Microsoft accounted yesterday that SQL server will plan run in Linux, with early private preview available today along with full launch for mid-2017. You can find the official blog here. If you like to sign up for the preview, you can apply here.
I assume that is the logical move for SQL server as a product, since other major DBMS can run on different platforms, for example, MySQL, Oracle and Sybase can all be running in Windows and Linux, making SQL server to be able to run on Linux will match the other DBMS. Apart from crossing the checklist of compatibility, does it really make a big different? Let’s explore the options here:
- Linux is more stable then Windows – With the newer version of Windows and Windows core, this difference become much insignificant then before. As less patch is required for Windows core, planned downtime become much less than before.
- Connectivity from application – If you are using Linux application, PHP or Java apps, they can connect to Windows base SQL server as of today, connecting to Linux base SQL might be using different drives (and potential performance grain), but it shouldn’t be the driving force of using it.
- Licensing – You can save the license cost of the OS since Linux is open source, but I don’t think SQL server will be open source in Linux platform, meaning you will still need to pay for license, I would assume the license cost would be the same with Window base. So you would only save the Windows license cost.
What does it mean for DBAs:
- First thing comes to mind will be you will need to know Linux, most DBA would know basic Windows and clustering, but might not know Linux, you will need to know the basic of Linux to perform troubleshoot and/or setup of SQL.
- Clustering setup – similar to the above, you will need to know how to setup SQL cluster in Linux world, along with all the troubleshoot along with it.
- Performance capturing – there will be different tools from Linux to capture server performance, compare to Windows perfmon where it captures all the disk usage and throughput, in Linux, you might need to use other tools.
With the above, I have not think of a very useful case to use SQL server on Linux yet, it would be interesting to see what is your options or plans for using it, please share it in the comments.
