To create an index online in SQL Server, add WITH (ONLINE = ON) to CREATE INDEX, and writers keep working while the index builds. This post compares a plain build with an online build on a three million row table, measures what each one blocks and what the online build costs, and shows where it still waits.
Application Locks With sp_getapplock: One Job at a Time
Use sp_getapplock to serialize a job, compare transaction and session ownership, handle return codes, and inspect a two-session test.
Counting Locks per Session With sys.dm_tran_locks
Count locks per session by resource and mode, connect open transactions, and recognize escalation while keeping snapshot limits clear.
Query Runs Slow Occasionally: Five Checks to Find Why
When a query runs slow occasionally, the cause sits around the query: a lock, a job, a file growth, tempdb or a shortage of worker threads. This post gives one tested T-SQL check for each and a small demo for blocking and file growth. It also says what to do when all five checks find nothing.
READPAST Query Hint: Skip Locked Rows in SQL Server
The READPAST query hint makes a read skip rows that another transaction has locked, instead of waiting for them. This post builds a five-seat demo, runs it in two windows and lists the cases where the hint does nothing. It also shows the work queue pattern that pairs READPAST with UPDLOCK.
Does NOLOCK Really Applies No Lock? – Interview Question of the Week #209
Does NOLOCK really applies no lock? No. NOLOCK takes a Sch-S, or schema stability, lock. Here is a simple script to see it for yourself.
Diagnosing Blocking Chains With sys.dm_exec_requests
Trace blocking chains from active requests to a candidate head session, inspect sleeping transactions, and verify the owner before action.