Uploaded image for project: 'ZABBIX BUGS AND ISSUES'
  1. ZABBIX BUGS AND ISSUES
  2. ZBX-17108

unreal spike logged for internal processes (more than 100% busy) on graphs, caused by major upgrade

    XMLWordPrintable

    Details

    • Team:
      Team A
    • Sprint:
      Sprint 59 (Dec 2019), Sprint 60 (Jan 2020), Sprint 61 (Feb 2020), Sprint 62 (Mar 2020)
    • Story Points:
      0.25

      Description

      Here is major 3.4 -> 4.4 upgrade log:

        5159:20191217:143705.304 Starting Zabbix Server. Zabbix 4.4.3 (revision 6cad6f888e).
        5159:20191217:143705.304 ****** Enabled features ******
        5159:20191217:143705.304 SNMP monitoring:           YES
        5159:20191217:143705.304 IPMI monitoring:           YES
        5159:20191217:143705.304 Web monitoring:            YES
        5159:20191217:143705.304 VMware monitoring:         YES
        5159:20191217:143705.305 SMTP authentication:       YES
        5159:20191217:143705.305 ODBC:                      YES
        5159:20191217:143705.305 SSH2 support:              YES
        5159:20191217:143705.305 IPv6 support:              YES
        5159:20191217:143705.305 TLS support:               YES
        5159:20191217:143705.305 ******************************
        5159:20191217:143705.305 using configuration file: /etc/zabbix/zabbix_server.conf
        5159:20191217:143705.317 current database version (mandatory/optional): 03050046/03050046
        5159:20191217:143705.317 required mandatory version: 04040000
        5159:20191217:143705.317 starting automatic database upgrade
      ...
        5159:20191217:144529.032 completed 100% of database upgrade
        5159:20191217:144529.033 database upgrade fully completed
        5159:20191217:144538.309 starting event name update forced by database upgrade
        5159:20191217:144539.126 completed 1% of event name update
      ... 
        5159:20191217:144732.434 completed 99% of event name update
        5159:20191217:144732.823 completed 100% of event name update
        5159:20191217:144732.829 event name update completed
        5159:20191217:144733.292 server #0 started [main process]
        5528:20191217:144733.294 server #1 started [configuration syncer #1]
        5529:20191217:144733.294 server #2 started [ipmi manager #1]
      ....
        5689:20191217:144734.088 server #158 started [preprocessing worker #1]
        5661:20191217:144734.299 cannot process proxy "zabbixproxy-01.xxxxx": protocol version 3.4 is not supported anymore
      .... normal operations
      

      (ignore initial dbversion 03050046, caused by first failed attempt)

      See legend on graphs, history:

      Of course it will stay is trends too:

      Such numbers may mislead, so zabbix should not calculate such wrong numbers.

      Maybe it's caused by specific change in 4.0 - exiting events name update.
      https://www.zabbix.com/documentation/4.0/manual/installation/upgrade_notes_400#problem_name_generation

      But, after the upgrade, I needed to restart zabbix server a few times. And every time after restart, server logs 100% busy spikes for many "internal process busy level" items.
      And what possibly would be useful and acceptable - just do NOT collect internal statistic during very first minute after daemon started. I.e start to collect internal statistic after 1 minute.
      It probably would be really useful for all the installations!

        Attachments

        1. strange values after upgrade.png
          84 kB
          Oleksii Zagorskyi
        2. strange values after upgrade trends.png
          62 kB
          Oleksii Zagorskyi

          Activity

            People

            Assignee:
            mabele Martins Abele
            Reporter:
            zalex_ua Oleksii Zagorskyi
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

              Dates

              Created:
              Updated:
              Resolved: