Ad Widget

Collapse

All history syncers stuck at every hour

Collapse
X
Collapse
 
  • Page of 1
  • Time
  • Show
Clear All
new posts
Previous template Next
  • Pedro.Almeida
    Junior Member
    • Sep 2014
    • 22
    #1

    All history syncers stuck at every hour

    We have a bit of old data on the DB (MySQL) for our 40K items.
    Issue here is, every hour data collecting seems to stop at minute 0.
    Even disabling housekeeper (running every hour) the problem.
    Increased the history syncers from 4 to 8, all get "stuck".

    Only thing ocuring at an hourly rate precision seems to be the trend aggregation. Which process is responsible for this?

    I have these, but none seem to match:

    Code:
    zabbix    1992  1984  0 11:05 ?        00:00:20 zabbix_server: configuration syncer [synced configuration in 2.305969 sec, idle 60 sec]
zabbix    1993  1984  0 11:05 ?        00:00:00 zabbix_server: db watchdog [synced alerts config in 0.000456 sec, idle 60 sec]
zabbix    1994  1984  0 11:05 ?        00:00:00 zabbix_server: poller #1 [got 0 values in 0.000003 sec, idle 1 sec]
... (several of previous)
zabbix    2044  1984  0 11:05 ?        00:00:00 zabbix_server: unreachable poller #1 [got 0 values in 0.000003 sec, idle 5 sec]
... (several of previous)
zabbix    2074  1984  0 11:05 ?        00:00:00 zabbix_server: trapper #1 [processed data in 0.000328 sec, waiting for connection]
... (several of previous)
zabbix    2084  1984  0 11:05 ?        00:00:00 zabbix_server: icmp pinger #1 [got 0 values in 0.000003 sec, idle 1 sec]
zabbix    2120  1984  0 11:05 ?        00:00:00 zabbix_server: alerter [sent alerts: 0 success, 0 fail in 0.000112 sec, idle 30 sec]
zabbix    2121  1984  0 11:05 ?        00:00:00 zabbix_server: housekeeper [removing old history and trends]
zabbix    2122  1984  0 11:05 ?        00:00:00 zabbix_server: timer #1 [processed 170 triggers, 0 events in 0.002817 sec, 0 maintenances in 0.000000 sec, idle 30 sec]
zabbix    2123  1984  0 11:05 ?        00:00:00 zabbix_server: http poller #1 [got 0 values in 0.000452 sec, idle 5 sec]
zabbix    2124  1984  0 11:05 ?        00:00:00 zabbix_server: discoverer #1 [processed 0 rules in 0.000606 sec, idle 60 sec]
zabbix    2125  1984  0 11:05 ?        00:00:01 zabbix_server: history syncer #1 [synced 0 items in 0.000292 sec, idle 5 sec]
zabbix    2135  1984  0 11:05 ?        00:00:00 zabbix_server: escalator [processed 0 escalations in 0.000131 sec, idle 3 sec]
zabbix    2136  1984  0 11:05 ?        00:00:00 zabbix_server: snmp trapper [processed data in 0.000009 sec, idle 1 sec]
zabbix    2137  1984  0 11:05 ?        00:00:00 zabbix_server: proxy poller #1 [exchanged data with 0 proxies in 0.000002 sec, idle 5 sec]
zabbix    2138  1984  0 11:05 ?        00:00:00 zabbix_server: self-monitoring [processed data in 0.000015 sec, idle 1 sec]
    Tags: None
    • Pedro.Almeida
      Junior Member
      • Sep 2014
      • 22
      #2
      Found the process...

      Guess it's the history sync. Can you confirm?
      Now it's passing the hour tick. DB performance issues, perhaps?
      Anyone been through this problem? Besides housekeeper working like a dog as you may notice...

        Comment

        • BDiE8VNy
          Senior Member
          • Apr 2010
          • 680
          #3
          Trends get written by history syncers, yes:
          main_dbsyncer_loop -> DCsync_history() -> DCmass_update_trends() -> DCflush_trends()

            Comment

            Previous template Next
            Working...