Database work
Notes on SQL Server performance, recovery and data migrations.
10 articles published
Featured article
SQL Server does not become slow simply because a database is old. Performance declines when data volume, workload shape, execution plans, and operating practices drift away from the assumptions under which the system was built.
Read articlePoor indexing is not limited to missing indexes. Redundant structures, weak key choices, oversized includes, and maintenance rules copied from old playbooks can make reads and writes slower at the same time.
Read articleQuery Store preserves the runtime and plan history needed to investigate regressions that disappear before anyone can capture them. The value comes from a disciplined workflow, not simply switching the feature on.
Read articleNOLOCK trades normal read consistency for the possibility of incorrect results, yet it does not remove every kind of blocking. Reliable fixes begin with the blocking chain, transaction scope, access paths, and isolation requirements.
Read articleA small business rarely needs a full-time database specialist, but certain reliability, performance, migration, and reporting problems become too risky to handle through occasional vendor support or guesswork.
Read articleA legacy SQL Server migration succeeds when the team understands application dependencies, proves behavior on the target, rehearses data movement, and treats cutover and rollback as engineered procedures.
Read articleBackup success messages can conceal serious recovery gaps. Companies need recovery objectives, complete backup chains, protected copies, monitored jobs, and regular restores that prove both the data and the runbook.
Read articleAzure SQL migrations are shaped by service-model fit, application dependencies, network latency, identity, operational responsibility, and realistic workload testing. Moving data is only one part of the work.
Read articleMore CPU, memory, or faster storage can be justified, but a workload review should come first. Otherwise new capacity may conceal inefficient queries briefly while cost and contention continue to grow.
Read articleA useful database health check is an evidence-based review of recovery, integrity, performance, security, capacity, and operations, delivered as prioritized actions rather than a long list of undocumented warnings.
Read article