没有拿到可用节点，是什么原因

[lanwood]

BASE_IMAGE="registry.cn-hangzhou.aliyuncs.com/hfai/hai-platform:latest"

hai-cli nodes 没有显示节点，scheduler.log 中找到 报错信息

scheduler 0.log:133:2023-12-26 21:25:41.014ERRORticks beater#1703597141000没有拿到可用节点，请检查

[lanwood]

hai-cli images load 自定义镜像 报错

AttributeError: 'UserImage' object has no attribute 'async load!

[lanwood]

hai-cli nodes 报错

scheduler 0.log:133:2023-12-26 21:25:41.014ERRORticks beater#1703597141000没有拿到可用节点，请检查

core.tml 中的 launcher.task_namespace 需改为 'hai-platform', 还有 manager_nodes 也需要调整

[lanwood]

自定义训练镜像满足 validate_image.sh 但无法使用

Volumes:
073c58e81e0042f88595e717c4ebf61d:
Type: HostPath (bare host directory volume)
Path: /data/mnt/hai-platform/workspace/haiadmin
HostPathType: Directory
marsv2-scripts-4:
Type: ConfigMap (a volume populated by a ConfigMap)
Name: marsv2-scripts-4
Optional: false
marsv2-entrypoints-4:
Type: ConfigMap (a volume populated by a ConfigMap)
Name: marsv2-entrypoints-4
Optional: false
start-scripts:
Type: ConfigMap (a volume populated by a ConfigMap)
Name: start-scripts-4-0
Optional: false
log-dir:
Type: HostPath (bare host directory volume)
Path: /data/mnt/hai-platform/workspace/log/haiadmin
HostPathType: DirectoryOrCreate
shm:
Type: EmptyDir (a temporary directory that shares a pod's lifetime)
Medium: Memory
SizeLimit:
QoS Class: Burstable
Node-Selectors: kubernetes.io/hostname=k8s-worker-node1
Tolerations: node.kubernetes.io/memory-pressure:NoExecute op=Exists
node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
Type Reason Age From Message

---

Normal Scheduled 5s default-scheduler Successfully assigned hai-platform/haiadmin-4-0 to k8s-worker-node1
Normal Pulled 4s kubelet Container image "python:3.10-slim" already present on machine
Normal Created 4s kubelet Created container hf-experiment
Normal Started 4s kubelet Started container hf-experiment
Warning FailedMount 1s (x3 over 3s) kubelet MountVolume.SetUp failed for volume "marsv2-entrypoints-4" : object "hai-platform"/"marsv2-entrypoints-4" not registered
Warning FailedMount 1s (x3 over 3s) kubelet MountVolume.SetUp failed for volume "marsv2-scripts-4" : object "hai-platform"/"marsv2-scripts-4" not registered
Warning FailedMount 1s (x3 over 3s) kubelet MountVolume.SetUp failed for volume "start-scripts" : object "hai-platform"/"start-scripts-4-0" not registered

[lanwood]

创建任务后如何配置不拉取新的训练镜像，使用本地已有的镜像

[yolunghiu]

hai-cli nodes 报错

scheduler 0.log:133:2023-12-26 21:25:41.014ERRORticks beater#1703597141000没有拿到可用节点，请检查

core.tml 中的 launcher.task_namespace 需改为 'hai-platform', 还有 manager_nodes 也需要调整

请问您这个问题解决了吗？貌似override.toml中的配置没有生效，改了core.toml重新构建了个镜像，还是获取不到node节点，scheduler_0.log报错：2024-04-12 16:54:36.031 | ERROR | ticks_beater#1712912076000 | 没有拿到可用节点，请检查