You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
when restart kubelet, the gpu-device-plugin will Restart and Re-register to the new kubelet, leads to pods that depend on the gpu-device-plugin restarting as well.
#199
Open
Soledao opened this issue
Jul 14, 2021
· 1 comment
because of this, usually, these pods were not available for about 60 seconds.
Normally, a pod that uses only CPU and memory will not restart when restarting kubelet. What solutions can be taken to ensure that pods rely on gpu device are not unavailable when restarting kubelet? Especially when I want to upgrade kubelet.
The text was updated successfully, but these errors were encountered:
because of this, usually, these pods were not available for about 60 seconds.
Normally, a pod that uses only CPU and memory will not restart when restarting kubelet. What solutions can be taken to ensure that pods rely on gpu device are not unavailable when restarting kubelet? Especially when I want to upgrade kubelet.
The text was updated successfully, but these errors were encountered: