Greatest Hits Archives - Software Engineering Daily · Greatest Hits Archives - Software Engineering Daily

Stateful Kubernetes with Saad Ali

·54 min·2 clips
The container storage interface moved storage vendors out of Kubernetes core and went GA in 1.13.
1. Software Engineering Daily’s "Stateful Kubernetes with Saad Ali" focuses on storage, persistent volumes, and Kubernetes stateful workloads, with Google and SIG Storage as the main frame. 2. Saad Ali is introduced as a senior engineer at Google and a member of the Kubernetes Storage SIG, which makes him the episode’s technical guide on Kubernetes storage. 3. The core thesis is that Kubernetes handles stateless containers well, but "stateful versus stateless workloads" still require PVCs, PVs, storage classes, and higher-level patterns. 4. Saad uses shopping carts, profile data, and container rescheduling to show how state disappears when a container is terminated or moved to another node. 5. He says container-attached storage is mostly about file and block storage, because POSIX and SCSI standardize the data path while Kubernetes controls provisioning, attachment, and mounting. 6. He also contrasts databases and object storage, noting that databases may write directly to block or file, while object stores lack a standardized data path across GCP, Amazon S3, and on-prem systems. 7. The conversation explains how application developers use persistent volume claims while cluster administrators manage persistent volumes and storage classes, especially through dynamic provisioning. 8. Saad describes the shift from built-in volume plugins to CSI, saying the old model created testing, crash, and privilege problems inside Kubernetes core binaries. 9. He adds that CSI 1.0 and Kubernetes 1.13 let storage vendors ship drivers independently, while operators and CRDs now handle upgrades, scaling, and other "day two operations." 10. Best for Kubernetes engineers, platform teams, and storage operators.

As heard by us

A measured guide to how Kubernetes helps applications request persistent storage across on-prem and hybrid environments.

Saad Ali frames storage as one of cloud infrastructure's stubborn problems: servers, networks, and software can all fail, but applications still need dependable places to keep state.

Read the full review in PlayNext →

Why you'd press play

Want Kubernetes to keep your app's state somewhere sturdier when servers, networks, or software have a bad day?

Read the full recommendation in PlayNext →
Listen to the show on