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
775 Commits 43 Branches 68 Tags
293d66bf710fb2a5b340b1c723b441a26a7bd0bb
Commit Graph

775 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Ian Barwick
293d66bf71 Fix minimum accepted value for "degraded_monitoring_timeout" 
Should be -1, the default.

Addresses GitHub #411.
 2018-03-30 09:12:17 +09:00
Ian Barwick
3e1f0ec168 repmgr: move demoted primary check to the final step during switchover 
This will give the demoted primary more time to start up as a standby,
during which "standby follow" can be executed on sibling nodes, if
specified.
 2018-03-27 16:41:13 +09:00
Ian Barwick
6f9a1f975e repmgr: poll demoted primary after restart during switchover 
During a switchover operation, once the demoted primary has been restarted
as a standby, repmgr attempts to reconnect to verify its status and drop
any redundant replication slots. However it's possible the standby may still
be in the startup phase, so poll for "standby_reconnect_timeout" seconds
before giving up.

Addresses GitHub #408.
 2018-03-27 15:58:18 +09:00
Ian Barwick
deea4f69f7 Fix "repmgr cluster crosscheck" output 
Addresses GitHub #398.
 2018-03-27 10:28:27 +09:00
Ian Barwick
37e53108a2 Consolidate connection closure calls 2018-03-27 08:52:23 +09:00
Ian Barwick
96cf06204c doc: add note about remote command execution 
When executing a command on a remote server, repmgr expects the remote binary
to be in the same location as the local binary. It's reasonable to assume
repmgr will be deployed in a unified environment; if not, the onus is on the
user to ensure repmgr can find the remote binary, e.g. by creating appropriate
symlinks.

Addresses query in GitHub #406.
 2018-03-27 08:47:56 +09:00
Ian Barwick
381e22c2c7 Misc tweaks to witness code 2018-03-26 20:59:38 +09:00
Ian Barwick
7e2af17783 repmgrd: tweak log notices when marking a standby as failed 
Announce what we're going to do (set the node record inactive) *before*
performing the action. Makes reading the log slightly easier.
 2018-03-23 13:27:37 +08:00
Ian Barwick
b4272853e7 Add event "repmgrd_failover_aborted" 2018-03-23 10:44:00 +08:00
Ian Barwick
562b6ddfc2 Add error code ERR_FOLLOW_FAIL 2018-03-23 10:34:19 +08:00
Ian Barwick
a15e5c9d52 Tidy up queries in dbutils.c 
- standardize formatting
- prefix various internal function calls with "pg_catalog.", to
  mitigate possible risks from CVE-2018-1058
 2018-03-23 10:33:28 +08:00
Ian Barwick
d9cc09cee4 repmgrd: fix typo 2018-03-21 12:36:51 +09:00
Ian Barwick
c4f6abe951 Update HISTORY 2018-03-21 06:51:56 +09:00
Martín Marqués
e454fb77d3 While reviewing 7cb6e5af8d before merging 
I noticed that besides the result cleanup added, there was still a missing
spot inside the if condition.

Adding the PQclear that was missing.
 2018-03-21 06:51:50 +09:00
Andrzej Nowicki
b76e5852d3 One more memory leak fixed 2018-03-21 06:51:43 +09:00
Andrzej Nowicki
0674364ffd Clear node list to avoid memory leak, fixes #402 2018-03-21 06:51:37 +09:00
Ian Barwick
b2eb9b8525 Correctly handle error message pointer when parsing strings. 
When parsing conninfo strings, ensure the error message pointer is
actually returned to the caller.

Not a criticial issue, just meant the contents of the error message
were not being displayed.
 2018-03-10 14:28:10 +09:00
Ian Barwick
71c5d10a8c doc: update 4.0.4 release date 2018-03-09 20:07:16 +09:00
Ian Barwick
1476b21cd4 doc: update release notes 
Add note about requiring 4.0.3 or later on all nodes when performing
a switchover from a noder running 4.0.3 or later.

Per report in GitHub #388.
 2018-03-09 09:46:58 +09:00
Ian Barwick
b17993abdb doc: update "repmgr primary unregister" description 
As noted by GitHub user yonj1e in GitHub #396.
v4.0.4 		2018-03-08 15:01:25 +09:00
Ian Barwick
8f68344f9a doc: update FAQ 
Additional clarification for "repmgr standby clone --recovery-conf-only"
 2018-03-08 10:04:30 +09:00
Ian Barwick
125ac6c297 doc: update FAQ 
Add entry about upgrading PostgreSQL
 2018-03-08 10:04:30 +09:00
Ian Barwick
955860923f Fix parsing of -k/--keep-history option 
GitHub #394.
 2018-03-07 19:14:18 +09:00
Ian Barwick
50626f90cc Add 4.0.4 release notes 2018-03-07 14:17:04 +09:00
Ian Barwick
9aea5b8aa7 repmgrd: fix failover handling in "manual" mode 
Regression was introduced in commit c7a585c555
 2018-03-06 22:35:51 +09:00
Ian Barwick
ed1bcb159e repmgrd: remove duplicate local record check in BDR mode 2018-03-06 12:31:07 +09:00
Ian Barwick
9c72c0d66e Add event "repmgrd_shutdown" 
Implements GitHub #393
 2018-03-06 10:59:54 +09:00
Emre Hasegeli
0ddc226c2a Add witness options to the main help 
GitHub #392
 2018-03-06 10:57:33 +09:00
Ian Barwick
93830cad61 Fix directory creation when cloning from Barman 2018-03-05 19:31:53 +09:00
Ian Barwick
bca1660d5e Improve repmgrd logging in BDR mode 
Also ensure interval status log line is shown as intended
 2018-03-05 15:05:40 +09:00
Ian Barwick
5a52917421 repmgrd: add debug log output for "monitor_interval_secs" sleep in all modes 2018-03-05 14:23:58 +09:00
Emre Hasegeli
70752d7d4a Add missing options to the main help 2018-03-05 09:52:04 +09:00
Ian Barwick
c29d1efc37 "standby clone": improve replication user selection 
Use the upstream node's replication user when checking the replication
connection.
 2018-03-02 16:21:32 +09:00
Ian Barwick
6fbbe2a97a "standby clone": fix --superuser handling 
get_superuser_connection() was erroneously using the local node record
to connect to as a superuser, which works when registering the primary
but obviously not when cloning a standby.

Addresses GitHub #380.
 2018-03-02 14:49:17 +09:00
Ian Barwick
ce42d6827e Update HISTORY 2018-03-01 15:51:09 +09:00
Ian Barwick
98384559a6 "standby clone": remove restriction on replication slots in Barman mode 
While it's preferable to avoid standby replication slots if Barman is in
use, there's no technical reason to prevent this.

Implements GitHub #379.
 2018-03-01 15:47:28 +09:00
Ian Barwick
4a1477343b repmgr: escape "restore_command" in generated recovery.conf 2018-03-01 10:39:04 +09:00
Ian Barwick
d2b9d20393 "standy clone": fix primary_conninfo when --upstream-conninfo provided 2018-03-01 09:18:40 +09:00
Ian Barwick
fe594c95ad repmgrd: retry standby connection after cascading standby failover 2018-02-28 21:15:11 +09:00
Ian Barwick
60e63feaca repmgrd: add configuration file parameter "standby_reconnect_timeout" 
This is used for determining a timeout when reconnecting to the standby
after executing the "follow_command". This will normally not need to be
set explicitly, but maybe useful in cases where the standby's startup
phase can last longer than usual.
 2018-02-28 18:56:33 +09:00
Ian Barwick
ae4d0f2622 repmgrd: fix main monitoring loop for witness server 
Missing "break" was breaking it when following a new primary.
 2018-02-28 16:30:14 +09:00
Ian Barwick
5e8b41e221 repmgrd: retry standby connection after "follow_command" executed 
It's possible that the standby is still starting up after the "follow_command"
completes, so poll for a while until we get a connection.
 2018-02-28 15:35:47 +09:00
Ian Barwick
c7a585c555 repmgrd: improve log output 
- emit explicit startup NOTICE
- emit NOTICE when falling back to degraded monitoring on a primary node
- improve log message and event notification details when monitoring
  a former primary which has been reconnected as a standby
 2018-02-28 12:35:13 +09:00
Ian Barwick
a27dd8c49c doc: document "primary_follow_timeout" configuration file parameter. 2018-02-27 10:09:40 +09:00
Ian Barwick
9365bf3474 "standby promote": make timeout values configurable 
This introduces following new configuration file parameters, which
were previously hard-coded values:

 - promote_check_timeout
 - promote_check_interval

Implements GitHub #387.
 2018-02-27 10:04:58 +09:00
Ian Barwick
e8ae0831fe doc: add <options> section for various commands 2018-02-26 16:54:54 +09:00
Ian Barwick
518866eba5 "node status": improve replication slot warnings 
Addresses GitHub #385
 2018-02-23 11:06:47 +09:00
Ian Barwick
ed0330c334 "standby clone": document --recovery-conf-only option 2018-02-23 10:54:42 +09:00
Ian Barwick
1f021dc9fa "standby clone --recovery-conf-only": display generated file with --dry-run 
Refactor the original code which generates "recovery.conf" to place the
output into a buffer, which can either be output as "recovery.conf"
or copied to a buffer specified by the caller.
 2018-02-23 10:16:47 +09:00
Ian Barwick
425839d764 Fix typo in function name 2018-02-22 15:48:41 +09:00