Huawei Cloud MRS Host Fault Diagnosis Skill

Overview

This skill diagnoses Huawei Cloud MRS (MapReduce Service) cluster faults. Given a service name and/or node name, it progressively localizes the root cause: quick log scan first, host troubleshooting when host issues are found, detailed investigation when no conclusion is reached.

Architecture: Caller (Agent) -> check_api_mode.py (Python, scripts/) determines the API mode -> either lakewatch_api_client.py -> LakeWatch API -> MRS cluster (node resource data, logs, MRS Manager proxy) or manager_api_client.py -> MRS Manager REST API (28443). Per-layer knowledge base (fault_layer/ + scenarios/ + propagation.md in lakewatch mode; fault_layer_manager/ + scenarios_manager/ + propagation_manager.md in manager mode) drives the diagnosis flow; per-component config under components/ is shared by both modes; three fault layers (host -> instance -> service) with propagation chain tracing.

Note on language: This SKILL.md and the documents under references/ are written in English per the repository spec. The knowledge base documents under fault_layer/, fault_layer_manager/, scenarios/, scenarios_manager/, components/, propagation.md, and propagation_manager.md are also in English. Commands and code blocks are English throughout.

Applicable Scenarios:

• A service is reported unhealthy and the root cause must be localized
• An instance is reported faulty on a specific node
• A host is reported unreachable or abnormal
• Progressive fault triage from quick scan to deep investigation