Limit of directory 'depth' for monitoring ?

[RobC]

Hi,

I'm trying to monitor the %free usage of some data partitions.

I have a mount point at
/data/postgres/9.1.3
with secondary mount points at
/data/postgres/9.1.3/data/db00
/data/postgres/9.1.3/data/db01
/data/postgres/9.1.3/data/db02
etc...

When monitoring /data/postgres/9.1.3 I get the expected results
When monitoring /data/postgres/9.1.3/data I get the expected results
But when monitoring /data/postgres/9.1.3/data/db00 I get no results (ZBX_NOTSUPPORTED)

if I mount /mnt/db00 to /data/postgres/9.1.3/data/db00 and point Zabbix at /mnt/db00 I get the expected results.

Does Zabbix have problems monitoring a 'deep' directory structure, or does it have issues monitoring a mount point within a mount point?

In a previous post I was looking at macro'ing the postgres version, as well as the 'DB' number (db00, db01, etc), but for the purposes of this monitoring I'm not using any macro's, just the direct path of /data/postgres/9.1.3/data/db00 - but trying to monitor the 'nested' mount point always fails with the "ZBX_NOTSUPPORTED" message.

Any clues?

Cheers,
Rob.

[RobC]

Anyone ?

[anpat89]

I believe you are correct when you say that you cannot set it to monitor a mount point within another mount point purely because it might create a loop or contradiction of what it has previously calculated.

I suppose as you found out yourself, the workaround is to create a tertiary mount point to your secondary mount points and link Zabbix to that.

I tried setting up mount points with lengthy directory listings and Zabbix was able to monitor these. However I have never had the need to monitor anything but root directory as there is no other set quota on folders or mount points.

Hope this helps.
Regards,
Andrew