Hotfixes for 2.16
The following is a list of the known and fixed issues for Run:ai V2.16.
Version 2.16.65¶
Internal ID | Description |
---|---|
RUN-21448 | Fixed an issue with degraded workload so the condition would reflect the actual state. |
RUN-20680 | Fixed an issue where the workload page did not present the requested GPU. |
Version 2.16.57¶
Internal ID | Description |
---|---|
RUN-20388 | Fixed an issue where cluster-sync caused a memory leak. |
Version 2.16.25¶
Internal ID | Description |
---|---|
RUN-17241 | Fixed an issue where the nodes page showed nodes as not ready due to "tookit not installed". |
Version 2.16.21¶
Internal ID | Description |
---|---|
RUN-16463 | Fixed an issue after a cluster upgrade to v2.16, where some metrics of pre-existing workloads were displayed incorrectly in the Overview Dashboard. |
Version 2.16.18¶
Internal ID | Description |
---|---|
RUN-16486 | Fixed an issue in the Workloads creation form where the GPU fields of the compute resource tiles were showing no data. |
Version 2.16.16¶
Internal ID | Description |
---|---|
RUN-16340 | Fixed an issue in the Workloads table where filters were not saved correctly. |
Version 2.16.15¶
Release content¶
- Implemented a new Workloads API to support the Workloads feature.
Fixed issues¶
Internal ID | Description |
---|---|
RUN-16070 | Fixed an issue where missing metrics caused the Nodepools table to appear empty. |
Version 2.16.14¶
Release content¶
*Improved overall performance by slowing down metrics updates from 10 seconds to 30 seconds.
Fixed issues¶
Internal ID | Description |
---|---|
RUN-16255 | Fixed an issue in the Analytics dashboard where the GPU Allocation per Node and GPU Memory Allocation per Node panels were displaying incorrect data. |
RUN-16035 | Fixed an issue in the Workloads table where completed pods continue to be counted in the requested resources column. |
Version 2.16.12¶
Fixed issues¶
Internal ID | Description |
---|---|
RUN-16110 | Fixed an issue where creating a training workload (single or multi-node) with a new PVC or Volume, resulted in the Workloads table showing the workload in the Unknown/Pending status. |
RUN-16086 | Fixed an issue in airgapped environments where incorrect installation commands were shown when upgrading to V2.15. |
Version 2.16.11¶
N/A
Version 2.16.9¶
N/A
Version 2.16.8¶
Release content¶
N/A
Version 2.16.7¶
Release content¶
- Added an API endpoint that retrieves data from a workloads's pod.
Fixed issues¶
N/A
Version 2.16.6¶
N/A