Trying unsuccessfully to get space usage percentage using Pool API (/api/v2.0/pool/)

[stewartLG]

I'm the author of a Nagios plug-in that uses the TrueNAS API to check a TrueNAS server for various kinds of health and status, and it uses the 2.0 API.

https://github.com/StewLG/check_truenas_extended_play

A user/fellow developer has submitted a PR to me that attempts to use the data coming back about the zpools from the /api/v2.0/pool/ API call to work out the used/free capacity for the zpool. It does this by adding up the topology.data.size and topology.data.allocated values for all a zpool's vdevs, and doing a percentage calculation.

This math isn't working out to agree with the TrueNAS GUI, and it's off enough that it's not yet fit for deployment (10-20% off).

Are the topology.data.size and topology.data.allocated values inclusive or exclusive of parity? Do they ignore or include the RaidZ level? How can I get reasonably close to the output I see in zpool_capacity.py, which is just calling 'zpool list -H -o cap'?

 

[ClassicGOD]

/api/v2.0/pool/ endpoint contains pool data - the values there are as ZFS sees them so depending on a pool configuration. For example for RAIDz2 pool it sums up space of all drives, for Mirror you see only "half". It's equivalent to "zpool list" output.

what you want is /pool/dataset endpoint. Every pool should contain top level dataset with the same name as the pool. There you'll find both .used and .available value in raw and parsed formats.

 

[stewartLG]

Thank you for responding!

I figured out that I had to use datasets already, and the script is now working properly. I am sorry for not following up here to answer my own question.