Ad Widget

Collapse

[zabbix-1.6.1] housekeeper issue

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • tekknokrat
    Senior Member
    • Sep 2008
    • 140

    #1

    [zabbix-1.6.1] housekeeper issue

    I have changed configuration for housekeeper from 365 to 20 days, item's history the same because our postgresql database grows to fast.
    So I hoped it would now also remove old entries but nothing happens. Database grows exactly the same. How can I resolve that?
  • Calimero
    Senior Member
    • Nov 2006
    • 481

    #2
    What did you change exactly ?

    The "Keep Days / Trends" parameters of each item ?

    Did you "vacuum" your zabbix database in pgsql ?

    Comment

    • tekknokrat
      Senior Member
      • Sep 2008
      • 140

      #3
      Originally posted by Calimero
      What did you change exactly ?

      The "Keep Days / Trends" parameters of each item ?

      Did you "vacuum" your zabbix database in pgsql ?
      Yes only the history, not the trends. General housekeeper settings.
      VACCUUM FULL done, also autovacuum is enabled.

      When grabbing the log files from the last seven days they always report:

      Code:
      20681:20081210:213450 Deleted 0 records from history and trends
      although housekeeping should be continous process, shouldnt it?

      Comment

      • tekknokrat
        Senior Member
        • Sep 2008
        • 140

        #4
        I would like to delete the history of all items ( via bulk ) . Why isnt there some option like clean history but only cleaning data before history intervall.

        Comment

        • tekknokrat
          Senior Member
          • Sep 2008
          • 140

          #5
          I read in v.16 016:

          Code:
          HousekkepingFrequency
          If PostgreSQL is used set the
          value to 24 as it will perform
          command VACUUM.
          So does this means that no normal zabbix internal houskeeping applies here
          or is VACCUUM additionally done?

          Comment

          • tekknokrat
            Senior Member
            • Sep 2008
            • 140

            #6
            OK got the problem temp. in control. I dropped the database and restored a dump.
            Database Size is now at 738MB. Also VACCUUM ANALYZE is run through a cron job (debian package) and I manually did some VACUUM FULL the size didnt decreased another way.
            What else can I do to stop bloating of db? Perhaps some overoptimisation in config?

            pgsql configuration file: http://pastebin.ca/1303690

            Comment

            Working...