Problem
KEP-5823 introduces native Pod-level checkpoint and restore in Kubernetes, adding CheckpointPod and RestorePod RPCs to the CRI RuntimeService. This shifts checkpointing from per-container states to taking a consistent snapshot of the entire sandbox/Pod boundary.
Because gVisor manages its own Sentry kernel and application state, gVisor's native checkpoint/restore model naturally aligns with Pod-level snapshotting much better than legacy per-container approaches (KEP-2008).
The gVisor community will partner with containerd community and other sandbox runtime for the support.
KEP/SIG-Node References
KEP(s): KEP-5823
stage: alpha
KEP Issue: kubernetes/enhancements#5823
KEP PR: kubernetes/enhancements#5851
K8s-Release: v1.37 (target for Alpha)
KEP-Owner: @rst0git @viktoriaas @adrianreber @andreyvelich
SIG-Node member liason: @haircommander
KEP-Shepherd: TBD
Containerd References
Kata References
Problem
KEP-5823 introduces native Pod-level checkpoint and restore in Kubernetes, adding CheckpointPod and RestorePod RPCs to the CRI RuntimeService. This shifts checkpointing from per-container states to taking a consistent snapshot of the entire sandbox/Pod boundary.
Because gVisor manages its own Sentry kernel and application state, gVisor's native checkpoint/restore model naturally aligns with Pod-level snapshotting much better than legacy per-container approaches (KEP-2008).
The gVisor community will partner with containerd community and other sandbox runtime for the support.
KEP/SIG-Node References
KEP(s): KEP-5823
stage: alpha
KEP Issue: kubernetes/enhancements#5823
KEP PR: kubernetes/enhancements#5851
K8s-Release: v1.37 (target for Alpha)
KEP-Owner: @rst0git @viktoriaas @adrianreber @andreyvelich
SIG-Node member liason: @haircommander
KEP-Shepherd: TBD
Containerd References
Kata References