The release information is reported second-hand by The New Stack.
According to The New Stack, Kubernetes 1.36 again guarantees reliable database backups. A coordinated group snapshot captures all volumes at the same point in time.
Previously, CSI standardized creating snapshots only for one PersistentVolumeClaim (PVC). For a database with separate data and log volumes, snapshots made at different times can prevent recovery.
Claim check:
- Kubernetes 1.36 again guarantees reliable database backups. (confirmed only by the carrying publication: evidence; «Kubernetes 1.36 restores a lost guarantee for database backups»)
- A coordinated volume group snapshot captures all volumes at the same point in time. (confirmed only by the carrying publication: evidence; «when you snapshotted the group, the array froze them all at the same instant. Every volume captured the same point in time.»)
- CSI standardized snapshots for one PersistentVolumeClaim (PVC) at a time. (confirmed only by the carrying publication: evidence; «The Container Storage Interface (CSI) standardized snapshots around a single object, the VolumeSnapshot, scoped to a single PersistentVolumeClaim (PVC). One PVC, one snapshot.»)
- For a database with separate data and log volumes, snapshots made at different times can prevent recovery. (confirmed only by the carrying publication: evidence; «Then Postgres refuses to start because the WAL on one volume references pages that were never captured in the data files on the other.»)
Publications:
- https://thenewstack.io/kubernetes-volume-group-snapshots
- https://thenewstack.io/mastering-deadman-alerts-to-prevent-silent-failures
- https://thenewstack.io/how-to-maintain-api-consistency-as-you-scale
- https://thenewstack.io/how-open-standards-enable-zero-trust-on-commodity-hardware
score 83.0 out of 100 · kind: release