[BUG] ProxMenux Monitor prevents HDD standby / wakes disks from standby

[Turtletrumpet]

Hi,

first of all: thank you for ProxMenux. The Monitor is a very useful tool and gives a great quick overview of the Proxmox host.

I noticed one issue with the Monitor on my system:

Environment

Proxmox VE 9
ProxMenux 1.2.2
Multiple SATA HDDs used for SnapRAID data/parity disks
HDD standby configured via hdparm and, for one Seagate disk, via hd-idle

Observed behavior

When the ProxMenux Monitor is running, my HDDs do not reliably enter standby, or they wake up again shortly after entering standby.

After stopping/disabling the ProxMenux Monitor, the disks enter standby correctly and remain in standby until there is an actual disk access.

Expected behavior

The Monitor should ideally avoid waking disks that are already in standby, or provide an option to disable disk/SMART/temperature polling for HDDs.

Possible cause

It looks like the Monitor may periodically query disk, SMART, or temperature information. Such checks can wake SATA HDDs or prevent them from becoming idle long enough to spin down.

A standby-safe approach could be to use something like smartctl -n standby where applicable, or to provide a setting to disable SMART/temperature polling for disks.

Workaround

Disabling the Monitor service resolves the issue:

systemctl disable --now proxmenux-monitor

It would be great if the Monitor could keep its current functionality while avoiding unwanted HDD spin-ups, especially for NAS/SnapRAID setups where disk standby is intentional.

Thanks again for the project.

[MacRimi]

Hi @Turtletrumpet, fixed in the upcoming build — thank you for the clean diagnosis and the workaround note, that helped a lot.

What was happening (your hypothesis was exactly right):

• The temperature poller ran smartctl -A -j /dev/sdX every minute for the chart, without -n standby. Every poll did a full SMART read, which any HDD interprets as activity → it either spun up from standby or never made it down.
• The health poller had the same shape on smartctl --health -j.

Both pollers now use smartctl -n standby ..., so a drive that's already parked exits with code 2 and no I/O reaches the disk. The one-shot smartctl -i for serial/model still runs once at startup and caches the result for the lifetime of the service — that's a single wake-up per boot, not once per minute.

Bonus: when the monitor detects a drive in standby, it now renders a small blue "⏻ Standby" badge in the Storage view in place of the temperature value. So if your SnapRAID disks are parked, the graph won't look broken — it'll explicitly say Standby until the drive wakes up by normal use, at which point the temperature comes back automatically.

smartctl -i on the modal you open by clicking a disk is intentionally unchanged — when you ask for SMART details, waking the drive is the right behaviour.

Available in the next beta, no setting needed (the new behaviour is the default).

Thanks again for the report and the kind words. 🙏