search

AliyunContainerService / gpushare-scheduler-extender

GPU Sharing Scheduler for Kubernetes Cluster
Apache License 2.0
1.39k stars 308 forks source link

I CAN'T GET GPU INFORMATION. The pod gpushare-device-plugin-ds is pending. #148

Open DoubleChen-cc opened 3 years ago

DoubleChen-cc commented 3 years ago

I cant understand the symbol "~=" in Prerequisites. Is it "not equal" or "approximate"? I use the nividia driver 450.80.02. I deployed it on my master node,then I found that it could not get GPU information.The pod gpushare-device-plugin-ds is pending. image

image image Is there anyone knows the reason?

（我不明白Prerequisites里的”~=“是什么意思，是不等于还是约等于？我用的英伟达驱动版本是450.80.02 我照着Installation Guide在主节点上部署后，发现不能get到GPU的信息。gpushare-device-plugin-ds这个pod的状态一直是pending。）

jax79sg commented 3 years ago

can you run these commands and post the outputs?

kubectl describe pod -n kube-system gpushare-device-plugin-ds-z4dsk
cat /etc/docker/daemon.json
timozerrer commented 3 years ago

I have the same issues, no events are registered. Pod remains pending. daemon.json is as intended