Feng Wang
Feng Wang
@yoheiueda This is interesting! I would love to know more about this project. A few questions, 1. How do you deal with the slowness of cloud VM provisioning? Would it...
Is there any performance data comparison after upgrading to iouring?
@LitFlwr0 Thanks for picking this up! Just to echo what @egernst commented, this is a hairy topic and there are various edge cases we need to consider. For example, we...
@fidencio I prefer 3PM UTC but 2PM is also doable :)
Exposing an interface through kata-runtime is potentially an option. Can you explain a bit more about the use case? Sounds like the GPU hotplug/unplug lifecycle is different from the pod/containers?...
@soulseen I'm think whether you can add another sidecar container that has the same lifecycle as the GPU, then you can use the container to manage the hotplug/unplug of the...
@Burning1020 First, thanks for fixing this. Can you address the github action failures? For example, create a tracking github issue and link it to the PR, make sure the commit...
>Binding vCPUs and CPUs. This allows containers to use the CPUs they really want. At the same time, scheduling overhead can be reduced and locality can be exploited for benefits....
We (Databricks) have a use case that are blocked this KEP. Without this, it's really awkward to use block devices with VM based containers.