The latest question of mine "when not to use virtualisation" got closed pretty fast.

The assumption for the close was that it would cause debate/discussion (my guess, after I reviewed the comment about KVM/VMWare) - I would rather have expected "not a real question", since the question is somewhat broad.

Anyway I think the question is still interesting, but I have no really good idea on how to reword it, so that answers will stay on focus without causing discussion.

I welcome any ideas about it - even if the idea is "this is the wrong place to ask that".

  • 4
    General commentary: If you fee your question needs something like Update I do not want a debate on this topic - just your facts. Optimal responses would come from people using a broad range of servers and a broad range of techniques. added to it, that's usually a sign it's not a good question. Good questions should not require inline explanation of what you (the asker) wants: Good Answers should proceed naturally from the question. If they don't there's usually a fundamental flaw in the question...
    – voretaq7
    Commented Nov 27, 2012 at 22:26
  • @voretaq7 I had an uneasy feeling about this question right when I asked it - it is in a very grey area - I knew that, but the question and answers should help professional sysadmins - and that is right on focus, is it not?
    – Nils
    Commented Nov 28, 2012 at 22:15

3 Answers 3


Hmm, I dunno, if we can allow these questions:

(ok so that's all I could find on a whim, but I know there are others in the same vein) then I'm willing to give your question a 2nd chance. I'll let others give their feedback on meta first and keep an eye on it before re-opening it though.

  • Thanks for the edit and re-open. The answers up until now go into the right direction.
    – Nils
    Commented Nov 28, 2012 at 22:09

What are you trying to ask?

"When should I not virtualize" is an enormously open-ended question with a huge breadth of possible answers. There's no possible way for an answer (or set of answers) to encompass all of them - it would degrade quickly into a forum-style "But that works for me" / "that doesn't work in environments where X" free-for-all.

As was pointed out in the comments, this is also not something that can be answered generically - the hypervisor(s) you're using and possibly even the guest OS will play important roles in the answer.

Bottom line for me, this question as written runs afoul of this chunk of our FAQ:

You should only ask practical, answerable questions based on actual problems that you face. Chatty, open-ended questions diminish the usefulness of our site and push other questions off the front page.

Your questions should be reasonably scoped. If you can imagine an entire book that answers your question, you’re asking too much.

The only way I can see to salvage it would be to reword the question as a specific application of virtualization technology (hypervisor, underlying server/storage/network subsystems) and intended workload.
Even that would be bordering on "too localized" -- it's a performance question that can really only be answered by benchmarking and deciding if your system's performance as a VM would be adequate. Every environment will generate its own rules on what can/can't virtualize based on in-house knowledge of the workloads and capability of the virtualization platform.

  • I've just done exactly as you suggested; re-worded it asking for specific metrics/answers rather than "What do you think?" Commented Nov 27, 2012 at 22:25
  • 1
    @MarkHenderson: It's still asking for a list(s) of <Things> which doesn't produce good authoritative answers.
    – user9517
    Commented Nov 27, 2012 at 22:38
  • 3
    @Iain - it's asking for a list of metrics, which is generally on-topic, as per the let's go shopping link we love to link. Don't ask "What's best?" ask "What makes something the best". I feel this is another variation of that. Commented Nov 27, 2012 at 22:40
  • @MarkHenderson Yeah I'm generally OK with it as a "What should I measure to see if virtualization is going to suck for me?" question
    – voretaq7
    Commented Nov 28, 2012 at 15:57
  • There were two reasons to ask this question. First - I came across that our Windows-Hyper-V-guys developed quite the same limits as we did with Linux/XEN. Second there was a question on how to reduce interrupts in a VM where the network-interrupts maxed out on a specific interface. IMHO this would max out on any VM-technique (if you stay with the same hardware).
    – Nils
    Commented Nov 28, 2012 at 22:12

It's a crap and highly subjective question and I've voted to close it again. There are at least two good close reasons:

  1. Not constructive
  2. Not a real question

Take your pick.

That fact that there are still some old equally crappy and subjective questions on the site is no reason to accept any more.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .