3

I'm after some advice. I asked the question below in good faith but it's been closed (as non-constructive) and I'm at a bit of a loss in figurng out how I can ask it the right way. Any help appreciated.

https://serverfault.com/questions/497491/common-reasons-why-virtually-hosted-backend-applications-might-fail

4

This one failed the "too broad" test, as there is no one true answer and could get a lot of brain-stormy answers. In short, it's focus is very open ended rather than narrow. Questions phrased like:

What are some common issues that affect...

Do badly here. Questions here need to be about one or a small set of already known issues that we can help resolve.

I'm not sure how to change your question to fit seeing as you're looking for "what have I missed" kinds of things.

  • Thanks, that helps :) Bit of a shame those sorts of questions don't do well, but at least I know why it was closed. – Adrian K Apr 9 '13 at 23:39
  • 3
    @AdrianK A good place to brainstorm this sort of stuff is the site chatroom. While it is often a wretched hive of scum, villainy, and cranky neckbeardy types a lot of the neckbeardy types that hang out there are pretty smart when they're not being cranky. – voretaq7 Apr 10 '13 at 5:59
  • @voretaq7 - thanks! BTW: great sales pitch, you should go into PR. – Adrian K Apr 11 '13 at 4:43
3

It's really hard to definitively answer the kind of question you're asking.
The crux of what you want to know is summarized toward the end of your question:

In what ways could an application fail that require "some time" to get the application back up and running (but not so fatal that DR [was] invoked)?

And the answer is "It depends" (on, among other things, your application, its failure modes, your downtime tolerance, and your DR plan). We could give you lists-of-things, but those lists are always going to be vague and incomplete.
Conversely if you come up with the list yourself (see the comment I left on the question for a point in the right direction) you'll have a better starting point to understanding what could go wrong and how you want to react (fix, flee (to DR sites), or F it all and sell the company).


Now once you've done a FMEA or fault tree analysis you'll probably have some new items you hadn't thought about ("What happens if a squirrel gets into the datacenter and chews through all the fiber cables?!") -- Those we can probably help you with, as they're more specific and narrowly scoped.

  • Cheers! FYI - The last option ("F it all...") not so viable in this case as we're talking about Govt sector ;) – Adrian K Apr 9 '13 at 23:47
  • 2
    Mmh, yes - the government (well at least MY government) is already on long-term leaseback to the lobbyists. They might get upset if you sold it while they're still using it :-) – voretaq7 Apr 10 '13 at 5:51

You must log in to answer this question.

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