[BUG] NVMe does not show capacity when using namespaces

[KyleSanderson]

Describe the bug nvme0

Expected behavior nvme0n

Screenshots

Log Files If related to missing devices or SMART data, please run the collector in DEBUG mode, and attach the log file. See /docs/TROUBLESHOOTING_DEVICE_COLLECTOR.md for other troubleshooting tips.

docker run -it --rm -p 8080:8080 \
-v `pwd`/config:/opt/scrutiny/config \
-v /run/udev:/run/udev:ro \
--cap-add SYS_RAWIO \
--device=/dev/sda \
--device=/dev/sdb \
-e DEBUG=true \
-e COLLECTOR_LOG_FILE=/opt/scrutiny/config/collector.log \
-e SCRUTINY_LOG_FILE=/opt/scrutiny/config/web.log \
--name scrutiny \
ghcr.io/analogj/scrutiny:master-omnibus

# in another terminal trigger the collector
docker exec scrutiny scrutiny-collector-metrics run

The log files will be available on your host in the config directory. Please attach them to this issue.

Please also provide the output of docker info

[AnalogJ]

sorry, I'm not sure I understand what you're proposing here:

Describe the bug
nvme0

Expected behavior
nvme0n

I do see that the detected capacity is 0b, can you attach the debug logs so I can understand if the issue is with Scrutiny or Smartctl?

[Tirarex]

Has same problem with SAMSUNG MZQL21T9HCJR-00A07 (1.7tb model). No errors in log or after "collector_1 | starting cron". Other sata ssd and hdd works fine.

[KyleSanderson]

Has same problem with SAMSUNG MZQL21T9HCJR-00A07 (1.7tb model). No errors in log or after "collector_1 | starting cron". Other sata ssd and hdd works fine.

Yeah, it's pretty trivial to see. NVMe uses namespaces, which the software doesn't support.

[Tirarex]

Yeah, it's pretty trivial to see. NVMe uses namespaces, which the software doesn't support.

it works fine with my other ssd's - samsung 980 pro / adata legend s70 / adata legend 960 / something from IRDM but nvme.

Looks like only enterprise grade ssd's are problem

[KyleSanderson]

Yeah, it's pretty trivial to see. NVMe uses namespaces, which the software doesn't support.

it works fine with my other ssd's - samsung 980 pro / adata legend s70 / adata legend 960 / something from IRDM but nvme.

Looks like only enterprise grade ssd's are problem

It depends on if namespaces are exposed. Some bad controllers don't follow the spec, others expose 1 as 0, I believe the limit is 255 per device.

[Hr46ph]

Does this require any more info, like debug logs or smart output? I have a stack of INTEL SSDPE2KX040T8 that aren't reporting their size in Scrutiny. Smart output reports it twice 😋 .

[uhthomas]

I also see this on a KIOXIA CD6.

[AnalogJ]

@Hr46ph @uhthomas @KyleSanderson

If you can provide me with the raw smartctl json, I can take a look at why Scrutiny is unable to extract your capacity information:

smartctl --xall --json {DEVICE}

[uhthomas]

@Hr46ph @uhthomas @KyleSanderson

If you can provide me with the raw smartctl json, I can take a look at why Scrutiny is unable to extract your capacity information:

smartctl --xall --json {DEVICE}

❯ k exec -it scrutiny-collector-94lbl -- smartctl --xall --json /dev/nvme4

{
  "json_format_version": [
    1,
    0
  ],
  "smartctl": {
    "version": [
      7,
      2
    ],
    "svn_revision": "5155",
    "platform_info": "x86_64-linux-6.1.69-talos",
    "build_info": "(local build)",
    "argv": [
      "smartctl",
      "--xall",
      "--json",
      "/dev/nvme4"
    ],
    "exit_status": 0
  },
  "device": {
    "name": "/dev/nvme4",
    "info_name": "/dev/nvme4",
    "type": "nvme",
    "protocol": "NVMe"
  },
  "model_name": "KCD61LUL3T84",
  "serial_number": "61Q0A05UT7B8",
  "firmware_version": "8002",
  "nvme_pci_vendor": {
    "id": 7695,
    "subsystem_id": 7695
  },
  "nvme_ieee_oui_identifier": 9233294,
  "nvme_total_capacity": 3840755982336,
  "nvme_unallocated_capacity": 0,
  "nvme_controller_id": 1,
  "nvme_version": {
    "string": "1.4",
    "value": 66560
  },
  "nvme_number_of_namespaces": 16,
  "local_time": {
    "time_t": 1706043228,
    "asctime": "Tue Jan 23 20:53:48 2024 UTC"
  },
  "smart_status": {
    "passed": true,
    "nvme": {
      "value": 0
    }
  },
  "nvme_smart_health_information_log": {
    "critical_warning": 0,
    "temperature": 57,
    "available_spare": 100,
    "available_spare_threshold": 44,
    "percentage_used": 0,
    "data_units_read": 10648491,
    "data_units_written": 11648374,
    "host_reads": 89057132,
    "host_writes": 277683503,
    "controller_busy_time": 175,
    "power_cycles": 33,
    "power_on_hours": 2040,
    "unsafe_shutdowns": 0,
    "media_errors": 0,
    "num_err_log_entries": 3546,
    "warning_temp_time": 0,
    "critical_comp_time": 0
  },
  "temperature": {
    "current": 57
  },
  "power_cycle_count": 33,
  "power_on_time": {
    "hours": 2040
  }
}

[uhthomas]

If you don't mind @AnalogJ, I'd like to take a try at fixing this!

[AnalogJ]

@uhthomas sure! Looking forward to the PR :)