Daniel Farina
da82829659
Avoid checking config file when performing standby clone action
...
Signed-off-by: Daniel Farina <daniel@heroku.com >
2011-02-10 02:10:58 -08:00
Dan Farina
ebbb7c3a47
Fix more quoting: used an overly-quoted copy
...
Signed-off-by: Dan Farina <daniel@heroku.com >
2011-02-07 03:59:18 -08:00
Dan Farina
630d716819
Use the proper freeing method
...
As according to
http://www.postgresql.org/docs/9.0/static/libpq-exec.html#LIBPQ-EXEC-ESCAPE-STRING
Signed-off-by: Dan Farina <daniel@heroku.com >
2011-02-07 03:59:18 -08:00
Dan Farina
c1b84fe9fc
Make some fast hacks to quote schema names
...
Signed-off-by: Dan Farina <drfarina@acm.org >
2011-02-07 03:59:18 -08:00
Dan Farina
bfd3aed33b
Fix a myriad of problems introduced by merging
...
Signed-off-by: Dan Farina <drfarina@acm.org >
2011-02-06 18:32:49 -08:00
Dan Farina
c356f9e97f
Merge remote-tracking branch 'tbrs/master' into heroku
...
Grab the configuration struct changes. It was expeditious to un-do
some of my by-hand line-wrapping that avoids 80 character limit,
though.
Conflicts:
config.c
config.h
repmgr.c
repmgr.h
repmgrd.c
2011-02-04 18:35:56 -08:00
Dan Farina
a52ebc412c
Merge remote-tracking branch 'origin/master' into heroku
...
The Great Whitespace Reconciliation
Conflicts:
check_dir.c
config.c
dbutils.c
repmgr.c
repmgr.h
repmgrd.c
Signed-off-by: Dan Farina <drfarina@acm.org >
2011-02-04 18:27:28 -08:00
trbs
57169f51cf
made rsync_options configurable
2011-01-08 01:46:39 +01:00
trbs
fc2405f17d
use struct for config file information
2011-01-07 01:36:46 +01:00
trbs
cb1192b912
fix strformat order reversal in error message
2010-12-23 15:02:02 +01:00
Dan Farina
156714f3f1
Fix quoting misbehavior
...
Move the single quote in the formatting string to the end of the line.
Signed-off-by: Dan Farina <drfarina@acm.org >
2010-12-23 01:04:05 -08:00
Dan Farina
dd5ac660bf
Comments and cleanup
...
Signed-off-by: Dan Farina <drfarina@acm.org >
2010-12-23 00:55:39 -08:00
Dan Farina
3a430397dc
Avoid overwriting the new password-in-recovery.conf formatting
...
Previously: did that work, then threw it away/overwrote it.
Signed-off-by: Dan Farina <drfarina@acm.org >
2010-12-23 00:55:39 -08:00
Dan Farina
1f098c60ac
Remove dead-ish code
...
This is a leftover after having fixed various problems in the verbose
output here.
Signed-off-by: Dan Farina <drfarina@acm.org >
2010-12-23 00:55:39 -08:00
Dan Farina
0bae682a0d
Fix off-by-one in passing buffer size to xsnprintf
...
This would cause a consistent panic (exit() call) without reason.
This should not enable any overruns (but check my work).
Signed-off-by: Dan Farina <drfarina@acm.org >
2010-12-23 00:55:32 -08:00
Dan Farina
f969dca821
Hack to get passwords in recovery.conf
...
Signed-off-by: Dan Farina <drfarina@acm.org >
2010-12-22 20:10:32 -08:00
Dan Farina
29c39c21f6
Avoid a use-after-free in verbose logging (again)
...
Previously, this print would use memory freed by PQClear previously.
Also allocate/free memory to prevent this tiny memory leak.
Signed-off-by: Dan Farina <drfarina@acm.org >
2010-12-22 20:10:15 -08:00
Dan Farina
620974ba04
Fix unsafe string handling
...
It looks like the old code would overflow in some cases.
Signed-off-by: Dan Farina <drfarina@acm.org >
2010-12-22 19:00:21 -08:00
Greg Smith
9843205a4f
Reformat all source code using astyle
...
astyle --style=ansi --indent=tab
2010-12-22 19:07:01 -05:00
Greg Smith
1ddb78ddd8
Merge branch 'license' of https://github.com/gbartolini/repmgr into gbartolini-license
2010-12-22 19:02:50 -05:00
Dan Farina
faddaed316
Avoid a use-after-free in verbose logging
...
Previously, this print would use memory freed by PQClear previously.
Signed-off-by: Dan Farina <drfarina@acm.org >
2010-12-22 15:03:05 -08:00
Dan Farina
7b0a142075
Merge remote branch 'gbartolini/gabriele-2010-12' into heroku
...
Conflicts:
config.c
dbutils.c
dbutils.h
repmgrd.c
Signed-off-by: Dan Farina <drfarina@acm.org >
2010-12-21 16:08:07 -08:00
Dan Farina
84b69b3bd4
Whitespace to adjust for longer snprintf identifier
...
This is done in a separate patch to try and reduce the sound and fury
of the patch that actually did the conversion from sprintf to
snprintf-alikes.
Signed-off-by: Dan Farina <drfarina@acm.org >
Signed-off-by: Peter van Hardenberg <pvh@heroku.com >
2010-12-21 15:19:28 -08:00
Dan Farina
916c0492fb
sprintf to snprintf conversion
...
Move out string operations to another file, and introduce a frontend
to snprintf for various situations. This change is important for
catching and eliminating sprintf overflows, which are as of now many
times silently corrupting memory.
Signed-off-by: Dan Farina <drfarina@acm.org >
Signed-off-by: Peter van Hardenberg <pvh@heroku.com >
2010-12-21 15:19:28 -08:00
Dan Farina
af2edf10a0
Attack of whitespace pedantry
...
pgsql conventions (tabs, four-spaces-wide, etc) applied all around.
Also tried to fix some very tiny capitalization errors, auto-fill
problems, and some inter-block vertical whitespacing issues.
Long strings in repmgr.c were left intact, though. They are rather
numerous and are less of a problem than tiny bits of function calls
and comments wrapping over a line; the latter kind of problem has been
mostly fixed.
Signed-off-by: Dan Farina <drfarina@acm.org >
Signed-off-by: Peter van Hardenberg <pvh@heroku.com >
2010-12-21 15:19:28 -08:00
Gabriele Bartolini
d88783a4d9
Changed pg_version() prototype in order to remove the small memory leak
2010-12-16 22:48:19 +01:00
Gabriele Bartolini
9b7a078e0e
Added license information in every source file
2010-12-16 21:50:15 +01:00
Greg Smith
56c65acd99
Add force option to help and docs on help
2010-12-06 00:33:02 -05:00
Jaime Casanova
6c484409bb
Added a --wal-keep-segments parameter which is the minimum value that the
...
PostgreSQL's GUC wal_keep_segments should have for "standby clone" to work
reliable. We default to 5000 but it could be changed for the user if it
knows what he is doing.
2010-11-19 08:36:15 -05:00
Greg Smith
3172ad97cf
Add rsync remote user name option. Expand docs to
...
include command line help examples.
2010-11-08 15:29:53 -05:00
Jaime Casanova
58498389dc
Temporarily remove the check if promote actually works, because
...
when the database is taking too long to recover and start it thinks
the database is not online.
2010-10-28 18:47:56 -05:00
Jaime Casanova
a70a6d2912
Really fix STANDBY FOLLOW this time.
2010-10-24 19:20:25 -05:00
Jaime Casanova
f45aa077b1
Add message to remember we should REINDEX any hash indexes, i decided
...
not to do it myself because it could take a long time if the index is
big.
2010-10-23 20:44:05 -05:00
Jaime Casanova
91276394e8
Check STANDBY PROMOTE actually promote the node
2010-10-23 19:54:45 -05:00
Jaime Casanova
028aeb452d
Fix STANDBY FOLLOW to correctly recreate recovery.conf
2010-10-23 19:44:07 -05:00
Jaime Casanova
ed3038ef6b
Fix a few typos in my last commit
2010-10-23 15:51:08 -05:00
Jaime Casanova
219b4431b4
Replace the function is_supported_version() with the function pg_version()
...
the main difference is that this new function doesn't return a bool but
a string representing the current major version of the postgresql server
or an empty string if it's earlier than 9.0.
Teach all commands that could connect to master and standby that they
both should have the same major version.
2010-10-23 15:27:24 -05:00
Jaime Casanova
8409727f0c
A few more comments
2010-10-22 18:41:31 -05:00
Jaime Casanova
45fdbbb426
Another segmentation fault caused by an uninitialized string
2010-10-22 18:38:35 -05:00
Jaime Casanova
e3d1580d76
In repmgrd there was the same problem that cause the segfault
2010-10-21 17:58:38 -05:00
Jaime Casanova
7ddc80ea92
I was having a segmentation fault because i was using a pointer to char
...
but never indicated a size for the compiler to reserve... malloc()
solves my problem
2010-10-21 14:49:26 -05:00
Jaime Casanova
ff511f19b4
Remove the second --port parameter, we now use repmgr.conf for get
...
conninfo of the standby
2010-10-21 01:04:23 -05:00
Jaime Casanova
14ae59c6b2
Add a function that checks the parameters based on the action.
...
Refactor some actions and use repl_nodes to get connections to
master.
2010-10-21 00:48:04 -05:00
Jaime Casanova
cd287f923a
I failed to add the schema "repmgr_%s" in all places where the repl_*
...
tables are used.
2010-10-18 09:34:14 -05:00
Jaime Casanova
378bdd7bd4
Add MASTER REGISTER and STANDBY REGISTER commands.
2010-10-18 02:21:55 -05:00
Jaime Casanova
6779a96b0b
The short option for --force is now -F, because i will use the other
...
to indicate a config file. also for simetry with repmgrd
2010-10-13 14:54:09 -05:00
Jaime Casanova
7dce793882
Add a --force option to upgrade an existing PostgreSQL instalation.
...
This could be useful, for readding recently failed nodes.
2010-10-08 02:55:52 -05:00
Jaime Casanova
d42f3e25f4
Fix a typo in the last commit that show the right directory in error messages
2010-10-07 00:29:10 -05:00
Jaime Casanova
f5e95dd46a
Exclude pid files from standby clone
2010-10-06 01:04:30 -05:00
Jaime Casanova
172caf8968
When the directory for a tablespace cannot be created or exists
...
but is not empty, it was sending the message using dest_dir instead.
2010-10-05 21:20:20 -05:00