One of our Linux Admin’s is trying to convince management that to improve our BigFix Environments performance (they don’t specify what performance metrics will be improved), all we need to do is switch the server from Windows based to Linux based.
Has anyone ever tried this in a production environment?
I have a growing list of reasons why I don’t think we should try it, but I’m looking for a reason why we might WANT to try it. If such a reason exists.
I have seen very large deployments run well on both Windows and on Linux-based deployments. For very large environments there are tuning recommendations, some of them quite detailed and dependent upon your environment.
When deciding which platform to use, I usually advise customers to base that on their tech skill levels on each platform. A lone Linux root server in a Windows-based environment will fare poorly, and vice-versa. You’ll want to ensure you have some depth of knowledge on the Linux/DB2 or Windows/SQL stack.
There are of course a few components that require Windows. The Console for one, and OSD Bare Metal Server (these can run fine with a Linux root server, but do need to execute on Windows platforms).
With an established environment, there is no tool or procedure to migrate your data from Windows to Linux. You’ll need to build the new root server independently, and then “Masthead Switch” your clients to the new environment, so you do lose some of your historic data and need to recreate/re-import your custom properties, analyses, fixlets, etc.
If you are trying to switch because your company wishes to invest in Linux rather than Windows, or your support infrastructure and skillbase is geared to Linux, then it could be worth the effort to migrate. If it’s just to resolve some performance issue, there are almost certainly easier ways to make improvements.
We managed 13K server mix of unix(10K+ wintel 5K) we have fair amount of expertise in both the OS. And we using the wintel for bigfix root server. Its working pretty good for the last 10years.