This week I will be highlighting the top items that I see on almost every SQL Server I run across that need to be changed in almost every instance.
Number 1 by a landslide:
Backups
We’ve been backing up SQL Server in the same way since at least SQL 7.0 from the late 90’s, yet…there are so many bad or missing backup plans:
- No backups at all
- Backups to the same drive as the databases
- Backups to the same server as the databases
- Backups with irregular frequency (manual?)
- Production databases with no transaction log backups (so no ability to restore to a point in time)
- VM backups only, and once a day at best
Every one of these exposes you to data loss that is probably very unacceptable to your customers.
If you think that because you are in the cloud, or have a great SysAdmin/Infra team you are covered, let me give you two reasons to completely re-think your strategy:
- User error
- Ransomware
Your expensive HA solution will not save you from accidentally deleting a table. Ransomware can bankrupt a small business.
Thanks for reading!
I am willing to be in most of those scenarios there has never been a discussion let alone a current discussion or documentation of RTO/RPO.