[Neil]

Hi

You can call me next week and ask for Neil and I can run you through in much more detail.

However....

You are correct though with Vmware HA the VMs are restarted on the other hosts if one fails, and if the failed host does not come back after 15 minutes a new host is added so you are back to full resources. The VMs are stored on the datastores and not the hosts.

The datastores are redundant as well, each one has a master and slave managed by us. The vLans are managed in vSphere, with the VMs networking you just add the network card to the vLan you want to use.

[robw]

Hi,

We currently have a VMWare private cloud in testing, and I have a few questions. We did speak to OVH on the phone about this, but the answers and the way they were given didn't fill me with confidence. So I'm hoping someone from OVH or someone else who has experience with the private cloud offering can answer a few queries...

If we take one of the basic packs as an example, we would have 2 nodes and 2 datastores.

In this scenario, should a node fail, we are told that a replacement server will be brought online and added to our cloud within 15 minutes. We were also told on the phone that any VMs running on the failed node would automatically be brought back online fully intact on the replacement node. Obviously the VM would be stored on the shared datastore, so I guess is possible if the VM config is stored centrally somehow too.. Is this definitely the case? We can fully utilise both nodes and not have to worry too much about redundancy as this will be handled for us should a node fail?

Is this also the case with the 2x 300gb datastores? They are both redundant and can be fully utilised without having to replicate data ourselves from one to the other?

Also, is there any guides on the networking/vRack side of things? The trial we've been given has 10 VLANs (10-20) set up in VSphere, but I can't find any documentation on how to configure it. I think I can guess how it's supposed to work, but the lack of any real docs does trouble me slightly.

Thanks
Rob