Day111 part3

Summary

Key Points

• ReplicaSets do not support rolling updates; only Deployments provide that capability. 0:15
• Blue‑green and canary deployments are alternative strategies that aim to reduce downtime, but they were only introduced briefly and will be covered later. 3:37
• The video demonstrates the rolling update process with kubectl apply -f deploy.yaml, showing pods being created and deleted sequentially. 8:31
• By default, a Deployment performs a rolling update, replacing pods incrementally instead of deleting all at once. 10:03
• Setting maxUnavailable to 25% means only one pod out of four can be unavailable at any time during the update. 10:26
• Setting maxSurge to 25% means one extra pod can be created above the desired replica count while the old pod is being terminated. 10:36
• Rolling updates can be customized using the maxUnavailable and maxSurge fields to control how many pods are taken down or added during the update. 10:45
• If maxUnavailable and maxSurge are both set to 100%, the Deployment behaves like a recreate strategy, deleting all old pods before creating new ones, causing downtime. 11:09
• The watch command can be used to monitor pod status (Creating, Running, Terminating, etc.) during the update. 16:23
• The overall goal of these strategies is to maintain high availability while deploying new container images. 19:25