Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Recuperação Desastres & Backups em Kubernetes

Carol
November 09, 2018

Recuperação Desastres & Backups em Kubernetes

Carol

November 09, 2018
Tweet

More Decks by Carol

Other Decks in Technology

Transcript

  1. Etcd - Replicação do log O líder notifica aos seguidores

    que o valor foi persistido. O cluster chega a um consenso do estado do sistema.
  2. Etcd - Quorum Consenso Concordar para tomar ação Mais do

    50%, (N/2)+1 Perder consenso Managers Quorum 1 1 2 2 3 2 4 3 5 3 6 4 7 4
  3. Etcd - Tolerancia falhas Redundância HA master com 5 nodes

    Managers Quorum Falhas 1 1 0 2 2 0 3 2 1 4 3 1 5 3 2 6 4 2 7 4 3
  4. Etcd - Recuperação desastre Resiliência : (N-1)/2 falhas log snapshots

    : /snap WAL : /wal etcdctl backup \ --data-dir %data_dir% \ [--wal-dir %wal_dir%] \ --backup-dir %backup_data_dir% [--backup-wal-dir %backup_wal_dir%] etcdctl snapshot save
  5. Etcd - Restore etcd \ -data-dir=%backup_data_dir% \ [-wal-dir=%backup_wal_dir%] \ -force-new-cluster

    \ etcdctl snapshot restore Automatizar o cluster …. Figure 4. Stateful. https://www.slideshare.net/mKrishnaKumar1/stateful-set-in-kubernetes-implementation-usecases
  6. Referências Demo: Demo backup using ark: https://github.com/krol3/backup-k8s The Log: What

    every software engineer should know about real-time data's unifying abstraction Understanding Distributed Consensus in etcd and Kubernetes - Laura Frank, CloudBees Disaster Recovery for your Kubernetes Clusters [I] - Andy Goldstein & Steve Kriss, Heptio