Once upon a time, neither long ago nor far way, repeated server crashes caused a client's IT management to replace about 40 x86 Linux servers with nice reliable Windows products.
Why? Well, according to a guy they didn't hire to run their Linux infrastructure the key player at the hiring interviews for that job was presented as the User Services Manager but really just ran the help desk - and asked questions like this:
How do you deploy patch kits? I mean, does this sh* even have remote boot?
The woman they eventually hired for the job is an Australian who, in her own words "didn't know Lingal from bingle" but was all hot to give it a go.
Unfortunately, the consequent bottom line on Linux in that organization was a cheque to Microsoft - and soon one to HP too, because now they're running into serious performance problems.
What happened? well, the fun part of writing about human barriers to change is that you can make categorical statements like: "Ignorance is its own ancestor" without much fear of being contradicted on the facts - and in the full knowledge that the statement implies something about the refutations you will get.
In this case senior management killed the Linux initiative before it even got going by ignoring a consultant's repeated warnings that systems management skills are not technology independent; accelerated its already inevitable defeat by allowed the Linux server group to be labelled an experiment within IT; and, doubled down on their earlier mistakes by allowing a Windows bigot to take charge of hiring the new Linux infrastructure manager months after the (contract) installer left and reliability, in the hands of eight different MCSEs with root access to all of the machines, started down the path to Windows.
There's a harsh lesson for Linux implementers here: don't be tentative about Linux, and don't start down that path without having the right people, with the right attitudes, in place first.