Hi Marcus, I'll give that a shot. I didn't know if those parameters specified the network CIDR or the guest CIDR.
Thank You, Logan Barfield Tranquil Hosting On Mon, Oct 6, 2014 at 3:15 PM, Marcus <[email protected]> wrote: > Do startip and endip createNetwork parameters not work for that (when > creating the network? That should carve out a subset of the network for > cloudstack use and leave the rest untouched. > On Oct 6, 2014 12:57 PM, "Logan Barfield" <[email protected]> wrote: > > > We have decided internally to set up a CIDR reservation with all new > > accounts to give us the ability to easily attach dedicated hosts to > > existing VM networks. > > > > We were thinking it would be easier to set up the reservation before > > deploying VMs. Setting up reservation after the fact can get complicated > > if a VM happens to be outside the intended reservation range. > > > > The issue we're having is that reservation is not allowed until the > network > > is in the "Implemented" state (i.e. after the first VM is deployed). > > > > Why is reservation not allowed upon initial network creation? If we try > to > > apply reservation after the first VM is online the command will fail > > occasionally because the first VM is deployed outside the CIDR range. > > > > Example: > > > > Guest Net: 10.1.1.0/24 > > Reserved CIDR: 10.1.1.0/25 > > > > - Attempt reservation before deploying a VM: Fails due to network not > being > > "Implemented" > > - Attempt reservation after many VMs are deployed: Fails due to VMs being > > outside Reserved CIDR (e.g., 10.1.1.150), and requires a lot of work to > > change the VM's IP > > - Attempt reservation after first VM is deployed: Either succeeds, or > fails > > if the first VMs IP is outside of the reserved CIDR. > > > > How can we fix this without hacking work arounds into the deployment > logic? > > (ex: Check network for 10.1.1.10, if it doesn't exist deploy the VM on > > that IP, if it already exists deploy it wherever.) > > > > Thank You, > > > > Logan Barfield > > Tranquil Hosting > > >
