Ad Widget

Collapse

Unreachable poller 100%

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • syndeysider
    Senior Member
    • Oct 2013
    • 115

    #1

    Unreachable poller 100%

    Hi

    Zabbix 2.4.7
    Redhat 7
    8GB Ram
    4 Core CPU

    Zabbix Proxy on seperate server to Zabbix Server.

    I'm finding that lately my zabbix_proxy process is routinely, around every 2 hours, choking on the unreachable poller process (pic).

    Slowly over 2 hours, more unreachable poller processes spawn and eventually the server chokes with 100% CPU utilization.

    I've attempted changing various zabbix_proxy Poller numbers/intervals etc. but to no avail. Any idea where i'd start to troubleshoot what's causing the Ureachable pollers to choke?
  • ingus.vilnis
    Senior Member
    Zabbix Certified Trainer
    Zabbix Certified SpecialistZabbix Certified Professional
    • Mar 2014
    • 908

    #2
    Hi,

    What about Zabbix performance graphs?


    What do you see in "Zabbix data gathering process busy graph" (and all other Zabbix server graphs?

    And why did you mention the proxy? What setup are you using there?

    Best Regards,
    Ingus

    Comment

    • syndeysider
      Senior Member
      • Oct 2013
      • 115

      #3
      hi

      I have a backend zabbix-server/mysql cluster that processes data/triggers/actions etc. The proxy is the single server in the DC that is responsible for collection/discovery etc. All 900+ devices (SNMP) and servers are configured to either be polled by or send data to the zabbix-proxy.

      I have always had this configuration, since 2.0.4, the growth rate of devices/items since 2.0.4 has been around 8%, which I believe i've factored into the number of pollers/memory etc. on the proxy.

      Attached is the proxies data gathering graph over the past 14d. I've been playing with the number of StartDiscovers for another issue, hence, the trend up for this Poller.

      Comment

      • syndeysider
        Senior Member
        • Oct 2013
        • 115

        #4


        Seem's this is related to my issue as strace ans lsof return same symptoms.

        Comment

        • ingus.vilnis
          Senior Member
          Zabbix Certified Trainer
          Zabbix Certified SpecialistZabbix Certified Professional
          • Mar 2014
          • 908

          #5
          Hi,

          Are you using IPMI monitoring?
          If yes, try temporary disabling them to see if that helps. Not sure if you are affected by ZBX-4823 though.

          Regarding Discoverers try setting them as many as you have discovery rules actually configured.

          Best Regards,
          Ingus

          Comment

          • syndeysider
            Senior Member
            • Oct 2013
            • 115

            #6
            Thanks. This has indeed helped. After some more investigation I've found that ZBX-4823 is definitely affecting me. This is ok for now though. IPMI is not as important as a stable Proxy right now.

            Comment

            Working...