Craig Ringer [Mon, 25 May 2015 14:38:09 +0000 (22:38 +0800)]
fixup! Add update path from 0.9.0.5 to 0.10.0.0
The Makefile.in must also know about the new files
Craig Ringer [Mon, 25 May 2015 05:57:33 +0000 (13:57 +0800)]
Add update path from 0.9.0.5 to 0.10.0.0
commit
401d117, bdr.table_set_replication_sets ...
introduces 0.9.0.5 and we need a path from it to 0.10.0.x.
Andres Freund [Mon, 25 May 2015 01:49:25 +0000 (03:49 +0200)]
bdr: Send pgstat upon replayed commit in apply process.
Previously it was only sent at apply worker shutdown.
Andres Freund [Mon, 25 May 2015 01:43:15 +0000 (03:43 +0200)]
bdr: Improve sequencer locking and pgstat reporting.
Up to now the sequencer locked its relations exclusively to prevent an
interlock with apply. That was problematic, because it would often
prevent auto vacuum/analyze from running, or interrupt it.
Instead of locking the underlying relations lock the bdr seqam object
and add code to apply to also acquire that lock. That allow autovacuum
to run undisturbed.
Testing brought an additional problem to light, namely that pgstat
records weren't sent in the sequencer until shutdown, which means that
in many situations autovacuum won't be trigger in time. This is a more
widespread problem than just the sequencer, but here only the sequencer
is fixed.
Andres Freund [Sun, 24 May 2015 23:46:02 +0000 (01:46 +0200)]
bdr: Adjust database name in debug message.
Andres Freund [Thu, 21 May 2015 21:36:41 +0000 (23:36 +0200)]
bdr: Add minimal checks for the cloning of databases with contents.
These should be expanded, but are better than nothing.
Petr Jelinek [Fri, 22 May 2015 13:40:16 +0000 (15:40 +0200)]
Add the 0.10.0.2 to Makefile
Craig Ringer [Fri, 15 May 2015 09:08:52 +0000 (17:08 +0800)]
bdr.table_set_replication_sets should permit unsafe commands
In
4005c1d bdr_replicate_ddl_command was changed not to override the
command filter, but this broke replication sets in UDR, since they have
to use SECURITY LABEL.
SET bdr.permit_unsafe_ddl_commands on bdr.table_set_replication_sets
since it knows what it's doing.
Fixes #67
Note: The ddl/replication_sets tests cannot be run as-is on UDR because
UDR does not replicate changes to replication sets. We really need
separate tests for this that cover UDR (or BDR+subscribe).
Andres Freund [Thu, 21 May 2015 21:03:30 +0000 (23:03 +0200)]
Don't specify the sequence am in UDR.
The initial cloning needs more tests, including data, but that requires
larger changes to several tests than I want to do in the fix commit
itself.
Fixes #69
Andres Freund [Fri, 15 May 2015 14:03:01 +0000 (16:03 +0200)]
Initial code to actually remove slots when parting a node.
This removes a slot when
a) It is of a remote node that has been parted.
b) We have been parted, and there are remote slots. It is not guaranteed
that we're notified of that case, but it seems worthwhile to handle
nonetheless.
Craig Ringer [Fri, 15 May 2015 09:26:08 +0000 (17:26 +0800)]
fixup! Add a test for node part
Didn't check in the expected/ data
Craig Ringer [Fri, 15 May 2015 09:18:47 +0000 (17:18 +0800)]
Fix UDR after BDRLockType introduction in
10b331fea
Craig Ringer [Fri, 15 May 2015 09:13:26 +0000 (17:13 +0800)]
Better error message if PG_CONFIG is nonsense
Petr Jelinek [Wed, 6 May 2015 14:51:53 +0000 (16:51 +0200)]
Use correct replication set and apply delay config in apply.
Craig Ringer [Wed, 6 May 2015 09:09:59 +0000 (17:09 +0800)]
Add a test for node part
This shows that the slots are currently _not_ correctly dropped.
Petr Jelinek [Tue, 5 May 2015 15:37:50 +0000 (17:37 +0200)]
Use the new lesser form of global lock for some of the DDL commands.
Petr Jelinek [Tue, 5 May 2015 15:35:30 +0000 (17:35 +0200)]
Add support for more levels of global locks.
The old ddl_lock was replaced by ddl_lock and write_lock.
The ddl_lock now only locks database against concurrent DDL but not
against writes anymore. The write_lock locks against writes the same way
old ddl_lock used to. We support upgrading the ddl_lock to write_lock at
some performance penalty (the lock needs to be negotiated again).
Petr Jelinek [Fri, 1 May 2015 13:56:02 +0000 (15:56 +0200)]
Fix broken supervisordb test
Petr Jelinek [Fri, 1 May 2015 13:51:37 +0000 (15:51 +0200)]
Make DML operation wait when there is DDL in progress.
Previously, we immediately raised ERROR when executing DML while DDL
lock was being acquired or held. Now we just keep them waiting until
either the lock is released or statement_timeout kills them.
Update isolation tests accordingly.
Petr Jelinek [Fri, 1 May 2015 13:23:40 +0000 (15:23 +0200)]
Don't immediately kill all transactions when acquiring DDL lock.
Previously we just canceled all transactions in database when acking DDL
lock. This code only cancels transactions which did any writes and gives
them some time (currently hardcoded to 10s) of grace period to finish
what they are doing.
Craig Ringer [Fri, 24 Apr 2015 07:17:40 +0000 (15:17 +0800)]
Prevent users from connecting to bdr_supervisordb
This won't stop bgworkers, but they should hopefully know better.
Craig Ringer [Fri, 24 Apr 2015 06:17:39 +0000 (14:17 +0800)]
Prevent attempts to create/drop/rename bdr_supervisordb
When the BDR extension is loaded, BDR its self can create
bdr_supervisordb, but applications cannot. This gives users
a better error message if they try to drop it, too, e.g.
DROP DATABASE bdr_supervisordb;
ERROR: The BDR extension reserves the database name bdr_supervisordb
for its own use
HINT: Use a different database name
Craig Ringer [Fri, 24 Apr 2015 04:38:26 +0000 (12:38 +0800)]
fixup! Prevent taking the DDL lock without explicit authorization
Per internal discussion, make bdr.permit_ddl_locking true by default.
Petr Jelinek [Thu, 23 Apr 2015 17:18:06 +0000 (19:18 +0200)]
Fix typo in previous commit
Petr Jelinek [Wed, 22 Apr 2015 19:13:34 +0000 (21:13 +0200)]
bdr_init_copy: insert correct bdr_node entry
Petr Jelinek [Wed, 22 Apr 2015 18:57:39 +0000 (20:57 +0200)]
bdr_init_copy: properly escape single quotes
Martín Marqués [Wed, 22 Apr 2015 18:50:02 +0000 (15:50 -0300)]
Typo in bdr_init_copy help information
Martín Marqués [Wed, 22 Apr 2015 18:48:03 +0000 (15:48 -0300)]
Revert "Typo in bdr_init_copy help information"
There was an unintended change in the commit
This reverts commit
1064034c898e98b210a8d87f4a65aa2c05d6666b.
Martín Marqués [Wed, 22 Apr 2015 17:44:17 +0000 (14:44 -0300)]
Typo in bdr_init_copy help information
Craig Ringer [Wed, 22 Apr 2015 09:25:50 +0000 (17:25 +0800)]
fixup! Prevent taking the DDL lock without explicit authorization
Missing <para> in the docs
Craig Ringer [Wed, 22 Apr 2015 06:24:19 +0000 (14:24 +0800)]
Make PG_CONFIG a declared precious configure variable
It'll show up in help and appear in the config cache.
Craig Ringer [Wed, 22 Apr 2015 06:21:19 +0000 (14:21 +0800)]
Fix incorrect AC_ARG_VAR usage, make PATH a precious var with decription
This makes configure consider the PATH important and documents that
it affects where we find pg_config .
Martín Marqués [Tue, 21 Apr 2015 17:42:42 +0000 (14:42 -0300)]
Typo in the docs
Craig Ringer [Tue, 21 Apr 2015 13:10:28 +0000 (21:10 +0800)]
Prevent taking the DDL lock without explicit authorization
Attempts to acquire the global DDL lock will now, by default,
ERROR with a message like:
ERROR: Global DDL locking attempt rejected by configuration
DETAIL: bdr.permit_ddl_locking is false and the attempted command would
require the global DDL lock to be acquired. Command rejected.
HINT: See the 'DDL replication chapter of the documentation.
Users/scripts which wish to execute DDL should:
BEGIN;
SET LOCAL bdr.permit_ddl_locking = true;
ALTER TABLE ....
COMMIT;
This will help prevent accidental DDL in things like user reports
from aborting transactions across the system and blocking user
writes. It's particularly useful where the DDL may be combined
with a transaction that does a large amount of row changes, as
there are performance issues with transactions that both do
DDL and change very large numbers of rows.
Future work will focus on progressively making the DDL lock finer
grained and less disruptive.
Craig Ringer [Tue, 21 Apr 2015 10:52:08 +0000 (18:52 +0800)]
Rename 'bdr' reserved db to 'bdr_supervisordb'
Craig Ringer [Tue, 21 Apr 2015 10:37:44 +0000 (18:37 +0800)]
Fix tests broken by
d7efbe8 and incorrect relation name
Craig Ringer [Tue, 21 Apr 2015 03:05:38 +0000 (11:05 +0800)]
More detail on DDL rep, mention use of replicate_ddl_command
Craig Ringer [Mon, 20 Apr 2015 12:28:06 +0000 (20:28 +0800)]
Clarify that you can only monitor lag from an upstream
Per query from the mailing list,
http://www.postgresql.org/message-id/
F38D1020-3B2C-41CF-BECD-
1F1FB5B19806@visi.com
Craig Ringer [Fri, 17 Apr 2015 07:23:59 +0000 (15:23 +0800)]
More docs on DDL locking
Craig Ringer [Thu, 16 Apr 2015 08:09:24 +0000 (16:09 +0800)]
The apply worker wasn't checking for postmaster death when paused
Fixes #58
Petr Jelinek [Sat, 18 Apr 2015 22:29:07 +0000 (00:29 +0200)]
Make sure node_name is not null and unique.
Petr Jelinek [Sat, 18 Apr 2015 22:11:51 +0000 (00:11 +0200)]
fix bdr_init_copy for UDR
Petr Jelinek [Sat, 18 Apr 2015 22:11:23 +0000 (00:11 +0200)]
pg_upgrade compatibility
Petr Jelinek [Sat, 18 Apr 2015 13:45:25 +0000 (15:45 +0200)]
Beginnings of 9.5 compat (UDR only so far)
Petr Jelinek [Sun, 12 Apr 2015 20:00:35 +0000 (22:00 +0200)]
Correct lock level in filter_AlterTableStmt()
Martín Marqués [Wed, 15 Apr 2015 22:39:44 +0000 (19:39 -0300)]
Missing separation coma in the bdr.bdr_group_join example
Martín Marqués [Wed, 15 Apr 2015 22:37:33 +0000 (19:37 -0300)]
Typo in the docs
Craig Ringer [Wed, 15 Apr 2015 03:39:49 +0000 (11:39 +0800)]
Lower log level of 'Found bdr_connections entry ' log lines
These are internal debug/informational messages.
Craig Ringer [Fri, 10 Apr 2015 05:03:34 +0000 (13:03 +0800)]
Fixes for UNLOGGED tables in the command filter
The command filter was checking only for temporary tables when deciding
whether to skip DDL locking, and when performing some operations, but
for others was permitting UNLOGGED tables too.
Consistently use UNLOGGED||TEMP throughout.
Fixes #44
Petr Jelinek [Wed, 8 Apr 2015 00:04:42 +0000 (02:04 +0200)]
Make bdr_replicate_ddl_command() obey the command filter.
Previously, bdr_replicate_ddl_command() would always execute any
command, no matter if it's considered safe or not. But this behaviour is
too much of a footgun for users of this function. Plus we already have
mechanism to override command filter - the
bdr.permit_unsafe_ddl_commands GUC, so it makes sense that
bdr_replicate_ddl_command() function should behave according to this GUC
just like plain DDL in BDR.
Petr Jelinek [Wed, 8 Apr 2015 00:03:34 +0000 (02:03 +0200)]
Rename locks check interface.
Craig Ringer [Tue, 7 Apr 2015 09:24:40 +0000 (17:24 +0800)]
Document pg_xlog_wait_remote_recieve and pg_xlog_wait_remote_apply
Petr Jelinek [Mon, 6 Apr 2015 14:30:44 +0000 (16:30 +0200)]
Don't acquire table locks in command filter.
The filter_AlterTableStmt() function was using
AlterTableGetLockLevel(astmt->cmds) to determine lock level which to use
on relation being ALTERed. This can cause deadlock against bdr_apply
when the apply is trying to catch up with changes from another node to
synchronize cluster for a DDL lock - any concurrent DML that is coming
from another node will never get the row lock grant because the table is
already locked by the backend performing the ALTER on it.
We now get the relid with NoLock which should be safe as we don't allow
concurrent DDL anyway so the table can't change under our hands.
Petr Jelinek [Fri, 3 Apr 2015 23:45:07 +0000 (01:45 +0200)]
Fix global lock crash recovery on multidb setup
Since bdr_locks_startup() is called from perdb worker we should not try
to load the locks for all databases but only for the database the perdb
worker is connected to.
Petr Jelinek [Wed, 1 Apr 2015 20:26:59 +0000 (22:26 +0200)]
doc: Fix typos in function names.
Petr Jelinek [Wed, 1 Apr 2015 16:45:54 +0000 (18:45 +0200)]
Fix bdr.bdr_replication_identifier table definition
The bdr.bdr_replication_identifier was mistakenly defined in extension
SQL file with two additonal columns, but the C code didn't expect them
this lead to memory corruption. It was identified by make check failures
with clang compiler.
Craig Ringer [Wed, 1 Apr 2015 07:28:15 +0000 (15:28 +0800)]
Swap "node management" and "configuration settings" chapters
Per suggestion from Martín.
Closes #38
Craig Ringer [Wed, 1 Apr 2015 06:55:07 +0000 (14:55 +0800)]
Quickstart should be using bdr-plugin/next
Petr Jelinek [Sun, 29 Mar 2015 14:09:47 +0000 (16:09 +0200)]
doc: Print BDR version in the title.
Petr Jelinek [Sun, 29 Mar 2015 13:38:16 +0000 (15:38 +0200)]
doc: Add optinal website-build flag
Petr Jelinek [Fri, 27 Mar 2015 21:50:13 +0000 (22:50 +0100)]
doc: Make the font bit nices
Martín Marqués [Fri, 27 Mar 2015 11:04:00 +0000 (08:04 -0300)]
Several "to to" typos found, which are now fixed.
Petr Jelinek [Thu, 26 Mar 2015 17:00:03 +0000 (18:00 +0100)]
Fix identifier test so that it works for UDR too.
Petr Jelinek [Thu, 26 Mar 2015 16:58:06 +0000 (17:58 +0100)]
Revert inadvertent test change done in
d8fc209d5
Petr Jelinek [Thu, 26 Mar 2015 11:29:19 +0000 (12:29 +0100)]
doc: Remove mentions of bdr-plugin/stable branch from docs
Craig Ringer [Thu, 26 Mar 2015 02:40:07 +0000 (10:40 +0800)]
Document tree structure for github README
Petr Jelinek [Wed, 25 Mar 2015 20:53:53 +0000 (21:53 +0100)]
Fix PK detection on inherited tables.
We don't allow UPDATEs or DELETEs on tables without primary key but the
code for detecting if table has primary key didn't work correctly when
the UPDATE/DELETE was run on parent table which had primary key but the
command has propagated to an inherited table didn't have primary key.
In passing, fix the primary key detection for a less common use case:
WITH foo AS ( UPDATE table_without_pk ... RETURNING * )
INSERT INTO table_with_pk SELECT * FROM foo;
Craig Ringer [Wed, 25 Mar 2015 07:49:57 +0000 (15:49 +0800)]
doc: Be consistent about sudo prefix
Christoph Moench-Tegeder [Wed, 25 Mar 2015 11:28:36 +0000 (12:28 +0100)]
fix link
Ian Barwick [Wed, 25 Mar 2015 02:28:23 +0000 (11:28 +0900)]
doc: misc typo fixes
Ian Barwick [Wed, 25 Mar 2015 02:22:49 +0000 (11:22 +0900)]
doc: fix a couple of xreflabels
Ian Barwick [Wed, 25 Mar 2015 01:47:04 +0000 (10:47 +0900)]
doc: Reword installation requirements slightly
Petr Jelinek [Wed, 25 Mar 2015 00:51:11 +0000 (01:51 +0100)]
doc: psql provided by BDR-PG has seqam support in \ds.
Petr Jelinek [Wed, 25 Mar 2015 00:30:30 +0000 (01:30 +0100)]
doc: Remove incorrect warning about bdr.bdr_replicate_ddl_command().
Petr Jelinek [Tue, 24 Mar 2015 22:11:29 +0000 (23:11 +0100)]
doc: Debian/Ubuntu installation docs
Craig Ringer [Tue, 24 Mar 2015 17:18:45 +0000 (01:18 +0800)]
doc: Write some conflict docs, partly cribbed from wiki, partly new
Craig Ringer [Tue, 24 Mar 2015 14:32:59 +0000 (22:32 +0800)]
Bump version to 0.10.0 after 0.9.0 release
Craig Ringer [Tue, 24 Mar 2015 12:14:00 +0000 (20:14 +0800)]
doc: Mention 2ndQuadrant prominently in top level, relnotes
Craig Ringer [Tue, 24 Mar 2015 12:07:33 +0000 (20:07 +0800)]
doc: Document apt repo key ID
Craig Ringer [Tue, 24 Mar 2015 11:11:49 +0000 (19:11 +0800)]
doc: Make it clear that conflicts and repsets docs are in dev
Craig Ringer [Tue, 24 Mar 2015 10:50:19 +0000 (18:50 +0800)]
doc: Sort relnotes newest to oldest (desc)
Craig Ringer [Tue, 24 Mar 2015 10:49:18 +0000 (18:49 +0800)]
doc: Mention Debian/Ubuntu packages
Craig Ringer [Tue, 24 Mar 2015 10:30:36 +0000 (18:30 +0800)]
doc: Flesh out 0.9.0 relnotes
Craig Ringer [Tue, 24 Mar 2015 10:09:00 +0000 (18:09 +0800)]
bdr: Flesh out 0.8.0 relnotes
Craig Ringer [Tue, 24 Mar 2015 09:49:07 +0000 (17:49 +0800)]
doc: Skeletal documentation on catalogs, conflicts, more on monitoring
Craig Ringer [Tue, 24 Mar 2015 07:02:34 +0000 (15:02 +0800)]
doc: Examples of node removal
Craig Ringer [Tue, 24 Mar 2015 06:47:31 +0000 (14:47 +0800)]
doc: Remove superceded distributed sequences docs
Craig Ringer [Tue, 24 Mar 2015 06:39:30 +0000 (14:39 +0800)]
Make bdr.queue_truncate() SECURITY DEFINER
Fixes #5.
This shows our testing doesn't cover DDL as non-superuser properly.
Abhijit Menon-Sen [Tue, 24 Mar 2015 05:23:21 +0000 (10:53 +0530)]
Fix typo: wheras → whereas
Craig Ringer [Tue, 24 Mar 2015 04:33:36 +0000 (12:33 +0800)]
doc: Convert and update global sequences documentation
Craig Ringer [Tue, 24 Mar 2015 02:37:01 +0000 (10:37 +0800)]
doc: Rearrange top level to have consistent part/chapter levels
Craig Ringer [Tue, 24 Mar 2015 02:33:53 +0000 (10:33 +0800)]
doc: add a stub "catalogs and views" section
Craig Ringer [Tue, 24 Mar 2015 02:29:30 +0000 (10:29 +0800)]
doc: First pass copy-editing and revision of monitoring docs
kforte90 [Mon, 23 Mar 2015 22:27:03 +0000 (15:27 -0700)]
"added information that seems helpful to monitoring your bdr system"
kforte90 [Mon, 23 Mar 2015 21:11:38 +0000 (14:11 -0700)]
"added examples of how to call node functions"
kforte90 [Mon, 23 Mar 2015 19:42:51 +0000 (12:42 -0700)]
" added bdr.bdr_nodes table for monitoring"
Craig Ringer [Mon, 23 Mar 2015 14:43:03 +0000 (22:43 +0800)]
bdr: doc: Add release notes for 0.7.x, 0.8.x, and (tentative) 0.9.x
Craig Ringer [Mon, 23 Mar 2015 13:09:37 +0000 (21:09 +0800)]
bdr: doc: Further node management docs work
Better structure for node management docs, some coverage of node removal
Craig Ringer [Mon, 23 Mar 2015 13:04:12 +0000 (21:04 +0800)]
bdr: add bdr.bdr_get_local_node_name() function
Add a helper function to look up the local node's name from
bdr.bdr_nodes using the local node identity details.
Christoph Moench-Tegeder [Mon, 23 Mar 2015 11:16:37 +0000 (12:16 +0100)]
bdr: make log_tuple() compile