http://www.microsoft.com/systemcenter/virtualmachinemanager/en/us/whats-new-R2.aspx

With VMM R2 you get “maintenance mode” which does in fact give you automatic evacuation of all VMs. Here’s the new features of VMM R2:

http://www.microsoft.com/systemcenter/virtualmachinemanager/en/us/r2.aspx

No religious wars here—just open, frank, and honest discussions.

Those are (obviously) all valid examples. Of course, I’d already mentioned the Update 2 timebomb (the first link above), so we only have 3 examples of “below standard” QA process. I’d agree that the QA process can improve (everyone’s can improve), but I don’t know that I would deem it “below standard”.

In any case, my comment about VMware “lessening” the impact is that it is, to my knowledge, still easier to mitigate the impact of patching. I’ll fully concede that I could be misinformed here, but does Microsoft have a solution to evacuate all VMs from a host before patching it? Or do VMs have to be manually evacuated? As consolidation ratios increase—and I’m assuming that you want higher consolidation ratios—this becomes more important. Live migration helps, most certainly, but there is more to this story than just live migration.

And you are absolutely correct about patching the guest OS inside the VMs; I’m glad you pointed that out. And I’m also glad to hear that Microsoft is making headway in producing patches that don’t require reboots.

Good discussion!

http://blog.scottlowe.org/2008/08/11/apparent-datetime-issue-with-update-2/
http://blog.scottlowe.org/2008/12/23/random-reboots-with-vmware-esx-35-update-3/
http://blog.scottlowe.org/2008/12/12/vmware-ha-problem-with-update-3/

I’m not trying to start a religious war. I’d hope you personally believe that VMware has the better approach. Otherwise I’d question why you are writing books about the product. Personally I think Hyper-V has the better approach for the simply selfish reason that the product has worked better for me and my needs. I simply didn’t find any value in the Virtualization Review piece. It lacked anything concrete and was simply a regurgitation of some VMware talking points. I’d expect more from “analysts”. I was confused as to why you linked to it and then made the incorrect statement of VMware lessening the impact of patches vs. Hyper-V. On a semi-related side note. One thing I’ve noticed since bumping some systems to 2008 R2 is an increase in patches that do NOT require a reboot. Again somewhat a moot point from the virtualization standpoint because of live migration, but I was happy having less need to reboot some of the app servers (which is something else lost in this conversation, no matter what you still need to patch the VMs themselves and therefore at times reboot). Of course that could simply be due to the “freshness” of the OS and we might see an increase in reboot requirements as it ages like all things.

Can you elaborate as to the specifics that lead you to state that VMware’s patch reliability “has been below standards”? We are all aware of the ESX 3.5 Update 2 fiasco, but if that’s all you have to go on that’s pretty weak, IMHO.

I agree that the addition of live migration to Hyper-V certainly changes the picture. I would also agree that converts are the primary targets here, not the existing installed base.

I would disagree with the “historical data” that shows VMware with a larger patch footprint than Microsoft; that is a fallacy foisted by Microsoft. This “larger patch footprint” is due to the fact that VMware updates (by choice, not by necessity AFAIK) the entire ESXi image.

Otherwise, good comments. Thanks for sharing Nate!

Historically speaking my HA and DRS related articles have scored high. Especially both the HA and DRS deepdive score at least 500 unique views per day.

This post was mentioned on Twitter by PlanetV12n: Virtualization Short Take #31 (blog.scottlowe.org) http://bit.ly/2JkfkI...