+1 lkvm should be taken as a separate activity ..prioritized based on it maturity ..
On 4/11/16, 1:08 PM, "Du, Fan" <[email protected]> wrote: >Thanks to bring *heavy* container to Mesos :) >I have a few comments as below: > >1. >One of reason adding Qemu/KVM is to support scenario where end user env >must be based on Windows, >while as we know Mesos also support Windows Server, if framework user >knows there is native Windows >based container present, there wouldn't need Qemu/KVM to support this. >So for efficiency, the question >(maybe off the topic already) is how/should to make this transparent to >framework user here? > >if mesos has native container based on Windows, then use it. Otherwise >use Windows spawned by Qemu/KVM. > >2. >There is MESOS-5163 to support LKVM Containerization, lkvm is just a >toy(learning vehicle) for virtualization, >it also does not have proven-record for production use cases, moreover >it has few community attention(search >kvm mail list). After Intel announced Clear Container, it has been hold >by ARM,please don't add lkvm here >unless you have compelling reason to do so. > > >On 2016/4/7 20:34, Abhishek Dasgupta wrote: >> Hi all, >> >> I uploaded a design doc for Mesos-2717: >> >> https://docs.google.com/document/d/1_VuFiJqxjlH_CA1BCMknl3sadlTZ69FuDe7qasDIOk0/edit?usp=sharing >> >> >> Need your views and comments on it. >>
