github/repmgr
1
0
Fork 0
mirror of https://github.com/EnterpriseDB/repmgr.git synced 2026-03-23 15:16:29 +00:00
Code Issues Packages Projects Releases 1 Wiki Activity
357 Commits 43 Branches 68 Tags
b4e83cf188d89eff91db79f6f5cd508e524fa8dc
Commit Graph

357 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Jaime Casanova
41dbc39527 Add release notes 2012-07-05 09:35:23 -05:00
Jaime Casanova
50b7147f15 Change Copyright date to cover 2012 2012-07-04 10:47:26 -05:00
Jaime Casanova
f5e57aa433 Add an option for "no-history" mode, where repmgrd just checks the 
conectivity of master but don't INSERT any data into it
 2012-07-04 10:07:31 -05:00
Jaime Casanova
ac5a9d1fd6 The release changed, just wait a little before setting it. 
Also make well known names in HISTORY be only names, without
last name
 2012-07-02 00:06:57 -05:00
Jaime Casanova
cb740b68be Add a check of the connection inside the CancelQuery() so it check 
that before trying to cancel a query, which can block.
 2012-06-26 11:29:02 -05:00
Jaime Casanova
d58ea77798 Add a quick setup for autofailover 2012-06-26 07:49:43 -05:00
Jaime Casanova
e3c3c22b6e Improve the version message to actually show the repmgr version not 
only postgresql's one
 2012-06-25 22:54:48 -05:00
Jaime Casanova
861a3c8f22 Fix CLUSTER CLEANUP, it needs to establish a local connection in order 
to look for the master
 2012-06-16 01:32:59 -05:00
Jaime Casanova
e51870b504 Force to enter a password for the superuser in the witness, this is 
in case we need to send a password to connect as stated in
master's pg_hba.conf.
 2012-06-15 13:51:45 -05:00
Jaime Casanova
5651720560 Remove a variable left in last commit 2012-06-15 09:46:01 -05:00
Jaime Casanova
d32a6cdb24 Remove kludge added to create user and db for witness. 
It's too fragile, almost always cause a "segment violation" and
don't seems to be very useful.
 2012-06-15 09:41:54 -05:00
Jaime Casanova
9e10987b90 Fix a few bugs introduced when merging features 2012-06-15 09:40:09 -05:00
Jaime Casanova
64fce88e99 Add a CLUSTER CLEANUP command to clean monitor's history, 
also include a --keep-history (-k) option to indicate how many
days of history to keep
 2012-06-13 00:39:54 -05:00
Jaime Casanova
7a76f1998c getMasterConnection() cannot avoid checking the same node that asks 
to find the master.
This was a micro optimization based on the fact that all commands that
needed to detect the master were executed from the standby but now that
we have CLUSTER level commands that is not true anymore
 2012-06-12 23:28:24 -05:00
Jaime Casanova
4db046a8ea Allow repmgr to obtain tablespace's locations from pg 9.2 and later 
in which we no longer have a spclocation column in pg_tablespaces
 2012-06-12 11:08:15 -05:00
Jaime Casanova
331eca447a STANDBY CLONE should be run by a SUPERUSER, otherwise we won't be able 
to retrieve data_directory and the other parameters we need by
querying the database.
 2012-06-12 09:42:50 -05:00
Jaime Casanova
b5b2f93f7e Merge branches 'master' and 'async' 2011-12-02 00:28:17 -05:00
Jaime Casanova
9d03d4a254 After checking that master is alive, is_pgup() should return not keep 
checking forever.
 2011-12-01 23:58:12 -05:00
Jaime Casanova
3b2ccc5b78 Add a master_response_timeout parameter and use it to limit the amount 
of time we spent a reponse from master before declaring the failure.
Also, change is_pgup() so it use PQsendQuery() instead of PQexec to
execute the check of master
 2011-12-01 01:20:33 -05:00
Jaime Casanova
89a1e2bcbd Not even consider old master as an option in failover 2011-11-27 19:17:59 -05:00
Jaime Casanova
7077a7c68f Add -w option to pg_ctl commands so we wait until command is finish. 
Or at least, we try. By default, after 60 seconds pg_ctl just return.
This make useless to wait ourselves after pg_ctl start of witness so
remove the sleep
 2011-11-27 18:38:53 -05:00
Jaime Casanova
9b8fb7e960 Remove last argument from log_err, left in commit 55c7ea4b5e. 
Also rephrase the sentence

Reported by Jeroen Dekkers
 2011-11-25 14:59:29 -05:00
Jaime Casanova
55c7ea4b5e Fix a wrong message. 
It was saying the problem is the version of the PostgreSQL server while
it actually is because the MASTER REGISTER command was running on a
standby node
 2011-11-10 09:38:12 -05:00
Jaime Casanova
0240cb2bfb Fix a typo introduced in commit cb764b180d 2011-11-03 12:50:51 -05:00
Jaime Casanova
2e64028f26 Add the improvement of the view repl_status to the history 2011-10-20 23:26:41 -05:00
Greg Smith
27396f462a Add strnlen on platforms that don't have it, such as OS X 2011-10-20 23:23:31 -05:00
Jaime Casanova
cb764b180d Let the clone happen in a session with synchronous_commit off. This 
is because in pg 9.1 the default configuration can easily allow sync
rep to be activated even if no standby is present and will block
pg_start_backup() and pg_stop_backup() in that case.
 2011-10-03 14:04:27 -05:00
Jaime Casanova
65cf045a63 Fix a few problems introduced in 020e17b059 
while adding standby_name
 2011-08-09 02:01:00 -05:00
Jaime Casanova
3d53661d97 Fix a typo introduced in the commit about standby_name 
(020e17b059)
 2011-08-09 01:14:50 -05:00
Jaime Casanova
7b3a9cbbe4 Rename standby_name configuration option as node_name 
per gripe by Cedric
 2011-08-09 01:01:37 -05:00
Jaime Casanova
276c947202 Add name to nodes and show it for the standby in repl_status. 2011-08-09 00:58:01 -05:00
Jaime Casanova
020e17b059 Add a new standby_name parameter that, when provided in repmgr.conf, 
is included in the primary_conninfo as application_name... This is
a simple first step towards making repmgr works fine with 9.1 and
sync standbys
 2011-08-09 00:57:23 -05:00
Jaime Casanova
10817ecffc Add "witness" as a node role in CLUSTER SHOW 2011-08-08 13:29:26 -05:00
Jaime2ndQuadrant
f26e37c087 Merge pull request #2 from 2ndquadrant-it/master 
Add function to show nodes with actual roles: repmgr cluster show
 2011-08-08 11:10:13 -07:00
Carlo Ascani
61c31eeb0c Added function to show nodes with actual roles: repmgr cluster show 2011-08-08 19:32:32 +02:00
Jaime2ndQuadrant
0b5d2418df Merge pull request #1 from 2ndquadrant-it/master 
Added function "write_primary_conninfo" which now adds the username to the primary_conninfo parameter in recovery.conf

Author: Gabriele and Marco
 2011-08-05 11:03:54 -07:00
Gabriele Bartolini
f44037088b Added function "write_primary_conninfo" which now adds the username to the primary_conninfo parameter in recovery.conf 
Signed-off-by: Gabriele Bartolini <gabriele.bartolini@2ndquadrant.it>
Signed-off-by: Marco Nenciarini <marco.nenciarini@2ndquadrant.it>
 2011-08-05 17:41:18 +02:00
Carlo Ascani
e05dd729f8 Fix a typo in documentation 2011-08-05 00:15:13 -05:00
Jaime Casanova
ab8d2673dc Remove useless second parameter from create_recovery_file() 
per gripe by Gabriele
 2011-08-04 17:45:28 -05:00
Jaime Casanova
28c5ac3d91 Make is_standby() and is_witness() work the same way, actually 
is_standby() can show the segmentation fault problem but for code
clearity
 2011-07-26 18:36:27 -05:00
Carlo Ascani
c983fdf83c Fix a possible double free that would cause a segfault in 
checkNodeConfiguration()
 2011-07-26 18:22:06 -05:00
Carlo Ascani
158214acae Only log that we have restored the connection when that already happens, 
this fix a flood in logs that make them unnecessarily big
 2011-07-26 18:16:22 -05:00
Jaime Casanova
04290c1f60 Another pass reviewing code: 
- remove a duplicate check for ssh connection and just exit if
  we can't connect to the remote host.
  stop_backup is only needed if pg_start_backup() has been
  already called
- remove a new connection to master in stop_backup label, AFAIC
  we hadn't close the one we already opened
- add a lot of PQfinish(), a few PQclear() and adjust code to what
  we used to do in 1.x
 2011-07-26 16:14:44 -05:00
Jaime Casanova
cedc5e20fb Add a check to ensure 'hot_standby=on' on master. While this parameter 
doesn't have an effect here it will be a reminder for users, besides
will reduce things user have to do on the standby
 2011-07-26 16:12:51 -05:00
Jaime Casanova
b61ac86b5f Move repmgr_update_standby_location() and 
repmgr_get_last_standby_location() functions to repmgr schema.
This will be important when we want to create this as an
extension, so better make it be in the right place from day 1.
 2011-07-26 16:12:11 -05:00
Jaime Casanova
64da72b48b repl_monitor.last_wal_standby_location should accept nulls because 
of the witness
 2011-07-26 16:11:26 -05:00
Jaime Casanova
c4778a83c5 Be consistent in the use of sqlquery_snprintf() for query strings, 
which was not used in the new code.
Also add a few "newline" characters in some messages, and other
minutia.
 2011-07-19 05:02:51 -05:00
Jaime Casanova
4d26e4d21e Redefining repl_status view to improve performance, also add an 
index on repl_monitor to speed up even more the view.
 2011-06-13 11:58:43 -05:00
Jaime Casanova
3f4ff5a862 A "HEAD" word remaining from a merge conflict was accidentally left 
in... removing it
 2011-06-10 21:16:07 -05:00
Simon Riggs
1d05467df1 README additions and improvements 2011-06-07 11:37:09 +01:00