Skip to content
Commits on Oct 2, 2012
  1. @steveschnepp
  2. @steveschnepp

    Merge pull request #30 from gentoo/devel

    steveschnepp committed Oct 2, 2012
    fw_conntrack: count both ipv4 and ipv6 connections together.
  3. @steveschnepp

    plugins: fix plugin vserver_cpu_ for 9+ cpu cores

    Jon Whiteman committed with steveschnepp Oct 2, 2012
    Issue is that grep "cpu 1" matches multiple CPU lines when CPU1 & CPU12 exists.
    
    Closes: D#681938
  4. @Flameeyes

    fw_conntrack: count both ipv4 and ipv6 connections together.

    Flameeyes committed Oct 1, 2012
    Without this change, only ipv4 connections are counted and they might
    float more than expected.
Commits on Sep 30, 2012
  1. @kenyon
Commits on Sep 29, 2012
  1. @steveschnepp

    Merge pull request #29 from bpineau/postgresql_fixes

    steveschnepp committed Sep 29, 2012
    PostgreSQL plugins fixes
Commits on Sep 28, 2012
  1. @bpineau

    Fix postgres_*prepared_xacts_ plugins.

    bpineau committed Sep 28, 2012
    postgres_oldest_prepared_xact_ and postgres_prepared_xacts_ uses the same
    "wildcardfilter" ("WHERE database=?") in their "configquery" and their
    "basequery".
    
    But those two queries uses two different tables, with different columns
    names to refer to the database name ; the "pg_database" table has a "datname"
    column, while the "pg_prepared_xacts" view has a "database" column.
    So we can't add a "WHERE $something=?" filter as is in both queries.
    
    Let's add a gratuitous JOIN in the "basequery" (joining pg_database, so
    we're sure any existing database name will be joined) ; then we can
    use a common "WHERE datname=?" filter on both queries.
    
    Those tables columns are still the same since xact statistics where
    introduced, in PostgreSQL 8.1, so this plugin's config had always been broken.
    http://www.postgresql.org/docs/8.1/static/release-8-1.html
    http://www.postgresql.org/docs/8.1/static/view-pg-prepared-xacts.html
    http://www.postgresql.org/docs/8.1/static/catalog-pg-database.html
  2. @bpineau

    Fix postgresql_* plugins for PostgreSQL 9.2 support.

    bpineau committed Sep 28, 2012
    The pg_stat_activity system view, used in many Munin postgresql_* plugins,
    changed in the recent PostgreSQL 9.2 release. In particular:
    - the 'procpid' column was renamed to 'pid'.
    - the 'current_query' column was renamed to 'query', and contains the last
      executed query (even when the connection is idling).
    - an additional 'state' field was added (replaces and completes the old
      current_query='<IDLE>' tests and similar).
    
    Cf. the new pg_stat_activity view structure documentation here :
    http://www.postgresql.org/docs/9.2/static/monitoring-stats.html#PG-STAT-ACTIVITY-VIEW
    as opposed to the structure for an older PostgreSQL release here :
    http://www.postgresql.org/docs/9.1/static/monitoring-stats.html#PG-STAT-ACTIVITY-VIEW
    
    The Munin::Plugin::Pgsql framework allows us to update the plugins to
    support pg 9.2 without breaking compatibility for older versions
    (changes tested on PostgreSQL 9.2 and 9.0).
Commits on Sep 25, 2012
  1. @steveschnepp
  2. @steveschnepp

    plugins: use Storable::nstore() instead of store()

    steveschnepp committed Sep 25, 2012
    Storable::store() isn't very reliable accross Perl versions.
    Since ::nstorable() is, well, we change.
    
    In the future, all those plugins shall use the state handling from
    Munin::Plugin.
    
    Closes: #1259, D#688528
  3. @steveschnepp
Commits on Sep 24, 2012
  1. @steveschnepp

    Merge pull request #27 from gentoo/devel

    steveschnepp committed Sep 24, 2012
    Gentoo fixes
  2. @steveschnepp

    Merge pull request #26 from kjmkznr/devel

    steveschnepp committed Sep 24, 2012
    Fix GOODSH bad interpreter
  3. @Flameeyes

    ejabberd_: allow overriding the ejabberd.cfg path and add /etc/jabber…

    Flameeyes committed Sep 24, 2012
    … to search paths.
    
    This fixes ejabberd_ plugin on Gentoo and solves one of the known bugs
    of the plugin.
  4. @Flameeyes
  5. @Flameeyes

    apc_nis: improve configuration.

    Flameeyes committed Aug 23, 2012
    Use a static title for the graph so that it's easier to integrate with
    Icinga/Munin. The host/port address pair is given in the graph's info
    instead.
    
    Read the LOTRANS/HITRANS value during configuration and use them as
    warning canaries for the line voltage.
  6. @Flameeyes
  7. @Flameeyes
  8. Makefile.config: fix bad interpreter

    Kazunori Kojima committed Sep 25, 2012
Commits on Sep 22, 2012
  1. @tgunkel @steveschnepp

    plugins: auto detect SATA in hddtemp_smartctl

    tgunkel committed with steveschnepp Sep 22, 2012
    It needs smartmontools 5.38 or higher, so it might not be
    compatible with older systems like RHEL4.
    
    Closes: D#497400
Commits on Sep 20, 2012
  1. @steveschnepp

    master: split rrd updates when using rrdcached

    steveschnepp committed Sep 20, 2012
    Commands sended to rrdcached are limited to 4KiB. For really large vector
    updates do to spoolfetching or supersampling, this gets too small, so we do
    unit cmd sending and let rrdcached buffer it out for us.
Commits on Sep 14, 2012
  1. @steveschnepp

    asyncd: fix --spooldir help

    steveschnepp committed Sep 14, 2012
Commits on Sep 13, 2012
  1. @steveschnepp

    plugins: fix apt_all plugin statedir for cron

    steveschnepp committed Sep 13, 2012
    The apt_all plugin has its state updated in cron. There the ENV var
    MUNIN_PLUGSTATE doesn't exist, so we need to set a default.
    
    Closes: D#687495
Commits on Sep 12, 2012
  1. @steveschnepp
  2. @ze42

    munin-master: htmlconf.storable is back (fix previous commit)

    ze42 committed Sep 12, 2012
    previous commit provided a broken state. now fixed.
  3. @steveschnepp
Commits on Sep 11, 2012
  1. @steveschnepp
  2. @ze42

    munin-master: htmlconf.storable is back

    ze42 committed Sep 11, 2012
    htmlconf.storable - how it should work
      munin-html writes it
      munin-cgi-html reads it
    
    munin-html (no cgi)
       written, never read
    munin-cgi-html (with munin-html called)
      written, and read to avoid lag generating it
    munin-cgi-html (without munin-html being called by cron)
      The file should NOT exist. Will be generated in ram on the fly.
Commits on Sep 8, 2012
  1. @steveschnepp

    plugins: fix import in ipmi_sensor_

    steveschnepp committed Sep 8, 2012
    Closes: D#686982
  2. @steveschnepp

    master: remove useless log lines, send to debug

    steveschnepp committed Sep 8, 2012
    Several log lines were only useful for debugging, but they stayed as INFO.
    Removal was necessary as the default is cron gfx again, so /var/log will fill
    very fast. It completes 789c59e that only
    cared about warnings.
    
    Closes: #1251
Commits on Sep 6, 2012
  1. @steveschnepp

    master: fix too many warnings in munin-graph.log

    steveschnepp committed Sep 6, 2012
    When graph_strategy was missing, a warning was emitted since the value is
    undef.
    
    Closes: #1251
Commits on Sep 5, 2012
  1. @steveschnepp

    Merge pull request #24 from ze42/devel

    steveschnepp committed Sep 5, 2012
    m::m::Utils - munin_*config*: rewritten functions
  2. @steveschnepp
  3. @steveschnepp

    node: remove File::Path as it's not needed

    steveschnepp committed Sep 5, 2012
    Having the import makes Perl5.8 break, as it doesn't have the make_path sub.
    Removing it since it's not even needed anymore.
Commits on Sep 3, 2012
  1. @ze42

    master: fix new config reading for m-u 1st run

    ze42 committed Sep 3, 2012
    munin_readconfig_part now behaves correctly on first run, when there
    isn't any datafile yet. Other function do need their file to be already
    generated, and might exit if not.
    
    Affects bootstrap for munin-update and munin-limits
Something went wrong with that request. Please try again.