Comments on: Virtualizing Even a Single Server https://smbitjournal.com/2015/04/virtualizing-even-a-single-server/ The Information Technology Resource for Small Business Sun, 19 Feb 2017 09:04:46 +0000 hourly 1 https://wordpress.org/?v=6.9.1 By: Scott Alan Miller https://smbitjournal.com/2015/04/virtualizing-even-a-single-server/comment-page-1/#comment-22462 Tue, 30 Jun 2015 18:05:50 +0000 http://www.smbitjournal.com/?p=716#comment-22462 @BS

Patching of the host is certainly a concern and worth considering, especially when choosing between host platforms. By and large hypervisors need very infrequent patches, they are very small, very light and do not have extensive “extras” so the amount to patch is low. (HyperV’s majority of patches come from use cases, not HyperV itself, you can get HyperV patching to be pretty rare if you want to stick with that.)

But there are many factors to consider:

* Patching of VM hosts is rare, is this really a concern at all? Most people patch quarterly, I believe. That’s generally not a major issue.
* Virtualization doesn’t imply consolidation. I’ve never heard of a patching concern that wasn’t someone confusing the two. Consolidation is a separate issue entirely and creates major patching concerns. Never use the caveats of consolidation to rule out virtualization, they are two unique decisions. You should always virtualize, but only consolidate most of the time.
* Because of the nature of limiting consolidation, you can, if need be, align patching between the host and the guest. If you are all Windows on HyperV, you can keep patching “in sync” and only need to patch HyperV when you also need to patch Windows and do them all together making for a single effort, single support window effectively eliminating the concern.
* In most environments of any size, an artifact of virtualization is often that you get transparent guest migration meaning that you can patch the firmware of the host server AND the hypervisor while workloads are shunting off to another box allowing virtualization to actually reduce patching issues rather than increasing them in those instances.
* Virtualization makes guest patching far easier due to the common, easy access to full system state rollbacks from snapshots taken before a patching process. This is just an awesome bonus for patching processes.
* Virtualization dramatically reduces the power cycle time of machines needing to cycle after a patch process also easing the pain of patching guests.

Patching is well worth considering. But with proper planning around workloads I have yet to see an edge case where virtualization introduced a patching concern. Poor planning once in a while will do this (running Windows on KVM so that you have mismatched patching is one reason. But effectively all concerns I have seen are consolidation patching concerns which is not related to virtualization patching concerns and is a completely different matter that should have no bearing on the decision to virtualize.

]]>
By: BS https://smbitjournal.com/2015/04/virtualizing-even-a-single-server/comment-page-1/#comment-22460 Sat, 27 Jun 2015 13:56:23 +0000 http://www.smbitjournal.com/?p=716#comment-22460 Thanks for sharing. I like and agree the idea of abstraction. But in a distributed environment like placing DCs worldwide and run the VM ontop of Windows Hyper-v role i only see the additional effort every patchday for example.
Maybe it makes more sense when VM host is running anything other than Hyper-V with lesser patch frequency. But than again i have the additional effort of another managing environment for managing the VM Host instead using the windows stuff (SCCM) which is already in place.
Do i get something wrong when i say i like Virtualization only in test environments and setups where i can use live migrations (hyper-v cluster setups for example) to avoid VM outage because of managing the host?!

]]>
By: Scott Alan Miller https://smbitjournal.com/2015/04/virtualizing-even-a-single-server/comment-page-1/#comment-22406 Mon, 04 May 2015 18:13:47 +0000 http://www.smbitjournal.com/?p=716#comment-22406 @Steve Just a coincidence 😉 I generally have articles in process for months before publication.

]]>
By: Dave Miller https://smbitjournal.com/2015/04/virtualizing-even-a-single-server/comment-page-1/#comment-22393 Thu, 23 Apr 2015 20:56:02 +0000 http://www.smbitjournal.com/?p=716#comment-22393 Spot on Scott. I am in the process of setting up a new server for a small business (35 users) and am using 2 VM’s for them on their host server. Backups, imaging, snapshots, etc. – make it a no-brainer.

]]>
By: Steve Flowers https://smbitjournal.com/2015/04/virtualizing-even-a-single-server/comment-page-1/#comment-22392 Tue, 21 Apr 2015 20:04:42 +0000 http://www.smbitjournal.com/?p=716#comment-22392 Scott, great post as usual. I almost feel like this came after a thread on Spiceworks where you advised me on my co’s P2V scenario, coincidence?
Anyhow, great points you highlight here. These will go great with any discussion an admin has with his boss about why to go virtual.

]]>
By: Jacob Joyner https://smbitjournal.com/2015/04/virtualizing-even-a-single-server/comment-page-1/#comment-22384 Thu, 16 Apr 2015 16:36:17 +0000 http://www.smbitjournal.com/?p=716#comment-22384 This is a very well timed article. We are currently in the process of moving P2V and one of the questions that was brought up to me was along the lines of “what if we decide to move to virtualization a little more gently by virtualizing to the servers that we are using now?”

I’ll be passing this one along to my boss and the CIO.

-jjoyner1985

]]>