LEARN. PRACTICE. GROW.Concept and mechanism
A network request crosses resolution, routing, transport, and application layers. Getent hosts observes the hosts database through NSS; a direct DNS query can ignore other system sources such as local hosts. After resolving an address, ip route get observes the routing decision for a destination without deleting routes. On the server, ss shows sockets and listeners. A 127.0.0.1 listener accepts local access but does not demonstrate exposure on a remote interface. Do not remove every firewall merely because a client failed. First identify the layer, compare intended and observed configuration, and retain test time, source, and destination.
Guided application
Continuity also needs clear criteria. RPO describes the data recovery point; RTO defines the system recovery-time objective. In a fictional example, accepting fifteen minutes of loss and two hours of recovery are different commitments. Backups, replication, and high availability are not synonyms. Replication can copy accidental deletion; recoverable versions or backups can return to an earlier state if retention and procedures support it. Restore needs access, dependencies, validation, and observed timing. The PM should confirm who executes and who accepts the result. An existing backup file and a running host do not alone demonstrate application recovery within the deadline.
Correct name resolution and a wrong route call for routing investigation rather than blind DNS changes.
Common pitfalls
Ping as application proof; localhost as remote exposure; replica as history; reversed RPO and RTO.
Related topics: Linux, shell, and permissions · Services, logs, and capacity · Cloud, responsibility, and cost
Locate the failure and measure recovery through to the required service.
Reference: Linux routing · LFCA domains and competencies updated2025-09-16; current page confirmed2026-09-30