CloudStack generally keeps its hands off of vms that do not conform to its naming scheme (i-x-abcde). If the name conforms to this naming scheme then the vm is likely to be killed by CloudStack since CloudStack has no record of this.
Capacity calculations will be off I think if you run non-CloudStack Vms (depending on the hypervisor). On 3/29/13 8:37 AM, "Wang Fei" <pytho...@gmail.com> wrote: >From the cloudstack installation guide, I know that I should have a "clean >hypervisor host" when deploy cloudstack. But I think that is ideally. >sometimes, we need to migrate host with existed VMs into cloudstack in >practice. of course, we can export the VMs from hypervisor and then >import >it after deploy cloudstack complete. It is very inefficient. I have >thought >about this issue for a long time. I found network and storage is the >biggest barrier for migrating. Existed VMs may use local storage. then can >not put it into a zone only with share storage. we can also not assign a >proper network to it since we need to keep the ip unchanged. But I think >we >can work around these issue. and I make a simple test to add a host with >some VMs into cloudstack. The happen after that surprise me. Cloudstack >connect to the host successfully and I can access all existed VMs in the >host as normal. Did it indicate that it is possible to migrate host with >VMs into cloudstack? If not, would you like to tell me what would happen >and what risk should I take. > > > >---- >best regards
