Dell OS10 Hardware Health and Optics Diagnostics

Most "switch is down" tickets turn out to be a power supply reporting a degraded state, a fan that has already failed redundancy, or an optic whose receive power has crept below the specification after a patch-panel change. Dell SmartFabric OS10 exposes all of that through a small set of show commands, but they are spread across different reference pages and are rarely collected together. This article gathers the health commands that matter, explains how to read the optical thresholds properly, and shows the sequence to run when you are handed a switch with an unknown history.

Chassis Health: Environment First

OS10# show environment
------------------ Unit: 1 ------------------
PSU  Status    Type        Fan Status      Fan Speed (rpm)  Airflow
---  ------    ----        --------------- --------------   -------
1    up        AC          up              8426             Front-to-Back
2    up        AC          up              8402             Front-to-Back

Unit  Temperature (Celsius)   Status
----  ----------------------  ------
1      38                      ok
1      47                      ok

Unit  Fan Tray  Status     Fan Speed (rpm)
----  --------  ---------- ----------------
1     1         up         8426

What to look for: a PSU showing up but a fan inside it down (redundancy already lost), a temperature that is "ok" but within a few degrees of the warning threshold, and mismatched airflow directions on a fresh install — front-to-back in a rack wired for back-to-front is a silent thermal failure that takes weeks to show up as random reboots.

OS10# show environment | grep -i -e fail -e down -e critical
OS10# show system
OS10# show inventory
OS10# show version | head -20
OS10# show boot

show inventory is the one to record in your asset register: it lists each FRU's part number and service tag, which is exactly what the replacement order needs. Collect show inventory and show version output into the ticket before you open it, not after the support engineer asks twice.

Optic Diagnostics with show interface transceiver

OS10# show interface ethernet 1/1/1 transceiver
Ethernet 1/1/1
  SFP Type             : SFP+
  Vendor Name          : DELL
  Vendor PN            : FTLX8574D3BCL
  Vendor SN            : FNS12345678
  Vendor Revision      : A
  Wavelength           : 850 nm
  Maximum Distance     : 300 m
  Media Type           : 10GBASE-SR
  Temperature          : 34.2 C
  Voltage              : 3.29 V
  Transmit Bias Current: 6.12 mA
  Transmit Power       : -2.11 dBm
  Receive Power        : -2.34 dBm

OS10# show interface ethernet 1/1/1 transceiver detail    # thresholds and alarm flags
OS10# show interface ethernet 1/1/1 phy                    # physical/lane level detail
OS10# show interface ethernet 1/1/1
OS10# show interfaces transceiver                         # all ports at once
OS10# show interfaces transceiver | grep -i -e alarm -e warn -e fault

How to read it without guessing:

• Receive power (Rx) is the single most useful value. Compare it against the module's specification for the media type — a 10GBASE-SR at -2.3 dBm is healthy; the same reading on a long-reach part may be marginal.
• Transmit bias current far above nominal indicates an ageing laser; a reading at zero means the transmitter is not being driven (usually a port that is administratively down rather than a failed optic).
• Temperature inside the module is always higher than the chassis value; consistent per-port outliers point at blocked airflow in a specific slot.
• One port in a pair reading -20 dBm or lower with a clean Rx on the far end is a dirty or damaged connector, not a dead module. Clean before you replace.

Dell also exposes tunable-wavelength optics on DWDM-capable interfaces, where the wavelength value must match the plan at both ends — a mismatch shows up as a port that links at some distances and not others.

Interface-Level Counters That Point at Layers 1 and 2

OS10# show interface ethernet 1/1/1 counters
OS10# show interface ethernet 1/1/1 counters detailed
OS10# show interfaces ethernet 1/1/1 | grep -i -e error -e drop -e "rate"
OS10# show mac address-table interface ethernet 1/1/1
OS10# show lldp neighbors interface ethernet 1/1/1

The pairing that resolves most disputes: physical errors (CRC, alignment) increase with traffic while FCS errors are zero on the far end — that is a bad patch lead or a dirty connector on the near end. If both ends count errors, suspect the medium or a speed/duplex mismatch rather than a single device.

When to Collect a Support Bundle

OS10# show tech-support
# or the scripted collector for a full bundle
OS10# system-support-collect
OS10# dir /var/log/

Collect it once, early, while the fault is present. Anything collected after a reload has lost the evidence that would have explained the reload.

Routine Health Check Sequence

1. show environment — PSUs, fans, temperatures, airflow direction.
2. show inventory + show version — record FRU IDs and image in the ticket.
3. show interfaces transceiver | grep -i -e alarm -e warn — optics outliers in one pass.
4. show interface ethernet <n> counters for every uplink.
5. show logging filtered for hardware and thermal events since the last reboot.
6. Save the output; the diff against last week's baseline is what actually finds slow-developing faults.

Related reading: optical transceiver DOM/DDM monitoring, Dell OS10 VLT troubleshooting and OS10 SNMPv3 monitoring configuration.

原文链接：https://www.dell.com/support/manuals/en-us/dell-emc-smartfabric-os10/smartfabric-os-user-guide-10-5-1/show-interface-transceiver-tunable-wavelength?guid=guid-ff2c2aa9-ef3d-418d-b4b9-c2b102b90dec&lang=en-us