Bug # 132 discussses issues where slon can't restart because
of a duplicate node lock. The watchdog instance of slon can kill the worker
child and restart it. The issue is that postgresql backend for the old
worker might still be around when the new worker tries to get the nodelock.
Prior to this patch we only had one shot at this, now we will keep retrying
until a user decides to kill this process or the other one goes away
(cherry picked from commit
acd46819bad1613764708b138ebcfa895467ac51)
Steve Singer [Thu, 5 Aug 2010 15:55:43 +0000 (11:55 -0400)]
Install the signal handlers after the fork.
sighandler is intended to be used by the watchdog instance not the worker
This can be considered part of bug 139 (non async safe functions being called
by a signal handler)
(cherry picked from commit
173e7786ddc8ec2ab2b9299fc991573d34e816ba)
Steve Singer [Wed, 4 Aug 2010 12:17:41 +0000 (08:17 -0400)]
Replace usage of alarm() for restarting slon with a KILL of the child
and a sleep before restarting.
Numerous instances were observed where the parent/watchdog slon instance
would exit on a SIGALRM. The various POSIX standards define rules about
what functions can be called by a signal handler. We suspect not following
these guidelines was part of the problem.
Steve Singer [Fri, 27 Aug 2010 18:28:40 +0000 (14:28 -0400)]
Bug 136
When a forwarding node goes away a dba might issue a subscribe set to reshape the
cluster so the receiver gets the data from a different provider.
It could be the case that the receiver never sees the SUBSCRIBE_SET event
because it is only listening on events from the origin from the forwarder
that has gone away.
If we are reshaping a cluster then slonik will contact the receiver
directly and run update sl_subscribe so the receiver can
listen from the correct source.
(cherry picked from commit
0f8305660488d508df6a6ea96bf8ebf57e2e77ea)
Conflicts:
src/backend/slony1_funcs.sql
Steve Singer [Fri, 30 Jul 2010 19:55:17 +0000 (15:55 -0400)]
When doing a fail node where the failed node is not a provider of the backup node
we need to delete any subscribe entries that are used to forward data to the backup
node. These will not be needed when the backup node becomes the new origin.
Until that happens they get in the way of the FAILOVER_SET message
that will be posted on the 'most ahead node' but iwth an origin of
the old origin from getting to this node.
(cherry picked from commit
d50fba024aac79020549973e98a4ea9d5bccdab5)
Steve Singer [Thu, 29 Jul 2010 19:04:25 +0000 (15:04 -0400)]
If a origin node fails a subscriber might
not be listening for events from the new origin. The FAILOVER event gets
processed on the subscriber but nothing after might be since the
sl_listen rows are setup so all events from the new origin get routed
through the node that used to be the provider(even if a direct
path from the new origin to the subscriber exists).
On processing the FAILOVER event we will mark the old origin as no
inactive. RebuildListenEntries will now not remove paths that are
dedundent if the (otherwise) preferred path flows through a disabled
node.
(cherry picked from commit
60566590d683b85733404ef290e6c1823c4c014c)
Jan Wieck [Fri, 27 Aug 2010 17:00:05 +0000 (13:00 -0400)]
Bug 124
Enforce that EVENT NODE is identical to EXECUTE ONLY ON node if
the latter is specified.
Also restore the session_replication_role in case ddlScript_prepare
changed it to local.
(cherry picked from commit
c6de292bbb3fea189cece38d6bf5bf7c22a9fa92)
Jan Wieck [Wed, 25 Aug 2010 20:41:26 +0000 (16:41 -0400)]
Bug #117
Change logging call to omit the write to stdout if we've got and are using only syslog (use_syslog = 2)
(cherry picked from commit
9eb5bce628d7feafb422c6f4b7e65518c8f1bc22)
Jan Wieck [Tue, 24 Aug 2010 21:59:48 +0000 (17:59 -0400)]
Bug 75
Call db_commit_xact() at the end of slonik's db_connect() to make
the session settings for datestyle and replication role permanent.
(cherry picked from commit
3d8606973c4ed5790787cb8b87e662edbb00c3ee)
Steve Singer [Mon, 23 Aug 2010 17:01:55 +0000 (13:01 -0400)]
Fix for bug #138
Fix synatx in autoconf m4 file to define GROFF
(cherry picked from commit
b4f202e3f67e66a311a4a32c6392bb2be3bc05a2)
Conflicts:
RELEASE
Steve Singer [Mon, 23 Aug 2010 16:56:32 +0000 (12:56 -0400)]
check fd for negatvie values. Bug 104 proper fix
(cherry picked from commit
65af4b6002e945c3df977d12b8613b98a379113a)
Christopher Browne [Tue, 17 Aug 2010 17:12:33 +0000 (13:12 -0400)]
generated SGML should be ignored by git
Christopher Browne [Tue, 17 Aug 2010 17:11:49 +0000 (13:11 -0400)]
Remove reference to logswitch_weekly(), per email from TAKATSUKA Haruka <harukat@sraoss.co.jp> 2010-06-24
Christopher Browne [Mon, 16 Aug 2010 21:24:19 +0000 (17:24 -0400)]
Bug #79 - add better documentation about configure options
Jan Wieck [Sat, 14 Aug 2010 03:45:01 +0000 (23:45 -0400)]
Making ducttape scripts executable again.
Jan Wieck [Fri, 13 Aug 2010 19:42:43 +0000 (15:42 -0400)]
Removed definition of PGAUTODOC. The variable is set in Makefile.global
by configure.
Jan Wieck [Fri, 13 Aug 2010 19:34:10 +0000 (15:34 -0400)]
Merge branch 'master', remote branch 'origin'
Christopher Browne [Fri, 13 Aug 2010 19:29:48 +0000 (15:29 -0400)]
More removals of references to TABLE ADD KEY
Christopher Browne [Fri, 13 Aug 2010 19:29:06 +0000 (15:29 -0400)]
cleanup event function HTML should be ignored
Christopher Browne [Fri, 13 Aug 2010 19:16:25 +0000 (15:16 -0400)]
TABLE ADD KEY was still indirectly referenced in the documentation in several places. Cleaning that out
Christopher Browne [Thu, 12 Aug 2010 17:25:20 +0000 (13:25 -0400)]
Merge branch 'master' into bug-146-HEAD
Conflicts:
src/backend/slony1_funcs.sql
tests/test1/generate_dml.sh
Merged HEAD changes
Christopher Browne [Thu, 12 Aug 2010 17:08:13 +0000 (13:08 -0400)]
Merge branch 'bug148-head'
Steve Singer [Wed, 11 Aug 2010 21:41:18 +0000 (17:41 -0400)]
Bug 42:
The problem is that postgresql_autodoc produces XML but the documentation is
processed using SGML. SGML does not process the empty tag notation <xref .../>
correctly. Instead it should be <xref ...> without slash. I have attached a
patch that fixes up the XML to be valid SGML so the documentation comes out
right.
Patch by Peter Eisentraut
Christopher Browne [Wed, 11 Aug 2010 21:36:39 +0000 (17:36 -0400)]
Merge branch 'master' of ssh://git.postgresql.org/slony1-engine
Steve Singer [Wed, 11 Aug 2010 21:13:36 +0000 (17:13 -0400)]
Removing admin guide entry for TABLE ADD KEY since the command
no longer exists.
See bug 147
(cherry picked from commit
a2c1cd68c242bc8833b599bb7edb8aef730ed27f)
Steve Singer [Wed, 11 Aug 2010 21:12:23 +0000 (17:12 -0400)]
Bug 147
Remove TABLE ADD KEY from altperl scripts.
Slony-I 2.0.x does not support TABLE ADD KEY so we should not
be generating slonik scripts that try and use it.
Instead generate errors if we find instances where it is required
(cherry picked from commit
9d827a40bf3a1fdaee599628c439870c7d410848)
Conflicts:
RELEASE
Steve Singer [Wed, 11 Aug 2010 21:00:38 +0000 (17:00 -0400)]
Fix for bug #127.
Generation of syns on leaf nodes (nodes that are not set origins) were disabled some time ago. This reverts
that commit and re-enables them
(cherry picked from commit
572581c4d80fd95206aed1910fb22d7119aa4be4)
Steve Singer [Wed, 11 Aug 2010 20:58:56 +0000 (16:58 -0400)]
Fix for Bug #120.
Call enableNode as part of the CLONE NODE processing. This will start the worker thread for the cloned node.
(cherry picked from commit
0d1a293927805e5728c39f8000031bd698ab49c4)
Christopher Browne [Wed, 11 Aug 2010 16:29:24 +0000 (12:29 -0400)]
Remove sl_log_1/2 from list of tables to be vacuumed, per comment from Jan in bug #78
Christopher Browne [Wed, 11 Aug 2010 16:23:52 +0000 (12:23 -0400)]
Bug #78
http://www.slony.info/bugzilla/show_bug.cgi?id=78
"Christopher Browne 2010-06-02 13:20:53 PDT
We should deprecate the "cleanup_deletelogs" parameter in 2.0.5 or so."
Christopher Browne [Tue, 10 Aug 2010 20:33:29 +0000 (16:33 -0400)]
Bug #148
When ShouldSlonyVacuumTable() asks if a table should be vacuumed, the first test should be:
Does the table exist altogether???
If not, then there's surely no need to vacuum it
'
Christopher Browne [Mon, 9 Aug 2010 19:59:19 +0000 (15:59 -0400)]
Bug #142 - Minimize security profile
The two places where superuser access is *truly required* have been
shifted to functions where they could readily be:
a) Turned into security definers, and
b) Locked down so only the slony management user(s) can run them
Also added documentation on this to "best practices."
Christopher Browne [Mon, 9 Aug 2010 19:58:44 +0000 (15:58 -0400)]
Additional HTML files to be ignored as they are generated from doc tree
Steve Singer [Fri, 6 Aug 2010 18:13:47 +0000 (14:13 -0400)]
Fix for bug # 122
Print a better error message when using a receiver node id that does not exist
(cherry picked from commit
03db9e0b2c1625aafdaedbf6bf3730c8889cbb2f)
Christopher Browne [Thu, 5 Aug 2010 16:02:30 +0000 (12:02 -0400)]
Bug #141 - removing alterTableRestore()
- Applying this to HEAD
Christopher Browne [Wed, 4 Aug 2010 20:14:00 +0000 (16:14 -0400)]
Merge branch 'master' of git://git.postgresql.org/git/slony1-engine into bug-146-HEAD
Conflicts:
config.h.in
Christopher Browne [Wed, 4 Aug 2010 20:13:19 +0000 (16:13 -0400)]
Initial swing at removing cleanup_deletelogs
Christopher Browne [Wed, 4 Aug 2010 20:12:33 +0000 (16:12 -0400)]
Make scripts executable
Christopher Browne [Wed, 4 Aug 2010 20:11:41 +0000 (16:11 -0400)]
Changing version number
Christopher Browne [Wed, 4 Aug 2010 20:08:21 +0000 (16:08 -0400)]
Make shell scripts executable - expected by tests
Christopher Browne [Wed, 4 Aug 2010 20:07:58 +0000 (16:07 -0400)]
Set version number to 2.1.0 for HEAD
Christopher Browne [Wed, 4 Aug 2010 20:06:49 +0000 (16:06 -0400)]
Add more generated files to .gitignore
Christopher Browne [Wed, 4 Aug 2010 18:58:31 +0000 (14:58 -0400)]
Add FAQ describing recent SSL-related problem experienced by Karl
Denninger
Steve Singer [Wed, 4 Aug 2010 14:13:48 +0000 (10:13 -0400)]
Documentation updates. Mostly typos and minor corrections.
These changes are based on a list of issues submitted by
Tomonari Katsumata, NTT Open Source Software Center
Thanks
Steve Singer [Thu, 29 Jul 2010 15:37:44 +0000 (11:37 -0400)]
Fixed some broken logic in failedNode()
It apparently never supported failing over to nodes that are not
a direct subscriber to the failed node.
Jan
(cherry picked from commit
2705654ec3dcb0381afef591756170c6de8f8ce2)
Conflicts:
src/backend/slony1_funcs.sql
Steve Singer [Wed, 28 Jul 2010 21:12:32 +0000 (17:12 -0400)]
Updating version number to 2.0.4
No longer generate warning of a unsupported version with Pg 9.0
(cherry picked from commit
2d645b3fa1716d649345e8edbb27b188d9505728)
Conflicts:
config.h.in
src/backend/slony1_funcs.sql
src/slonik/slonik.c
Steve Singer [Wed, 28 Jul 2010 21:09:27 +0000 (17:09 -0400)]
Bug #95 - slony.info/bugzilla/show_bug.cgi?id=95
There are two calls to the postgres function subscribeSet_int(integer, integer,
integer, boolean, boolean) in slony1-engine/src/slonik/slonik.c
Both calls do not pass in the last boolean input (the new OMIT COPY parameter)
and hence you get errors. We have only seen it when using slonik_store_node
but looking at the code it could also happen in slonik_failed_node.
(cherry picked from commit
ba90f82da5778648e53d0d742b1fa89901a794e8)
Conflicts:
src/slonik/slonik.c
Steve Singer [Wed, 28 Jul 2010 21:06:12 +0000 (17:06 -0400)]
Add in OMIT COPY option to SUBSCRIBE SET in support of upgrading from
elder Slony-I versions.
(cherry picked from commit
dc3f318983213f8fb90eec0545e60dff61f1f55d)
Conflicts:
doc/adminguide/adminscripts.sgml
doc/adminguide/intro.sgml
doc/adminguide/slonyupgrade.sgml
src/backend/slony1_funcs.sql
src/slon/remote_worker.c
src/slonik/parser.y
src/slonik/scan.l
src/slonik/slonik.c
src/slonik/slonik.h
tests/run_test.sh
Steve Singer [Wed, 28 Jul 2010 20:27:19 +0000 (16:27 -0400)]
Perform a va_copy and use the copy in the loop that uses va_args.
You are not allowed to use va_args multiple times.
Fix for bug # 55
Patch submitted by Paul Cannon
(cherry picked from commit
f98048352362c990d51a694583bfaa78736cd33b)
Conflicts:
src/slon/misc.c