Greg Sabino Mullane [Thu, 30 Oct 2014 21:02:50 +0000 (17:02 -0400)]
Merge dbname2, host2, etc. into the new array method.
Greg Sabino Mullane [Thu, 30 Oct 2014 17:40:36 +0000 (13:40 -0400)]
Quote values in error message when replicate_row fails
Christoph Berg [Mon, 8 Sep 2014 15:05:35 +0000 (17:05 +0200)]
Fix bloat check to use correct SQL depending on the server version.
Previously it used the psql version which might differ when multiple
versions are installed locally, or when querying remote servers.
The hot_standby_delay check is also affected, but more involved to fix,
so leave that for later.
Adrian Vondendriesch
Greg Sabino Mullane [Mon, 14 Jul 2014 16:52:47 +0000 (12:52 -0400)]
Show OS on failure-to-parse errors
Greg Sabino Mullane [Mon, 14 Jul 2014 16:50:34 +0000 (12:50 -0400)]
When we fail to parse the psql output, show the entire output string rather than just the failing line, to better aid in debugging as recently seen on the list.
Greg Sabino Mullane [Mon, 14 Jul 2014 16:21:26 +0000 (12:21 -0400)]
Tweak testing output
Andrew Yochum [Thu, 3 Jul 2014 14:15:27 +0000 (10:15 -0400)]
Improved multi-slave support in replicate_row.
Greg Sabino Mullane [Thu, 3 Jul 2014 11:24:35 +0000 (07:24 -0400)]
Merge branch 'space_error' of https://github.com/jeffjanes/check_postgres
Greg Sabino Mullane [Thu, 3 Jul 2014 11:22:39 +0000 (07:22 -0400)]
Merge branch 'master' of https://github.com/terrorobe/check_postgres
Conflicts:
check_postgres.pl
Greg Sabino Mullane [Thu, 3 Jul 2014 10:53:16 +0000 (06:53 -0400)]
Merge branch 'query-time-show-query' of https://github.com/petere/check_postgres
Christoph Berg [Thu, 3 Jul 2014 08:31:36 +0000 (10:31 +0200)]
Declare POD encoding to be utf8
Greg Sabino Mullane [Thu, 12 Jun 2014 21:49:51 +0000 (17:49 -0400)]
Quick stub to show port when testing. Always 5432 for now (with a custom socket dir)
Josh Williams [Thu, 12 Jun 2014 21:24:24 +0000 (17:24 -0400)]
Add xact timestamp support to hot_standby_delay
Allow the hot_standby_delay check to accept xlog byte position or
timestamp lag intervals as thresholds, or even both at the same time.
Peter Eisentraut [Wed, 11 Jun 2014 20:04:38 +0000 (16:04 -0400)]
Show actual long-running query in query_time output
Michael Renner [Wed, 11 Jun 2014 15:10:33 +0000 (17:10 +0200)]
fix & extend hot_standby_delay documentation
This should also account for all places where the refactoring of
the numbered dbparameters (host2, port2, etc.) was referenced in the
documentation. Drop the TODO entry.
Michael Renner [Wed, 11 Jun 2014 15:08:40 +0000 (17:08 +0200)]
Merge branch 'master' of github.com:terrorobe/check_postgres
Jeff Janes [Wed, 21 May 2014 17:48:01 +0000 (10:48 -0700)]
Don't swallow space before the -c flag when reporting errors
Greg Sabino Mullane [Mon, 5 May 2014 15:50:39 +0000 (11:50 -0400)]
Quick TODO item
Greg Sabino Mullane [Tue, 18 Feb 2014 01:11:20 +0000 (20:11 -0500)]
Start 2.21.1 notes
Jeff Frost [Tue, 18 Feb 2014 01:08:24 +0000 (20:08 -0500)]
Add an ORDER BY to the slony_status check to account for cases in which
there is more than one node in replication; we want to report on the
most lagged server.
Greg Sabino Mullane [Wed, 9 Oct 2013 18:14:05 +0000 (14:14 -0400)]
'no warnings' makes Perl::Critic cry. Try a different way.
David E. Wheeler [Tue, 24 Sep 2013 23:27:22 +0000 (16:27 -0700)]
--host2 and friends no longer work.
David E. Wheeler [Tue, 24 Sep 2013 23:22:44 +0000 (16:22 -0700)]
Eliminate some "Use of uninitialized value" warnings.
David E. Wheeler [Tue, 24 Sep 2013 23:20:46 +0000 (16:20 -0700)]
Get rid of number opts in docs.
David E. Wheeler [Tue, 24 Sep 2013 22:30:21 +0000 (15:30 -0700)]
`curl` does not like `=` between an option and its value.
This gets the `t/02_new_version_cp.t` to pass on systems that use `curl`.
Greg Sabino Mullane [Tue, 24 Sep 2013 19:12:35 +0000 (15:12 -0400)]
Signatures for 2.21.0
Greg Sabino Mullane [Tue, 24 Sep 2013 19:10:18 +0000 (15:10 -0400)]
Whitespace
Greg Sabino Mullane [Tue, 24 Sep 2013 19:08:30 +0000 (15:08 -0400)]
Remove the no warnings hack.
Greg Sabino Mullane [Tue, 24 Sep 2013 19:00:04 +0000 (15:00 -0400)]
Version bump, better binmode
Greg Sabino Mullane [Tue, 24 Sep 2013 18:58:28 +0000 (14:58 -0400)]
Spelling tweaks
Greg Sabino Mullane [Tue, 24 Sep 2013 18:57:59 +0000 (14:57 -0400)]
Version bump
Greg Sabino Mullane [Tue, 24 Sep 2013 18:34:32 +0000 (14:34 -0400)]
Skip a few more files in the manifest
Greg Sabino Mullane [Mon, 23 Sep 2013 20:22:23 +0000 (16:22 -0400)]
Spelling fixes
Greg Sabino Mullane [Mon, 23 Sep 2013 19:46:09 +0000 (15:46 -0400)]
Typo
Greg Sabino Mullane [Mon, 23 Sep 2013 19:44:38 +0000 (15:44 -0400)]
Perl::Critic insipred fixes
Greg Sabino Mullane [Mon, 23 Sep 2013 19:12:43 +0000 (15:12 -0400)]
Minor translation fixes
Christoph Berg [Mon, 23 Sep 2013 13:12:18 +0000 (15:12 +0200)]
Document testsuite, notably the DBD::Pg version required
Greg Sabino Mullane [Sun, 22 Sep 2013 18:53:40 +0000 (14:53 -0400)]
Fix regex for cp_version testing.
Greg Sabino Mullane [Sun, 22 Sep 2013 18:47:24 +0000 (14:47 -0400)]
Bump version to 2.20.2
Greg Sabino Mullane [Sun, 22 Sep 2013 18:33:44 +0000 (14:33 -0400)]
Adding a jstkind to the dummy pgagent schema, so the recent changes that reference it can work.
Christoph Berg [Thu, 19 Sep 2013 10:58:12 +0000 (12:58 +0200)]
t/02_relation_size.t: Sleep 0.5s for the MRTG test
Sleeping 1s created a race condition because we also check after 1s, so
it was only luck whether the result of that check is "1" or "2"
Christoph Berg [Thu, 19 Sep 2013 10:11:06 +0000 (12:11 +0200)]
Fix same_schema check on 9.3 by ignoring relminmxid differences in pg_class
Christoph Berg [Thu, 19 Sep 2013 09:48:50 +0000 (11:48 +0200)]
t/02_slony_status.t: Fix to work with 9.3
The error was "schema does not exist" until 9.2, and is now "table does
not exist", so just look for "does not exist".
Christoph Berg [Thu, 19 Sep 2013 09:21:17 +0000 (11:21 +0200)]
Setting SKIP_NETWORK_TESTS will skip the new_version tests
Build daemons for Debian packages might not have network access, so
provide a way to disable the tests depeding on networking
Christoph Berg [Thu, 19 Sep 2013 09:10:44 +0000 (11:10 +0200)]
t/02_last_vacuum.t: Reorder commit calls to make test more stable
Christoph Berg [Thu, 19 Sep 2013 09:07:47 +0000 (11:07 +0200)]
Install man page in section 1
Peter Eisentraut, bug 53, github issue 26
Christoph Berg [Wed, 18 Sep 2013 15:44:40 +0000 (17:44 +0200)]
Add a changelog entry for the locks fix
Christoph Berg [Wed, 18 Sep 2013 15:37:11 +0000 (17:37 +0200)]
Skipping disk_space tests if df doesn't work
df might fail in chroot environments, e.g. on build daemons where
check-postgres packages are built
Christoph Berg [Wed, 18 Sep 2013 15:04:38 +0000 (17:04 +0200)]
Sort lock types so the output ordering is predicatable
Greg Sabino Mullane [Wed, 28 Aug 2013 15:27:05 +0000 (11:27 -0400)]
Credit recent change.
mintsoft [Wed, 28 Aug 2013 09:03:17 +0000 (10:03 +0100)]
Update check_postgres.pl
Fixing issue with SQL steps in check_pgagent_jobs for jobs which perform deletes, these return non-0 yet are still successful.
Behaviour should now be that <>0 on batch is a failure, but only -1 is a failure for sql steps.
Christoph Berg [Thu, 4 Jul 2013 11:17:57 +0000 (13:17 +0200)]
README: Fix two typos
Christoph Berg [Tue, 2 Jul 2013 13:27:58 +0000 (15:27 +0200)]
t/02_settings_checksum.t: Accept hex digits
Previously, this check only worked by chance if the returned checksum
started with a plain digit.
Greg Sabino Mullane [Tue, 25 Jun 2013 02:59:13 +0000 (22:59 -0400)]
Signatures for 2.20.1
Greg Sabino Mullane [Tue, 25 Jun 2013 02:56:34 +0000 (22:56 -0400)]
Final html doc tweak
Greg Sabino Mullane [Tue, 25 Jun 2013 02:55:50 +0000 (22:55 -0400)]
Let's get 2.20.1 out the door
Greg Sabino Mullane [Tue, 25 Jun 2013 02:50:20 +0000 (22:50 -0400)]
Cannot test for bad sequences in that manner: we already exclude temp schemas!
Greg Sabino Mullane [Tue, 25 Jun 2013 02:37:24 +0000 (22:37 -0400)]
No tabs please.
Greg Sabino Mullane [Tue, 25 Jun 2013 02:36:14 +0000 (22:36 -0400)]
Move issue tracking from Bugzilla to github
Greg Sabino Mullane [Tue, 25 Jun 2013 02:34:37 +0000 (22:34 -0400)]
Bump version to 2.20.1; update docs
Greg Sabino Mullane [Tue, 25 Jun 2013 02:32:26 +0000 (22:32 -0400)]
Update credits.
Greg Sabino Mullane [Tue, 25 Jun 2013 02:28:31 +0000 (22:28 -0400)]
Better "simple" output
Greg Sabino Mullane [Mon, 24 Jun 2013 12:57:50 +0000 (08:57 -0400)]
Don't compare the new 'relallvisible' ccolumn
Greg Sabino Mullane [Mon, 24 Jun 2013 12:42:24 +0000 (08:42 -0400)]
Update TODO file
Greg Sabino Mullane [Wed, 8 May 2013 21:32:20 +0000 (17:32 -0400)]
Credit recent change
Josh Williams [Wed, 8 May 2013 19:26:10 +0000 (15:26 -0400)]
Remove host reordering in hot_standby_delay check
This was resulting in the master queries being sent to the replica and
vice versa, if the replica is listed first. Will probably still want to
reorder the hosts in output (or just list them all) later.
Thanks to Jacobo Blasco for debug assistance.
Greg Sabino Mullane [Wed, 1 May 2013 02:08:10 +0000 (22:08 -0400)]
Credit the 9.2 fixes
Greg Sabino Mullane [Wed, 1 May 2013 01:55:35 +0000 (21:55 -0400)]
Typo
Greg Sabino Mullane [Wed, 1 May 2013 01:35:55 +0000 (21:35 -0400)]
Do things the right way and run different versions of the bloat query depending on the target version.
Greg Sabino Mullane [Wed, 1 May 2013 01:18:37 +0000 (21:18 -0400)]
Do not use pg_stats "inherited" column in bloat query, if the database is 8.4 or older.
Technically, we are not checking the target's version.
Greg Sabino Mullane [Wed, 1 May 2013 01:07:54 +0000 (21:07 -0400)]
Credit last change
Michael Renner [Wed, 3 Apr 2013 10:40:39 +0000 (12:40 +0200)]
bloat query: fix totalwastedbytes
totalwastedbytes returned only the number of wasted pages - fix that by
multiplying with the blocksize
Greg Sabino Mullane [Wed, 1 May 2013 01:00:08 +0000 (21:00 -0400)]
Credit last change
Michael Renner [Wed, 3 Apr 2013 10:40:39 +0000 (12:40 +0200)]
bloat query: fix totalwastedbytes
totalwastedbytes returned only the number of wasted pages - fix that by
multiplying with the blocksize
Nathaniel Waisbrot [Mon, 1 Apr 2013 15:33:18 +0000 (11:33 -0400)]
cause the --reverse flag to invert string comparisons in custom queries
Greg Sabino Mullane [Sat, 23 Mar 2013 16:14:06 +0000 (12:14 -0400)]
Credit last change
Dominic Hargreaves [Sat, 23 Mar 2013 16:09:34 +0000 (12:09 -0400)]
Make sure connection checks return CRITICAL rather than UNKNOWN.
Patch by Dominic Hargreaves with minor tweaks by Greg Sabino Mullane.
Thanks to Christoph Berg for the forward (this came in via the Debian bugs system)
Greg Sabino Mullane [Wed, 13 Mar 2013 18:48:00 +0000 (14:48 -0400)]
New signatures
Greg Sabino Mullane [Wed, 13 Mar 2013 18:41:43 +0000 (14:41 -0400)]
Signature for 2.20.0
Greg Sabino Mullane [Wed, 13 Mar 2013 18:41:27 +0000 (14:41 -0400)]
Manifest tweaks.
Greg Sabino Mullane [Wed, 13 Mar 2013 18:38:12 +0000 (14:38 -0400)]
Minor tweaks to get ready for 2.20.0
Greg Sabino Mullane [Wed, 13 Mar 2013 18:37:56 +0000 (14:37 -0400)]
Don't barf when schema does not exist.
Greg Sabino Mullane [Thu, 21 Feb 2013 02:25:49 +0000 (21:25 -0500)]
Don't assume that everyone has a home. Per github issue #23
Greg Sabino Mullane [Mon, 11 Feb 2013 21:27:40 +0000 (16:27 -0500)]
Force slave to be defined, per github issue 19: https://github.com/bucardo/check_postgres/issues/19
Greg Sabino Mullane [Mon, 11 Feb 2013 21:24:31 +0000 (16:24 -0500)]
Adjust docs from earlier remote merge.
Greg Sabino Mullane [Mon, 11 Feb 2013 21:23:46 +0000 (16:23 -0500)]
Force tests to use port 5432 in case PGPORT is conflicting.
Greg Sabino Mullane [Mon, 11 Feb 2013 21:22:25 +0000 (16:22 -0500)]
Better debug output: check for dbdir and datadir separately.
gp [Mon, 14 Jan 2013 12:45:43 +0000 (13:45 +0100)]
Fix the check_bloat SQL to take inherited tables into account, and assume 2k for non-analyzed columns.
Greg Sabino Mullane [Mon, 11 Feb 2013 20:04:09 +0000 (15:04 -0500)]
Credit recent change.
Henrik Ahlgren [Fri, 14 Sep 2012 07:11:20 +0000 (10:11 +0300)]
Get rid of "uninitialized value" warnings.
Greg Sabino Mullane [Mon, 11 Feb 2013 19:55:33 +0000 (14:55 -0500)]
Credit recent change
Mika Eloranta [Wed, 30 Jan 2013 19:23:15 +0000 (21:23 +0200)]
check_txn_idle: fix query_time --excludeuser query
Parenthesis are required:
"a OR b AND usename <> 'foo'" -> "(a OR b) AND usename <> 'foo'"
Fixes at least "query_time --includeuser/--excludeuser".
Greg Sabino Mullane [Mon, 11 Feb 2013 19:44:24 +0000 (14:44 -0500)]
Credit recent change
Michael van Bracht [Mon, 11 Feb 2013 10:38:32 +0000 (11:38 +0100)]
Fix userclause handling in check_txn_idle
For databases > Postgrs 9.1, irrelevant records are discarded
with two ORed conditions. This makes the optional userclause
associate with the right hand side of the condition. The expression
needs to be grouped to get the desired effect regardless of the
evaluation of the previous condition.
Greg Sabino Mullane [Sat, 9 Feb 2013 04:24:06 +0000 (23:24 -0500)]
Bump year to 2013 and some minor HTML adjustments.
Guillaume Lelarge [Thu, 27 Dec 2012 08:57:55 +0000 (09:57 +0100)]
Make sure hot_standby_delay doesn't return negative values
Before this patch, the hot_standby_delay action could return negative values.
To minimize the risk, we first query the slave, then the master. And to make
sure, we never get negative values, we set them to zero if we have a negative
value.
Report and ideas from Filip RembiaĆkowski.
Greg Sabino Mullane [Thu, 4 Oct 2012 03:43:54 +0000 (23:43 -0400)]
Credit recent change
Greg Sabino Mullane [Thu, 4 Oct 2012 03:41:44 +0000 (23:41 -0400)]
Make sure SQL2 is defined, per github issue 16
Josh Williams [Tue, 11 Sep 2012 16:58:29 +0000 (12:58 -0400)]
Postgres 9.2 compatibility
check_txn_idle and check_backends (and the backends test) have been
updated to use the new pg_stat_activity. And check_disk_space now uses
pg_tablespace_location().
Guillaume Lelarge [Thu, 23 Aug 2012 17:50:39 +0000 (19:50 +0200)]
Fix disk_space docs
Comments from Lance Raymond.