Ad Widget

Collapse

buffer is full, cannot store persistent value

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • djp210
    Junior Member
    • Aug 2016
    • 15

    #1

    buffer is full, cannot store persistent value

    We are using Zabbix 3.0. On a couple Windows 2008 R2 servers, we have the Zabbix 3.0.0 agent installed. Among other things, the agent is monitoring a large, busy log file. One server, the file is about 56MB, and on the other, it's 69MB.

    In the zabbix_agentd.log file on both servers, we're occasionally seeing these messages:

    2836:20170303:133128.541 buffer is full, cannot store persistent value
    2836:20170303:133149.650 buffer is full, cannot store persistent value
    2836:20170303:133210.666 buffer is full, cannot store persistent value
    2836:20170303:133231.681 buffer is full, cannot store persistent value
    2836:20170303:133252.681 buffer is full, cannot store persistent value
    2836:20170303:133313.713 buffer is full, cannot store persistent value

    I've been unable to find anything online that explains this message. I'm mentioning the size and activity of the monitored file because I don't know if that has anything to do with the issue we're seeing.

    It seems that this monitor is running behind, in that the Latest Data screen for one of the two servers currently shows a last check of 12:43 p.m. and the log entry it found includes a time stamp of 11:28 a.m. (The Item does not include a log time format setting, mainly because Zabbix doesn't seem to support a text-based month. So the Latest Data screen would seem to indicate it found a log line written at 11:28 a.m. at 12:43 p.m.)

    Any ideas what's happening and what can be done to resolve the issue?

    Thanks.
  • jan.garaj
    Senior Member
    Zabbix Certified Specialist
    • Jan 2010
    • 506

    #2
    It looks like you are trying to send a lot of logs to the zabbix server.

    Try to set increase BufferSize - https://www.zabbix.com/documentation.../zabbix_agentd
    or add more conditions into your log monitoring items (for example log only errors, not everything) - https://www.zabbix.com/documentation...ypes/log_items
    Devops Monitoring Expert advice: Dockerize/automate/monitor all the things.
    My DevOps stack: Docker / Kubernetes / Mesos / ECS / Terraform / Elasticsearch / Zabbix / Grafana / Puppet / Ansible / Vagrant

    Comment

    • djp210
      Junior Member
      • Aug 2016
      • 15

      #3
      BufferSize

      OK. Our agents are set to the default BufferSize=100.

      We already have the log items collecting only error messages. There just happens to be a lot of them in these logs.

      I've updated BufferSize to 300 on these two servers and restarted their agents. Let's see if that helps.

      Thanks.

      Comment

      Working...