[BUG] SCSI SAS device temperature

[vector-co-uz]

Describe the bug

I'm using Scrutiny Collector on Unraid as docker container. In Scrutiny Web UI my sas device temperature not functional. My Unraid dashboard shows the temperature, but on Scratch Web it shows infinity.

Expected behavior

---

Screenshots

Log Files

Environment

No response

[Starosdev]

Hi, I investigated this against current master and develop, and Scrutiny already includes the known SAS/SCSI fallback for drives where temperature.current is 0 and the real value is under scsi_environmental_reports.temperature_1.current.

To narrow down what your device is reporting, please share the raw SMART JSON for the affected disk. On the host running the collector, run:

smartctl --xall --json /dev/sdX

If the disk is behind a controller or needs a device type, use the same arguments Scrutiny uses for that device, for example:

smartctl --xall --json -d sat /dev/sdX
smartctl --xall --json -d scsi /dev/sdX
smartctl --xall --json -d megaraid,N /dev/bus/0

Please include:

• The full smartctl --xall --json output for the affected SAS/SCSI disk
• Your Scrutiny version/image tag
• Whether the incorrect value appears in the device details page, the dashboard/history graph, or both
• If possible, the API response for the affected device from:

/api/device/<device-id>/details
/api/summary
/api/summary/temp

You can redact serial numbers and WWNs if needed, but please keep all temperature-related sections intact. That will let us determine whether the issue is in SMART parsing, stored history, or frontend rendering.

[Starosdev]

I rechecked this against current master and develop, and the known SAS/SCSI fallback is already present in both branches for devices where temperature.current is 0 and the real value is under scsi_environmental_reports.temperature_1.current.

At this point, the next step is to confirm what your device and API are actually returning.

Please share:

• The full raw SMART JSON for the affected disk from the host running the collector:

smartctl --xall --json /dev/sdX

If the disk is behind a controller or needs an explicit device type, please use the same arguments Scrutiny uses for that device, for example:

smartctl --xall --json -d sat /dev/sdX
smartctl --xall --json -d scsi /dev/sdX
smartctl --xall --json -d megaraid,N /dev/bus/0

• Your Scrutiny version or image tag
• Whether the incorrect infinity value appears in the device details page, the dashboard/history graph, or both
• If possible, the API responses for the affected device from:

/api/device/<device-id>/details
/api/summary
/api/summary/temp

You can redact serial numbers and WWNs if needed, but please keep all temperature-related sections intact. That will let us determine whether this is a SMART parsing gap, stale stored temperature history, or something else upstream of the frontend formatter.

[Starosdev]

A second pass against current master and develop confirms the known SAS/SCSI temperature path is already covered. Current code reads SCSI/SAS temperatures from temperature.current, and when that value is 0, falls back to scsi_environmental_reports.temperature_1.current. That path is covered by regression tests and is present in v1.62.2 and current branch heads.

The screenshot shows the dashboard card receiving/rendering a non-finite value (∞°C), but the current backend summary API stores temperature as an integer, so we still need the actual payload from the affected system to identify where the bad value is entering.

Please update to the latest available image, then share the following for /dev/sdd:

smartctl --xall --json /dev/sdd

If the disk requires an explicit device type, please use the same arguments Scrutiny uses for that device, for example:

smartctl --xall --json -d scsi /dev/sdd
smartctl --xall --json -d sat /dev/sdd
smartctl --xall --json -d megaraid,N /dev/bus/0

Also include:

/api/summary
/api/device/<device-id>/details
/api/summary/temp

Please include the Scrutiny image tag/version as well. Serial numbers and WWNs can be redacted, but keep all temperature-related fields intact.

[Starosdev]

The known SCSI/SAS fallback is present in current develop and covered by regression tests, but the reported infinity value is not yet explained. Please update to a current image and attach:\n\n- Full smartctl JSON for the affected disk, using the same device arguments the collector uses\n- API responses from /api/summary, /api/device//details, and /api/summary/temp\n- Scrutiny image tag or version\n\nKeep all temperature-related fields intact; serial numbers and WWNs may be redacted. This will show whether the non-finite value enters during SMART parsing, storage, or frontend rendering.