How to realize Pod in-place update in Kubernetes?
KubeBlocks version 0.9 introduces in-place updates for improved system availability, leveraging Kubernetes Pod API capabilities. It supports CPU and memory updates for Kubernetes 1.27+, with potential application restarts needed.
Read original article
KubeBlocks has introduced an in-place update feature starting from version 0.9 to enhance system availability during instance updates. Previously, updates to StatefulSets required a complete recreation of pods, which was not ideal for database management. The new feature allows for in-place updates by leveraging the Kubernetes Pod API's capabilities, supporting specific fields such as annotations, labels, and resource requests for CPU and memory. For Kubernetes versions 1.27 and above, additional support for in-place updates of CPU and memory resources is available through the PodInPlaceVerticalScaling feature. KubeBlocks automatically detects the Kubernetes version and feature switches to enable these capabilities. It is important to note that after resizing resources, some applications may need to be restarted to recognize the new configurations, necessitating adjustments to the container restart policy. The upper-level APIs related to instances, including Cluster and ComponentDefinition, also support in-place updates, allowing for a more efficient update process without significant downtime.
- KubeBlocks version 0.9 introduces in-place updates to improve system availability.
- In-place updates leverage Kubernetes Pod API capabilities for specific fields.
- Support for CPU and memory in-place updates is available for Kubernetes 1.27 and above.
- Applications may require restarts after resource resizing to recognize new configurations.
- Upper-level APIs like Cluster and ComponentDefinition facilitate in-place updates.
Related
A skeptic's first contact with Kubernetes
The author shares their journey from skepticism to understanding Kubernetes, highlighting its workload management, key components, and challenges like complexity, scaling, and configuration, while recognizing its architectural value.
Introduction to Kubernetes (k8) – a quick lookback
Kubernetes has grown significantly since 2018, increasing to 109,000 stars and 3,674 contributors by 2024. It offers features like scaling, automated rollouts, and self-healing, encouraging hands-on learning for newcomers.
0 commentsRelated
A skeptic's first contact with Kubernetes
The author shares their journey from skepticism to understanding Kubernetes, highlighting its workload management, key components, and challenges like complexity, scaling, and configuration, while recognizing its architectural value.
Introduction to Kubernetes (k8) – a quick lookback
Kubernetes has grown significantly since 2018, increasing to 109,000 stars and 3,674 contributors by 2024. It offers features like scaling, automated rollouts, and self-healing, encouraging hands-on learning for newcomers.