So vSphere 6 is finally here!! VMware has given us some pretty big new features as well as some great updates to some already existing. I will go over these at a high-level here and will be diving into each of these, and then some in the next few articles. So lets dive right into the cool stuff.
First off, VMware has made installing vCenter MUCH nicer. Those of you running the Windows version today will be happy to know that that is not going away. In fact its now easier and is more powerful to install. There are some really nice custom options you can now choose, paths, ports and better error messages (which is something i’ve been complaining about for some time). VMware has now also made it a lot easier to automate the installs, as scripting and automation was actually considered when creating the installers and not just an afterthought. Also, there are some big changes for those who want to run the vCenter Server Appliance (vCSA). UPDATE: The installer will be changing from what we know, but it is NOT the same as the Beta.
One of the next huge things with this release is we now finally see VVol’s for real. I could spend a whole article on VVols, and i will, but for now think of them as a single VM per an individual LUN/Share. With VVOLs you are no longer creating a giant VMFS datastore and dumping in all of your VMs. Think of it as each VM gets its own “datastore”. They utilize new APIs (VASA) to create VM-Aware volumes. This is a feature that must be supported by your storage provider as these are new APIs and could allow for WAY more LUNS/Shares then arrays typically support. . This is a major feature and something that will change the way we think of snapshots, backup and other storage related tasks.
The other large storage related thing that has been a HUGE thorn in my side is VMware is finally supporting NFS 4.1. Since 4.1 has been around since 2010, this is a long time coming. What this will allow is Cluster-Aware NFS and potentially pNFS. ESX also now allows using Kerberos as an authentication method for NFS, so its no longer secured just by an IP address. This is major, in many of my deployments customers have either gone with iSCSI because they need the multi-pathing or they have had to buy expensive HA-type storage arrays where the NFS IP fails over to another head, which can take time and causes other issues.
There are also some big enhancements to some existing features that will make things much more scale-able and easier to use. The first enhancement is to vMotion. VMware now allows vMotions to occur across vCenter Servers. They also allow you to vMotion between different Datacenter Objects, Folders and even Virtual Switches within the same vCenter. This makes your VMs much more mobile within your envrionment.
Because of the ability to vMotion between vCenters and Datacenter Objects, one really neat ability to do “Long Distance vMotion”. This means you can vMotion a VM between your vCenters as long as the link has less then 100ms latency times. Now there are obviously some gotchas as your data change rate will become a factor in if the VM will ever be able to really catch up and actually vMotion. If the change rate is too high, the “replication” may never be able to keep up and thus it’ll never finish the vMotion, and then timeout. However VMware no long disallows you to try it. As somebody who has lived in the backup/recovery world, please do not use this as your sole BCDR strategy.
Next up is something that has been asked for for some time, and rumored for almost as long. You can now do Multi-Processor Fault Tolerance (SMP-FT), which is really awesome. This now allows you to do FT for real workload VMs, not just simple single vCPU ones. There is a current limitation of 4 vCPUs on a FT enabled VM though, so your giant database servers may not be able to fit. However, you may be able to protect more then a simple Windows 7 VM now. One thing to note is if you are doing FT on a 4x vCPU VM you better have 10gb+ Ethernet. In doing some testing i’m seeing some pretty high network throughput numbers on my 10gb pipes at home, which makes sense since CPU instructions are being replicated between servers. Update: vCenter Server FT will be supported in some limited cases, this is still being worked out.
VMware has also brought vCloud Air integration into vSphere, which will allow failover/failback to and from vCloud Air to your environment. In addition they have improved vSphere Replication to now support more VMs per vCenter, 2000VMs are supported. They have also increased the possible RPO to 5 minutes which is pretty impressive for any product, especially one that isn’t designed as a full-blown backup/recovery product.
One of the last things i’ll mention that again has been rumored for some time and actually took me by surprise is around the legacy C# host client. It is not actually going away. It still exists and connects to the ESXi hosts just fine. In fact they actually gave it a feature update for the first time in a while, it can now do some work with VMs running Virtual Hardware 8 through 11. You no longer are forced to use a web-client to work with them. So it can be used on 5.0-6.0 hosts and VMs which is a really nice change. However, like current 5.x version new features will still need the Web Client.
There is a plethora of other minor changes and updates and new features out there in this release. This is truly a x.0 release as it has a TON of new features and really brings things to the next step. I highly recommend checking out my other 6.0 articles as well as all the other well written ones out there as there is a ton of new information available for this release.