Y2K did not end in worldwide disaster. Its repairs, failures, and recoveries show why, and what database teams can learn about proving maintenance matters.
Why ‘Max Server Memory’ Isn’t Always the Limit
What if I told you that setting a limit with “max server memory (MB)” doesn’t always mean SQL Server will stick to that limit?
Latch Waits at the Top: Reading sys.dm_os_latch_stats
Sample SQL Server latch waits by class, then connect ACCESS_METHODS_DATASET_PARENT, FGCB_ADD_REMOVE, and LOG_MANAGER to the next check.
Before You KILL a Session: Estimating the Rollback Cost
Before you KILL a session, inspect its transaction age, log bytes, current command, and likely rollback work, then monitor the undo safely.
Capturing sp_WhoIsActive Snapshots to a Table Every Minute
Capture sp_WhoIsActive snapshots every minute, keep seven days, and query past blockers and long requests from a logging table.
Performance Condition Alerts: Warnings Before a Log File Fills
Build performance condition alerts for log usage, tempdb free space, and blocking, with sensible delays and an operator response.
Splitting DBCC CHECKDB Across the Week for a Large Database
Keep a large database under regular integrity checks by splitting DBCC CHECKDB work across seven nights and auditing coverage.