SMART Tests on "All Disks" Does Not Appear to Execute on NVMe Disks

TrueNAS General
,
1

I am running TrueNAS Scale ElectricEel-24.10.2. All of my pools are running the latest ZFS as far as I can tell (i.e. absence of Upgrade button).

I see now that Electric Eel can execute and show SMART test results for NVMe drives, or at least mine. However, I have “Periodic S.M.A.R.T. Tests” scheduled to run on “All Disks” - a long one and a short one. The tests are running as scheduled for the HDDs, but aren’t running at all for the NVMe drives - I have to run these manually. Does “All Disks” not include NVMe drives?

2

Indeed, TrueNAS still does not support SMART tests on NVMe.
Look into the Multi-Report script:

1 Like
3

What exactly do you mean by “support”? I’m confused by this without clarification because:

  1. The release notes for Electric Eel state: “UI support for NVMe S.M.A.R.T. tests.”
  2. I’ve manually executed SMART tests on my NVMe drives in TrueNAS. They do take some time to finish, so they’re doing something

I’m not suggesting you’re wrong/TrueNAS does support SMART tests for NVMe - I’m just seeing conflicting things.

4

TrueNAS still does not support scheduling SMART tests on NVMe.

1 Like
5

The GUI looks like you can schedule an NVMe SMART Test however the test is never actually run. This remains true in 25.04.RC1 as well and I suspect for the final release too. TrueNAS will monitor for SMARTD received errors but that is only part of the coverage you need. In other words, if TrueNAS detects a bad sector on an NVMe, it should report it to you.

6

I run Scrutiny instead.

1 Like
7

I understand that scheduling isn’t working, but if you manually run a SMART test from the web UI (Storage → Disks) on an NVMe, is it actually running the test and properly reporting results?

8

Short answer - No, not for NVMe.

9

Good to know - thank you.

10

I definitely test and track this with each TrueNAS release. Once it works, all will know.

1 Like
11

Any updates on this? I’m experiencing the same problem and would love to restart using the SMART setting in “Data protection”.

In the meantime, I’m using cronjobs to execute tests on all my four nvmes in the middle of the night using these commands on separate days (my nvmes being named nvme0n1 … nvme3n1):

Short tests:
for d in /dev/nvme{0..3}n1; do sudo smartctl -t short "$d"; done

Long tests:
for d in /dev/nvme{0..3}n1; do sudo smartctl -t long "$d"; done

This seems to log the test results properly in the GUI (“Storage” > “Disks” > nvme…n1 > “S.M.A.R.T. Test Results”) but I’m not sure if the notification on SMART errors (via mail) works alright. Does anyone have any experience on this?

12

You can install the “Scrutiny” app to collect SMART data of all of your drives.