For some time now, I have been fighting against CIOs who are imposing increasingly frequent shutdowns of database servers under the pretext of security. I nevertheless think that these regular shutdowns of RDBMS type servers like Microsoft SQL Server go completely against security, especially when they are systematic, as is the case with Windows UPDATEs. Let me explain by taking the case of SQL Server.... Each time we stop the SQL Server service, or the virtual machine or reboot the physical server, the database files, which are locked for the exclusive use of SQL Server, become vulnerable at the moment of the stop. One of my clients who had a massive attack on his RDBMS saw all the files of his databases under Linux encrypted, while the files of the SQL Server databases under Windows were not... Thinking he was doing the right thing , supposedly to protect MS SQL Server instances he stopped them and those files were encrypted!
I have just written an paper on this subject in order to warn about this practice which I consider very bad.
Of course you have to pass the Windows and SQL Server patches... but there is no point in passing them if these patches do not contain a CVE or do not correct a problem that we obviously have. .. In other words there is no urgency to pass these tnat patches that we are not concerned in terms of vulnerabilities or a correction of the functional.
The question is are you pro or con a systematic shutdown of servers ?