We have Zabbix 5.0.2 on Centos 7.8. We monitor VMWARE (6.5) virtuals WITH docker installed on Centos 7. Monitoring tens of virtuals without problems with one exception.
Problem:
VMWare virtual machine with Docker monitored by VMWare templates.
Discovery templates detect/discover mounted FileSystems correctly and corectly detect ALL mounted docker temporary FileSystems also. After deletion of temporary filesystem the items in zabbix are correctly marked as "not available".
BUT for ALL docker temporary filesystems are data written into Zabbix database as it would exists.
After some time we have in Zabbix database hunderts items for mounted filesystem WITH the same values. After 30 days this nonexistent filesystems are correctly removed from Zabbix database.
In short - for this Centos/Docker system we have many EMPTY /var/lib/docker/devicemapper/mnt/xxxx itmes (see picture) AND Zabbix are getting THE SAME data for all such "phantom" filesystems from VMWare.
I think that source of problem lies in length of filesystem name, but I cannot find more usable details.
Problem:
VMWare virtual machine with Docker monitored by VMWare templates.
Discovery templates detect/discover mounted FileSystems correctly and corectly detect ALL mounted docker temporary FileSystems also. After deletion of temporary filesystem the items in zabbix are correctly marked as "not available".
BUT for ALL docker temporary filesystems are data written into Zabbix database as it would exists.
After some time we have in Zabbix database hunderts items for mounted filesystem WITH the same values. After 30 days this nonexistent filesystems are correctly removed from Zabbix database.
In short - for this Centos/Docker system we have many EMPTY /var/lib/docker/devicemapper/mnt/xxxx itmes (see picture) AND Zabbix are getting THE SAME data for all such "phantom" filesystems from VMWare.
I think that source of problem lies in length of filesystem name, but I cannot find more usable details.
Attached Files