My prodding seems to have awakened the Stack Exchange dragons -- It looks like Server Fault will be getting our own survey similar to what has been done for Stack Overflow for a while now.

I've got a few ideas on what I would like to see appearing on the survey, but I know I'm not thinking of even a tenth of the useful stuff we can come up with, so I'm sticking this here:

  • What questions would YOU like to see on a ServerFault survey?

These questions can help us better define and understand our community (and give Stack Exchange valuable data to help get more relevant ads for the site :).

  • 5
    Favourite cheese?
    – Chopper3
    Feb 14, 2012 at 23:42
  • @Chopper3 CHEESE ROLL CALL!
    – voretaq7
    Feb 15, 2012 at 14:56
  • @voretaq7 Your prodding may have gotten a survey, but did we ever get any results from said survey? Sep 7, 2012 at 1:04

6 Answers 6


I think we need to ask something about the size of organizations we support in addition to the basic demographic question, about how many people are employed where you work.

Depending on how you count there I work at an organization with ~200 staff. But we provide regional support for our clients (k12 schools) which have somewhere from ~5 staff and 2 computers up to 500 staff an ~5,000 computers.

I think this is somewhat important since I have been one of the people in favor of some of the more SOHO questions because I support many small organizations. I would like to know how many people are supporting the Fortune 500 companies, and how many are working at or supporting small businesses.

  • 1
    Size does matter -- I know from the questions we get (and the fact that I work in a relatively small organization) that we have a lot of small-office SAs coming in.
    – voretaq7
    Feb 14, 2012 at 22:43
  • Organization size for schools needs to include student count. Both total and FTE.
    – freiheit
    Feb 15, 2012 at 18:59
  • 1
    Such a question would reveal the diversity of our roles, which if nothing else would help some people to realise that not everyone here works in a similar environment to their own. That's gotta be a good thing. Feb 15, 2012 at 20:32

A few of the ideas I'm throwing out there:

  • What operating systems are in use on desktops? On servers?
    (I think a breakdown of Windows, Linux, BSD, Solaris, AIX, Other would be enough here, though it may make sense to break down the Windows chunk into XP, Vista, 7, Other for desktop & 2003, 2008, Other for servers)

  • How many machines are running each OS?

  • How many sysadmins are working in your organization?

  • How many developers are working in your organization?

  • What kind of storage systems are in use in your organization
    (SAN, NAS, Local Disk (DAS))?

    • How much total storage is managed?
  • What kind of virtualization solutions are deployed in your organization?
    VMWare ESX, VMWare Workstation, Microsoft Hyper-V, VirtualBox, Linux KVM, Other

  • What percentage of systems are virtualized
    (may be better to ask as number virtualized using each of the above technologies)?

  • When I read your list it looks too much like one of those surveys we get presented with far too often by trade publications and the like. I don't think any of us get much pleasure out of doing those. The point being that this may be counter productive by discouraging people form participating. I would suggest a far less detailed list. Feb 15, 2012 at 20:38
  • I'm definitely not married to any specific level of detail (though I tend to think that more is better) - I want to come away knowing more about what we do: How many people are Windows versus Unix? How many are using virtualization (and what kind)? etc. and whether we get down to the nitty-gritty or not is secondary to having a broad picture.
    – voretaq7
    Feb 15, 2012 at 20:45
  • What kind of virtualization solutions are deployed in your organization - could also add "None" as I expect there are some shops that are just running plain servers.
    – tombull89
    Feb 17, 2012 at 15:24
  • @tombull89 Good point. I don't know any company that's not even dabbling with VirtualBox, but I'm sure there's one.
    – voretaq7
    Feb 17, 2012 at 15:24


Building a set of questions for a meaningful survey is hard. Questions build bias, bias is built into the questions, and analyzing the results is its own discipline. Best for us is to build a list of things we'd like to know about and then let someone who knows how to write surveys build it for us.

A lot of what we're looking for probably overlaps with the SAGE/USENIX surveys. If we can, pull elements out of that to found sections like demographics (who/what/where) and environments (how-much) we'll save time.

Overall areas I know I'm interested in:

  • Diversity of operating-system environments supported. Whatall are you dealing with? We got a LOT of tags out there. We have some AIX people, some pure Windows, and I'm sure some pure Solaris still out there. And people who do all three and more.
  • Diversity of virtualization frameworks supported. Not everyone uses VMware. Or HyperV. Or KVM.
  • Diversity of network environments supported. We need to disambiguate this from 'storage networks', but Ethernet technologies, WAN technologies, routing, advanced foo.
  • Diversity of storage environments supported. I'm thinking technologies rather than vendor ecosystems, so things like Fibre Channel, iSCSI, FCoE, SCSI, SAS, Infiniband, as well as the major blocks such as storage routers, arrays, and suchlike.
  • Diversity of scripting/automation environments supported. Where SF runs into SO-land, but it's still useful to know.
  • Size of environment supported. This will be a major section of its own, but mainly useful for cross-tabs.
  • Career items
    • Method of payment, Hourly, salary, independent contractor, pony-rides.
    • Number of peers in title/position in your organization. This is fuzzy especially since we have contractor-types who work with a lot of entities where they may be peers. So, good question-smithing will be needed.
    • Rough number of internal users supported. To get an idea of organizational size of responsibility. I'm purposely ignoring 'total internal users in the organization as a whole' as I don't think that's useful. I may be wrong.
    • Rough number of external users supported. For our Google SRE's, this may be "all of humanity", but for the office-IT people it may very well be zero. Building the right options on this question will be an interesting task.

I have some interest in level of certifications, but I believe other surveys handle this better.

  • 1
    all surveys have inherent bias. If we recognize our bias that's a good start :)
    – voretaq7
    Feb 16, 2012 at 5:01

Salary... Trends in local tech environment (hypervisor choice, configuration management tools, networking), future projects... BACKUP PROCEDURES...

  • 1
    maybe salary in conjunction with an area or region, as it can vary from location to location Feb 15, 2012 at 3:10
  • Region is an important one (I think the SO survey asked for this, we definitely should - It makes a lot of the other data more meaningful)
    – voretaq7
    Feb 15, 2012 at 15:59
  • But on backups... that's a big one because I have no idea how most places accomplish their backups.
    – ewwhite
    Feb 15, 2012 at 16:02
  • Region and salary are already on it :); although i think backup procedures might be a bit much for a multiple choice survey
    – Zypher
    Feb 15, 2012 at 18:14
  • 1
    -1 for the salary suggestion. It's so incredibly localised as to be absolutely meaningless on a global site. Discussion of salaries also bring out too much pride, envy/jealousy and discussion. It might almost be bearable on a forum but is definitely not healthy on a Q&A site. Feb 15, 2012 at 20:35
  • Salary, while would be interesting (and likely to make me cry), would depend on the job as well as location. A Junior-level sysadmin in a school would make a lot less that one working in say, a bank, or SME.
    – tombull89
    Feb 17, 2012 at 15:25
  • True, but I think the expectations are a big thing. I've seen job ads offering $45k for the skillset of a $120k engineer. Sysadmins should know what they're worth.
    – ewwhite
    Feb 17, 2012 at 15:27
  • A question about the technical level of the users we deal with

This may be useful when interpreting other items. e.g. Highly technical users generally need fewer admins per N machines/users/servers than say a building full of bean counters.


How about the type of organization, such as municipal, state or federal government? School district? Independent chartered bodies such as port and transit authorities? Non-profit? Private industry?

You must log in to answer this question.

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