| 1.17.11 | v1.25 v1.27 v1.28 v1.29 v1.30 v1.31 v1.32 | Supported CCE clusters v1.32. |
| 1.16.17 | v1.25 v1.27 v1.28 v1.29 v1.30 v1.31 | Supported even scheduling in virtual GPUs. |
| 1.15.11 | v1.23 v1.25 v1.27 v1.28 v1.29 v1.30 | Fixed some issues. |
| 1.15.6 | v1.23 v1.25 v1.27 v1.28 v1.29 v1.30 | Resources can be oversubscribed based on pod profiling. |
| 1.13.3 | v1.21 v1.23 v1.25 v1.27 v1.28 v1.29 | - Supported scale-in of custom resources based on node priorities.
- Optimized the association between preemption and node scale-out.
|
| 1.13.1 | v1.21 v1.23 v1.25 v1.27 v1.28 v1.29 | Optimized scheduler memory usage. |
| 1.12.18 | v1.21 v1.23 v1.25 v1.27 v1.28 v1.29 | - Supported CCE clusters v1.29.
- The preemption function is enabled by default.
|
| 1.11.21 | v1.19.16 v1.21 v1.23 v1.25 v1.27 v1.28 | - Supported Kubernetes v1.28.
- Supported load-aware scheduling.
- Changed the image OS to Huawei Cloud EulerOS 2.0.
- Optimized CSI resource preemption.
- Optimized load-aware rescheduling.
- Optimized preemption in hybrid deployment scenarios.
|
| 1.11.6 | v1.19.16 v1.21 v1.23 v1.25 v1.27 | - Supported Kubernetes v1.27.
- Supported rescheduling.
- Supported affinity scheduling of nodes in node pools.
- Optimized the scheduling performance.
|
| 1.10.7 | v1.19.16 v1.21 v1.23 v1.25 | Fixed the issue where the local PV add-on fails to calculate the number of pods pre-bound to nodes. |
| 1.10.5 | v1.19.16 v1.21 v1.23 v1.25 | - The volcano agent supports resource oversubscription.
- Added the verification admission for GPUs. The value of nvidia.com/gpu must be less than 1 or a positive integer, and the value of volcano.sh/gpu-core.percentage must be a multiple of 5 and less than 100.
- Fixed the issue where pod scheduling is slow when PVC binding fails.
- Fixed the issue where newly added pods cannot run when there are terminating pods on a node for a long time.
- Fixed the issue where Volcano restarts during concurrent PVC creation or mounting to pods.
|
| 1.9.1 | v1.19.16 v1.21 v1.23 v1.25 | - Fixed the issue where the counting pipeline pod of the networkresource plugin occupies supplementary network interfaces.
- Fixed the issue where the binpack plugin scores nodes with insufficient resources.
- Fixed the issue of processing resources in the pod with an unknown end status.
- Optimized event output.
- Supported HA deployment by default.
|
| 1.7.2 | v1.19.16 v1.21 v1.23 v1.25 | - Supported clusters v1.25.
- Improved the scheduling performance of Volcano.
|
| 1.7.1 | v1.19.16 v1.21 v1.23 v1.25 | Supported clusters v1.25. |
| 1.4.7 | v1.15 v1.17 v1.19 v1.21 | Deleted the pod status Undetermined to adapt to cluster Autoscaler. |
| 1.4.5 | v1.17 v1.19 v1.21 | Changed the deployment mode of volcano-scheduler from StatefulSet to Deployment, and fixed the issue where pods cannot be automatically migrated when the node is abnormal. |
| 1.4.2 | v1.15 v1.17 v1.19 v1.21 | - Resolved the issue that cross-GPU allocation fails in certain scenarios.
- Supported the updated EAS API.
|
| 1.3.7 | v1.15 v1.17 v1.19 v1.21 | - Supported hybrid deployment of online and offline jobs and resource oversubscription.
- Optimized the scheduling throughput for clusters.
- Fixed the issue where the scheduler panics in certain scenarios.
- Fixed the issue where the volumes.secret verification of the volcano job in the CCE clusters v1.15 fails.
- Fixed the issue where jobs fail to be scheduled when volumes are mounted.
|
| 1.3.3 | v1.15 v1.17 v1.19 v1.21 | Fixed the scheduler crash due to a GPU error and a privileged init container admission failure. |
| 1.3.1 | v1.15 v1.17 v1.19 | - Upgraded the Volcano framework to the latest version.
- Supported Kubernetes v1.19.
- Added the numa-aware plugin.
- Fixed the Deployment scaling issue in the multi-queue scenario.
- Adjusted the algorithm plugin that is enabled by default.
|
| 1.2.5 | v1.15 v1.17 v1.19 | - Fixed the OutOfcpu issue in some scenarios.
- Fixed the issue where pods cannot be scheduled when some capabilities are set for a queue.
- Made the log time of the volcano component consistent with the system time.
- Fixed the issue of preemption between multiple queues.
- Fixed the issue where the ioaware plugin cannot behave as expected in some scenarios.
- Supported hybrid clusters.
|
| 1.2.3 | v1.15 v1.17 v1.19 | - Fixed the training task OOM issue caused by inaccuracy.
- Fixed the GPU scheduling issue in CCE clusters of v1.15 and later versions. Rolling upgrades of CCE versions during task distribution are not supported.
- Fixed the issue where the queue status is unknown in certain scenarios.
- Fixed the issue where a panic occurred when a PVC is mounted to a job in a specific scenario.
- Fixed the issue where decimals cannot be configured for GPU jobs.
- Added the ioaware plugin.
- Added ring controller.
|
