vSphere 5.1 might technically be a “minor” update, but just like vSphere 4.1 there are some very significant new features – many around increased scalability and improved operations. I had the opportunity to spend a little bit of time with the vSphere 5.1 beta and I thought I’d quickly share some of the new feature highlights and perhaps drill into some of these in more detail in the future.
vSphere 5.1 now includes vShield Endpoint for AV protection! For more details on what’s new in vSphere 5.1 also see this whitepaper from VMware
The maximum number of hosts which can share a file in read-only mode has been increased from 8 to 32. These improvements in VMFS locking will enable larger clusters in environments which use linked clones, such as VDI and vCD environments.
The “Monster VM” has also been taking his vitamins as 64 – yes, sixty-four! – vCPUs can now be assigned to a virtual machine. VMware also announced the end of vRAM as a licensing component which means organizations will be able to utilize these features in larger VMs without an additional license penalty.
And finally improvements have been made in MSCS support to enable Failover Clusters with up to 5 (five) nodes.
Host vMotion can now take place in the absence of shared storage. Needless to say a low latency environment (such as a Metro Area Network or better) is ideal, but this can empower new scenarios for migrations and many other scenarios.
Also Storage vMotion is now supported for up to 4 parallel operations across 8 volumes.
- Install ESXi onto an FCoE LUN
- Suport for 16GB HBAs
- Improvements for handing All Paths Down (APD) conditions
- Improved Storage DRS for Linked Clones
- Install ESXi to FCoE disks
- Space Efficient Sparse Virtual Disks for Linked Clones
The last is especially interesting as in the past you had to use SDELELTE and manual cumbersome steps to reclaim space from a VM. Recall that Windows does not delete blocks when a file is deleted, it only removes the directory entry. Now VMware Tools can initiate a scan in the OS of unused blocks and reorganize those blogs to leave a contiguous collection of blocks at the end of the disk. Then a SCSI UNMAP command will be sent to the SAN allowing the space to be reclaimed from your thin disks.
Another advantage on the new sparse disk model is that snapshot chains no longer consume space when deleting snapshots. In the past it was possible to get into a situation where the volume was full and you didn’t have any free space available to consolidate any open snaps. Because now each snap in the chain consolidates directly into the base disk (as opposed to other snaps and then the base last) no additional free space is required to consolidate snapshots.
At times, physical networks uplinks on the ESX host can be configured in such a way that proper function is not possible. vSphere 5.1 includes a new health monitor for networks which checks teaming, VLAN and even MTU settings such that an alarm is created if a network configuration might not deliver the intended results.
In many environments, backups of switch and router configurations are maintained and always at the ready in the event a prior state needs to be restored. Now your vDS switches and their port groups can be backed up and restored, which could come in useful if your vCenter server were to become unavailable. The system can also automatically rollback to the previous networking state if networking is found to have been disrupted.
In addition vDS will now support both Port Mirroring and LACP.
This is a nice feature which can help to organize various vSphere building blocks (VMs, networks, volumes, etc.) and relate them to applications, teams or whatever groupings might be helpful. It works much like tagging pictures for anything to do with “cooking”, “bird watching” or “family” for example. You can tag virtual machines that exist across different vSphere clusters as being part of the same application, business unit or whatever construct you find useful. Then when you search for a tag, you can quickly bring up a list of all the objects which possess that tag.
The vSphere Web Client has gone through a complete overhaul and I think that most will be pleasantly surprise at just how much of the vSphere GUI functionality is now in the web client. I tested the web client with Google Chrome and it was fast, response and an enjoyable experience such that the web client did not feel like a second-class citizen.
Single Sign On
Single Sign On is now provided across the web client and the vCloud infrastructure suite, without having to login to the components individually.
VMware Tools & Upgrades
For those that remember the pre-virtualization days, it was often a chore to update hardware driver components (think HP SIM, Dell OpenManage, etc.). Often times you’d get notification that a certain driver was not the current release, but did that mean there was a compelling reason to upgrade? Now VMware Tools are yesterday’s hardware drivers and with the version changing with some ESX patches, how important is it to update VMware Tools and incur a reboot on your guests?
The new model hopes to reduce this by mapping VMware Tools to the virtual hardware version (now “virtual machine compatibility” in the Web UI. Furthermore, reboots will be required less often in future VMware Tools releases (after the update to 5.1) due to improved driver management. I think that reboot-free VMware Tools upgrades will be a popular feature in many environments.
VMware View environments will benefit from the ability to leverage GPUs to increase the quality of virtual desktops, especially in the areas of full motion video, 3D graphics, and more.
Auto Deploy now supports stateless caching to enable operations to continue when an Auto Deploy server becomes unavailable. Also a new Stateful Install option can make it possible to deploy an ESX host more quickly in several scenarios.
Those are some of the bigger highlights I was able to capture . I’ll be looking for even more details as VMworld progresses.