Ad Widget

Collapse

Error doing snmp_open() or +40000 items in queue

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • sducoin
    Junior Member
    • Dec 2011
    • 9

    #1

    Error doing snmp_open() or +40000 items in queue

    Hi all,
    I've been using zabbix for a couple of years now, monitoring both servers with zabbix agents and snmp switches/appliances.
    All worked fine until the last two weeks. The queue length started to grow and no item was being updated. I finally found that setting the value SourceIP=zabbix_server_ip in zabbix_server.conf solved the problem... except for snmp monitoring. All the snmp items became unsupported - I found lots of "Error doing snmp_open()" in the zabbix_server log file.
    I suspect a network issue but I have no clue
    Any help will be appreciated

    My config :
    zabbix 1.8.9
    ubuntu maverick
    postgresql 8.4

    Cheers.
  • sducoin
    Junior Member
    • Dec 2011
    • 9

    #2
    disabling option SourceIP in zabbix_server.conf doesnt solve the problem

    If I disable option SourceIP I got 40000+ item in queue
    If I enable it, all the snmp items become unsupported
    regards,

    Sebastian

    Comment

    • dima_dm
      Senior Member
      • Dec 2009
      • 2697

      #3
      disable option SourceIP
      Use
      zabbix[process,<type>,<mode>,<state>]
      zabbix[wcache,<cache>,<mode>]
      zabbix[rcache,<cache>,<mode>]
      to collect statistics and adjust the values wcache, rcache, process in zabbix_server.conf.
      http://www.zabbix.com/documentation/...l/config/items
      Last edited by dima_dm; 07-12-2011, 15:48. Reason: add zabbix_server.conf

      Comment

      • sducoin
        Junior Member
        • Dec 2011
        • 9

        #4
        disabled option SourceIP in zabbix_server.conf :

        here are the values in the attached files.
        We can see zabbix suddenly stopped updating data.
        Attached Files

        Comment

        • dima_dm
          Senior Member
          • Dec 2009
          • 2697

          #5
          Is running process zabbix_server in this time?

          Comment

          • sducoin
            Junior Member
            • Dec 2011
            • 9

            #6
            yes it is

            I got about 100 zabbix_server processes and 28 startPoller in zabbix_server.conf

            regards,

            Sebastian

            Comment

            • dima_dm
              Senior Member
              • Dec 2009
              • 2697

              #7
              Can you check database server performance and error log?

              Comment

              • sducoin
                Junior Member
                • Dec 2011
                • 9

                #8
                postgresql + pgouncer

                I use postgresql 8.4 and pgbouncer as a connections pooler and it works just fine :

                got auto-vacuum and analyse logs for posgtresql :

                Code:
                2011-12-05 14:54:27 CET LOG:  automatic analyze of table "zabbix.public.history" system usage: CPU 0.34s/0.43u sec elapsed 30.13 sec
                2011-12-05 15:54:55 CET LOG:  automatic analyze of table "zabbix.public.history_uint" system usage: CPU 0.44s/0.57u sec elapsed 56.97 sec
                2011-12-06 01:24:43 CET LOG:  automatic analyze of table "zabbix.public.history" system usage: CPU 0.26s/0.53u sec elapsed 31.87 sec
                2011-12-06 03:33:00 CET LOG:  automatic analyze of table "zabbix.public.history_uint" system usage: CPU 0.47s/0.58u sec elapsed 45.47 sec
                2011-12-06 16:48:11 CET LOG:  automatic analyze of table "zabbix.public.history" system usage: CPU 0.38s/0.54u sec elapsed 36.00 sec
                2011-12-06 17:59:27 CET LOG:  automatic analyze of table "zabbix.public.history_uint" system usage: CPU 0.45s/0.51u sec elapsed 50.07 sec
                2011-12-07 19:57:39 CET LOG:  automatic analyze of table "zabbix.public.history" system usage: CPU 0.36s/0.49u sec elapsed 34.65 sec
                2011-12-07 20:26:52 CET LOG:  automatic analyze of table "zabbix.public.history_uint" system usage: CPU 0.44s/0.49u sec elapsed 47.36 sec
                and some connections stats for pgbouncer :

                Code:
                2011-12-08 13:19:28.387 1289 LOG Stats: 1384 req/s, in 86546 b/s, out 129232 b/s,query 875 us
                [...]
                2011-12-08 13:20:36.423 1289 LOG C-0xc1aff0: zabbix/[email protected]:36543 login attempt: db=zabbix user=zabbix
                2011-12-08 13:20:36.439 1289 LOG C-0xc242e0: zabbix/[email protected]:36545 login attempt: db=zabbix user=zabbix
                2011-12-08 13:20:36.453 1289 LOG C-0xbf9f80: zabbix/[email protected]:36547 login attempt: db=zabbix user=zabbix
                2011-12-08 13:20:36.501 1289 LOG C-0xc26280: zabbix/[email protected]:36549 login attempt: db=zabbix user=zabbix
                2011-12-08 13:20:37.421 1289 LOG C-0xc1aff0: zabbix/[email protected]:36543 closing because: client close request (age=0)
                2011-12-08 13:20:37.440 1289 LOG C-0xc1e010: zabbix/[email protected]:36541 closing because: client close request (age=1)
                2011-12-08 13:20:37.489 1289 LOG C-0xbf9f80: zabbix/[email protected]:36547 closing because: client close request (age=1)
                2011-12-08 13:20:37.847 1289 LOG C-0xc242e0: zabbix/[email protected]:36545 closing because: client close request (age=1)
                2011-12-08 13:20:38.020 1289 LOG C-0xc26280: zabbix/[email protected]:36549 closing because: client close request (age=1)
                2011-12-08 13:20:43.270 1289 LOG C-0xc26280: zabbix/[email protected]:36562 login attempt: db=zabbix user=zabbix
                2011-12-08 13:20:43.565 1289 LOG C-0xc26280: zabbix/[email protected]:36562 closing because: client close request (age=0)
                2011-12-08 13:20:43.853 1289 LOG C-0xc26280: zabbix/[email protected]:36564 login attempt: db=zabbix user=zabbix
                2011-12-08 13:20:43.949 1289 LOG C-0xc26280: zabbix/[email protected]:36564 closing because: client close request (age=0)
                2011-12-08 13:20:50.934 1289 LOG C-0xc26280: zabbix/[email protected]:36571 login attempt: db=zabbix user=zabbix
                Last edited by sducoin; 08-12-2011, 17:01.

                Comment

                • sducoin
                  Junior Member
                  • Dec 2011
                  • 9

                  #9
                  internal postgresql statistics :

                  I use postbix to monitor my postgresql server
                  we can see values not being updated after setting the SourceIP value... except for postbix values - zabbix trapper is not impacted by this issue.

                  regards

                  Sebastian
                  Attached Files

                  Comment

                  • dima_dm
                    Senior Member
                    • Dec 2009
                    • 2697

                    #10
                    Check Timeout option.
                    /etc/zabbix/zabbix_server.conf
                    Code:
                    ### Option: Timeout
                    #       Specifies how long we wait for agent, SNMP device or external check (in seconds).
                    #
                    # Mandatory: no
                    # Range: 1-30
                    # Default:
                    Timeout=30
                    And zabbix_agentd.conf Timeout option.
                    Last edited by dima_dm; 08-12-2011, 17:23.

                    Comment

                    • sducoin
                      Junior Member
                      • Dec 2011
                      • 9

                      #11
                      already set

                      here is my zabbix_server.conf :

                      Code:
                      StartPollers=28
                      StartPollersUnreachable=16
                      StartTrappers=16
                      StartPingers=12
                      StartDiscoverers=6
                      StartDBSyncers=16
                      #SourceIP=10.4.1.34
                      HousekeepingFrequency=22
                      MaxHousekeeperDelete=2000
                      SenderFrequency=30
                      DebugLevel=3
                      Timeout=30
                      UnreachablePeriod=300
                      UnavailableDelay=120
                      PidFile=/var/run/zabbix/zabbix_server.pid
                      LogFile=/var/log/zabbix-server/zabbix_server.log
                      AlertScriptsPath=/etc/zabbix/alert.d/
                      ExternalScripts=/home/zabbix/bin
                      DBHost=127.0.0.1
                      DBPort=6432
                      DBName=zabbix
                      DBUser=zabbix
                      DBPassword=xxxx
                      CacheSize=384M
                      TrendCacheSize=128M
                      CacheUpdateFrequency=300
                      LogSlowQueries=5000

                      Comment

                      • sducoin
                        Junior Member
                        • Dec 2011
                        • 9

                        #12
                        Dirty ipam dns

                        Hi,
                        The problem came from our dns server (ipam 3.5) who was in a dirty state. Still responding but...
                        Anyway after a cold reboot of the ipam dns server and SourceIP commented, zabbix worked fine and good as the old good time.

                        Personal reminder : think about ipam fine monitoring.

                        Thank you dima_dm for your help.
                        have a good day

                        cheers

                        Comment

                        Working...