Straight from chapter 3 of Huawei's CloudEngine 16800, 9800, 8800 and 6800 series maintenance manual — the routine check for data-center core and spine switches: environment, basic information, running status and interface content. What to look at in each step, what counts as a pass, the recommended cadence, and the exact display commands used to check it.
By the AtlasCommTech engineering team — 13 years of carrier & enterprise network deployments · Updated July 2026
Chapter 3 of Huawei's CloudEngine 16800/9800/8800/6800 series maintenance manual organizes routine maintenance into four sections, not five — and the chassis behind it is a different animal entirely.
Our enterprise switch routine health check note follows Huawei's Sx3-series manual for campus and access-layer switches — wiring closets, office floors, one switch failing affects one segment. This note follows the manual for the CloudEngine 16800, 9800, 8800 and 6800 series instead: core and spine switches sitting in a data center, where one bad reading behind a core box can affect every rack downstream of it, not just one wing of a building.
Chapter 3, '例行维护' (Routine Maintenance), organizes everything into four sections: device environment check, device basic information check, device running check, and interface content check. Two of the four chassis families — CE16800 and CE8800/6800 — even get their own separate environment tables, because their rated voltage and humidity tolerances genuinely differ. What follows is that four-step routine, the pass standard and display command for every item, a DC-appropriate cadence, and a note on configuration backup discipline that the check items alone don't cover.
Same order the source manual uses — environment first, interfaces last.
Diagram labels are kept in English for engineering clarity.
The source manual gives the CE16800 series and the CE8800/6800 series two separate tables — their voltage and humidity tolerances genuinely differ, and this step has no CLI commands, only physical checks.
| Check Item | Pass Standard |
|---|---|
| Equipment room temperature | Operating temperature 0°C to 40°C (0m to 1800m altitude); above 1800m to 5000m, the rating drops 1°C for every 220m of additional altitude. Storage temperature -40°C to +70°C. |
| Equipment room humidity | Operating relative humidity 5%RH to 85%RH, non-condensing. If room humidity runs consistently too high, install dehumidification; if too low, install humidification. |
| Cleanliness | The device itself should show no visible dust buildup; watch the dust-screen condition and clean or replace the dust cotton promptly so cabinet ventilation and cooling aren't affected. |
| Cooling / airflow | Fans must run normally during operation (except while being cleaned); the chassis air intake must not be blocked by excess dust — clean the dust screen if heavily soiled, recommended quarterly. |
| Cable routing | Power cables routed separately from business cables; both power and business cabling neat and orderly. |
| Cable labeling | Cable labels clear, accurate and standards-compliant. |
| Power supply system | Stable supply required. Rated voltage 110V AC / 220V AC; rated range 100V AC-120V AC / 200V AC-240V AC; maximum range 90V AC-290V AC. |
| Check Item | Pass Standard |
|---|---|
| Device placement | Placed in a ventilated, dry environment, sitting firm and level. |
| Equipment room temperature | Same as CE16800: operating 0°C to 40°C (0m-1800m, -1°C per 220m above 1800m to 5000m); storage -40°C to +70°C. |
| Equipment room humidity | Operating relative humidity 5%RH to 95%RH, non-condensing — a wider tolerance than the CE16800's 85%RH ceiling. |
| Room air conditioning | Air conditioning runs continuously and stably, keeping room temperature within the device's tolerance. |
| Grounding and ground resistance | Working ground, protective ground and building lightning ground kept separate where possible (a shared ground is acceptable if room conditions require it); outdoor-installed equipment must be grounded — an ungrounded outdoor unit is highly vulnerable to lightning damage. |
| Power connection | Power cable correctly connected to the device's designated position and firmly seated; the power indicator stays solid green. |
| Power supply system | Stable supply required. Rated range 100V AC-240V AC, 50/60Hz; maximum range 90V AC-290V AC, 45Hz-65Hz — a wider rated band than the CE16800's fixed 110V/220V. |
| Cooling airflow obstruction | No clutter around the device; recommended cooling clearance between two units is at least 1U. |
| Cooling alignment | The power module's and fan module's cooling airflow direction must be consistent with each other, and with the cabinet's own cooling airflow direction. |
| Fan module status | Fan module indicator normal when slow-blinking green. |
| Device running status | Device SYS indicator normal when slow-blinking green. |
Checks whether basic information — software version, patch information, system time, configuration, license — is correct. We recommend monthly, tightened to align with any change-control cycle your data center already runs.
| Check Item | Check Method | Pass Standard |
|---|---|---|
| Software version running | display version | The sub-card's PCB version number and software version number match what is required. |
| Software package in use | display startup | The file names of the product software and configuration file currently in use, and to be loaded at next startup, are correct. |
| Patch information | display patch-information | The patch file matches actual requirements — Huawei recommends loading the latest patch published for the running product version. |
| System time | display clock | Time matches local actual time (difference no greater than 5 minutes), so faults can be pinpointed precisely by timestamp. If not, run clock datetime or configure NTP to synchronize network time. |
| Configuration correctness | display current-configuration | Review the currently effective configuration parameters to verify device configuration is correct. |
| License information | display license | Review the activated License file's details — name, version, validity period, control items — to determine whether the version needs upgrading to support more features. |
<HUAWEI> display version
<HUAWEI> display startup
<HUAWEI> display patch-information
<HUAWEI> display clock
<HUAWEI> display current-configuration
<HUAWEI> display licenseThis chapter's own basic-info check only verifies that the running configuration is currently correct — it says nothing about whether that configuration would survive a reboot, a board reset, or a power event. Once display current-configuration passes, run save to persist it to the startup configuration, and keep a scheduled off-box copy as well — an FTP/SFTP export or a configuration-management pull kept somewhere that survives the device itself being unavailable. save alone protects against the next reload; it doesn't protect against the device, its storage, or the operator being the actual point of failure. Some remediation steps that come out of a failed check — a reboot, a board reset, a delete on a full storage device — carry real risk of their own; see our dangerous commands checklist before running any of them.
Checks board status, fan and power status, temperature, CPU and memory — the items with the most immediate blast radius on a core switch. We recommend daily.
| Check Item | Check Method | Pass Standard |
|---|---|---|
| Board running status (CE16800 only) | display device | Focus on board presence and status: normal means board 'Online' = 'Present', 'Power' = 'PowerOn', 'Register' = 'Registered', 'Status' = 'Normal'. |
| Fan status | display device fan | 'Status' = 'Normal' indicates normal. |
| Power status | display device power | 'State' = 'Supply' indicates normal. |
| Device temperature | display device temperature | 'Status' = 'Normal' indicates normal. |
| CPU status | display cpu | 'Status' = 'Non-overload' indicates normal. |
| Memory usage | display memory | 'Status' = 'Non-overload' indicates normal. |
<HUAWEI> display device
<HUAWEI> display device fan
<HUAWEI> display device power
<HUAWEI> display device temperature
<HUAWEI> display cpu
<HUAWEI> display memoryChecks port error packets, negotiation mode and Up/Down status. We recommend weekly.
| Check Item | Check Method | Pass Standard |
|---|---|---|
| Port error packets | display interface | While business is running, check for port errors, including CRC errors. |
| Port negotiation mode | display interface | Negotiation mode correct and consistent on both ends — half-duplex mode is not acceptable. |
| Port status | display interface brief | Port Up/Down status meets the planned requirement. |
<HUAWEI> display interface 100GE1/0/1
<HUAWEI> display interface brief'Check for CRC errors' is a one-line pointer, not the full story — see our packet loss troubleshooting note for the complete breakdown of inbound CRC/Giants/Runts errors, optical-module bit errors and outbound Discard congestion, with the exact fields and remediation for each.
Every command above is read-only — display only — but a DC core box has a few ways of reading the result wrong that a campus switch doesn't.
RISKdisplay cpu and display memory report a Status of Non-overload rather than a numeric percentage the way the campus Sx3-series manual's 80%/60% thresholds do. Treating 'Non-overload' as the whole picture misses a slow climb toward overload that hasn't crossed the line yet.
SAFER PRACTICEPair the Non-overload status with the numeric detail available inside the fuller display cpu/display memory output if you want to trend it over time, and check it more often than monthly given how many services sit behind a DC core box.
RISKThe board-presence running-status item scopes explicitly to CE16800-series only. On CE8800/6800 its absence is correct — but on a mixed fleet it's easy for a technician to genuinely lose track of whether a missing item was a deliberate model-based exclusion or something skipped by mistake.
SAFER PRACTICERecord which chassis model each check ran against, not just pass/fail, so a mixed CE16800 + CE8800/6800 estate carries an audit trail explaining every item marked not applicable.
RISKCE16800 environment tolerance allows 5%RH-85%RH and a rated 110V/220V supply; CE8800/6800 allows up to 5%RH-95%RH and a wider 100V-240V rated range. Copying the tighter CE16800 humidity band onto an 8800/6800 chassis, or the reverse, produces a false fail or a false pass depending on which direction the mistake runs.
SAFER PRACTICEConfirm the chassis family before applying either environment table, and keep the two as separate checklists rather than merging them into one generic switch-environment standard.
RISKThis item only verifies the running configuration is currently correct — it says nothing about whether it has actually been saved to the startup configuration a reload will load next. Unlike the campus Sx3-series routine, this manual's basic-info check doesn't carry a separate compare configuration step to catch the gap.
SAFER PRACTICETreat save — and an off-box configuration backup — as mandatory every time this check passes, not an optional extra; see the configuration backup discipline note above, and our campus switch health-check note for the equivalent compare configuration habit.
RISKThe interface content check's port error-packet item flags CRC errors generically, but doesn't distinguish an inbound physical-layer problem from an outbound congestion problem, or check the optical module behind the interface at all — a routine pass here can still be sitting on a climbing Discard counter or a module close to its power threshold.
SAFER PRACTICEWhenever this item flags anything beyond a clean zero, follow through with the full three-source breakdown — optics, inbound errors, outbound Discard — in our packet loss troubleshooting note, rather than treating a CRC mention as the complete diagnosis.
Pulled straight from the field — the ones worth having an answer ready for.
Mostly, but not entirely — chapter 3 gives each chassis family its own environment-check table, since voltage and humidity tolerances differ, and the board-presence running-status item is CE16800-only. Basic-info, CPU/memory and interface checks apply the same way across all four series.
The source manual's chapter 3 doesn't assign a cycle per item the way the campus Sx3-series manual does — see our honest limits note below. Given how much traffic and how many services sit behind a DC core switch, we recommend environment and running-status checks daily (or continuously via monitoring), basic-info monthly, and interface content weekly — tightening any of these where your own change-control or compliance requirements call for more frequent evidence.
That note follows Huawei's Sx3-series manual, with items and cycles specific to campus and access-layer switches. This note follows the CloudEngine 16800/9800/8800/6800 series manual for data-center core and spine switches — different chassis families, different environmental tolerances, and — notably — no built-in per-item cycle, which we address with a DC-appropriate cadence instead.
No — it's a one-line pointer to a much longer diagnosis. See our packet loss troubleshooting note for the full breakdown of inbound CRC/Giants/Runts errors, optical-module bit errors and outbound Discard congestion, with the exact display interface fields and remediation for each.
save only protects the running configuration against being lost on the next reload — it doesn't protect against the device itself failing, a corrupted storage module, or a mistaken overwrite nobody notices until well after the fact. Pair save with a scheduled off-box copy — an FTP/SFTP export, or a configuration-management pull — kept somewhere that survives the device itself being unavailable.
This routine is built from chapter 3 of Huawei's CloudEngine 16800/9800/8800/6800 series maintenance manual. Unlike the Sx3-series manual behind our campus switch note, this V300 chapter doesn't assign a recommended cycle per check item — the daily/monthly/weekly cadence above is our own operational-practice recommendation for a data-center core role, not a figure the manual itself states. The chapter also doesn't include a separate compare-configuration step, which is why we've added the configuration backup discipline note above as a deliberate supplement. This is a health check, not a fault-diagnosis procedure — for packet loss specifically, see our packet loss troubleshooting note.
Tell us your CloudEngine models and how many devices are in scope — we'll help you turn this into a recurring checklist with a cadence that matches your change-control requirements.