Надежность и отказоустойчивость ИТ-инфраструктуры

Уровень 0 · материалов: 5

В кластер входят документы, посвященные методам обеспечения стабильной работы и надежности серверных и сетевых систем, и не входят документы, не касающиеся эксплуатации ИТ-инфраструктуры.

Общие признаки: обеспечение аптайма, управление сетевой инфраструктурой, стратегии предотвращения сбоев, системная устойчивость

Группа выше: Отказоустойчивость и балансировка нагрузки

Материалы

• #23981 — Digital Entropy and the Danger of Long Server Uptime — Представитель этого кластера

  Цифровая энтропия: почему ваш сервер с аптаймом в 1000 дней — это ходячий мертвец, и как с этим жить — AdminFuture

  Смысл: The main idea is that long server uptime is a dangerous illusion of stability; true reliability comes from the ability to reboot or replace systems at any time without failure, achieved through automation and immutable infrastructure.

  High server uptime is a liability, not an achievement, because it hides underlying decay and prevents critical updates, necessitating a shift toward immutable infrastructure.

• #7647 — NetWare 3.12 Server with 16.5 Years Uptime

  Сервер на NetWare 3.12 с аптаймом 16,5 лет — atomlib

  Смысл: The main idea is to showcase an extreme example of system reliability, where a NetWare 3.12 server maintained an uptime of 16.5 years, eventually stopped only due to physical hardware noise and wear.

  A NetWare 3.12 server ran continuously for 16.5 years from 1996 until it was finally shut down due to excessive mechanical noise from its aging hard drives.

• #27340 — Chaos Monkey: Improving Uptime through Intentional Failure

  Злая обезьянка повышает аптайм — mifa

  Смысл: The main idea is that intentionally introducing failure into a system (Chaos Engineering) forces the creation of a more robust, redundant, and reliable infrastructure, ultimately increasing overall uptime.

  Netflix uses a tool called Chaos Monkey to randomly shut down servers, forcing their engineers to build a system that can survive any single point of failure.

• #20807 — Strategies for Systematic Network Equipment Upgrades

  Обновление сетевого оборудования — Insspb

  Смысл: The main idea is that network infrastructure requires proactive, cyclical maintenance and upgrading based on security, reliability, and scalability to prevent catastrophic business failures and enable technological growth.

  System administrators should move from reactive 'firefighting' to a planned, cyclical network upgrade strategy to ensure security, stability, and business scalability.

• #13843 — Challenges and Dependencies of MT_FREE Public Wi-Fi

  Будни MT_FREE: несколько историй про влияние сторонних сервисов на работу публичного Wi-Fi — iVolynkin

  Смысл: The main idea is that maintaining a massive public Wi-Fi network involves managing complex, often invisible dependencies on third-party software, OS updates, and external infrastructure, requiring proactive monitoring and systemic resilience.

  MT_FREE explores how external factors like OS updates, partner outages, and cellular failures impact the stability and traffic of Moscow Metro's public Wi-Fi.