Kubernetes Podcast from Google: Leader Election, with Mike Danese

[castrojo]

Kubernetes makes it easy to run distributed workloads, but how do you make sure that replicas don’t conflict with one another? You elect one as the leader. Mike Danese, chair and TL of Kubernetes SIG Auth, joins a vegan and a carnivore to explain how Kubernetes implements leader election.

Do you have something cool to share? Some questions? Let us know:

• web: kubernetespodcast.com
• mail: kubernetespodcast@google.com
• twitter: @kubernetespod

Chatter of the week

• What is a staycation?
• What is steak?
  • Beefsteak, vegetarian/vegan restaurant
• Nachos

News of the week

• Chaos Mesh 1.0
• Azure news:
  • AKS comes to Azure Stack HCI (Preview)
  • AKS adds stopping/restarting clusters, Kubernetes 1.19, confidential compute nodes (Preview)
  • Bridge to Kubernetes is GA
• Istio Steering Committee election results
• OpenServiceMesh joins the CNCF Sandbox
• Odo 2.0.0 GA
  • Odo from Deep Space 9
• Determined AI on Kubernetes
• Cloud Run for Anthos adds events
• KubeAcademy Pro from VMware
• KubeCon EU 2020 transparency report
• Scholarships for KubeCon NA 2020 are open for application

Links from the interview

• Wet labs and dry labs
• Threads
  • What is the difference between processes and threads?
• Mutex or lock
  • What is a mutex?
• Critical section
• Compare-and-swap
• Gas station bathroom keys
• Futex
• Lock server:
  • Chubby
  • etcd
• Optimistic concurrency
• Resource versions
• Regional clusters in GKE
• Leader election
• Leader election client in Kubernetes’ client-go
  • An example of using it by Carlos Becker
  • The new Lease API
• Paxos and Raft
• Deadlock
• Split brain
• Mike Danese on Twitter and GitHub

Original Source: https://kubernetespodcast.com/episode/123-leader-election/