- Introduction
- General details
- Transaction Isolation Levels
- Lock types
- Locking optimizer hints
- Lock Escalation
- Deadlocks
- Viewing locks (sp_lock, sp_lock2)
- Literature
Introduction
In this article, I want to tell you about general details of
SQL Server 7.0 locking, about Transaction Isolation Levels,
what kind of Transaction Isolation Level exists, and how you
can set the appropriate Transaction Isolation Level, about
Lock types and Locking optimizer hints, about Lock Escalation,
about Deadlocks, and about how you can view locks with sp_lock
or sp_lock2 stored procedure.
General details
There are two units of data storage in SQL Server 7.0:
- Page
- Extent
SQL Server 7.0 stores data on the 8Kb data pages. Only 8060 bytes are
used to store the user's data, other space is used by SQL Server to
store system information.
When you insert a new row, and there is no space on the current data
page to store this row, then new page will be created.
Extent consists of eight 8 KB pages. When you create new table, then
new Extent will be generated. In SQL Server 7.0, different objects can
share an extent or an object can have its own extent(s). A table and
index both have a minimum of two pages.
SQL Server 7.0 can lock the following types of items:
- RID
- Key
- Page
- Extent
- Table
- Database
RID is a row identifier. It is used to individually lock a single row
within a table.
Key is a row lock within an index. Used to protect key ranges in
serializable transactions.
Page lock is a lock, when entire 8-KB data page or index page will
be locked.
Extent lock is only used for allocation. When it's used, entire
extent
will be locked.
Table lock is used when a large percentage of the table's rows are
queried or updated. This lock includes all table's data and indexes.
Database lock is used when you restore the database.
Transaction Isolation Levels
There are four isolation levels:
- READ UNCOMMITTED
- READ COMMITTED
- REPEATABLE READ
- SERIALIZABLE
SQL Server 7.0 supports all of these Transaction Isolation Levels and can
separate REPEATABLE READ and SERIALIZABLE.
Let me to describe each isolation level.
READ UNCOMMITTED
When it's used, SQL Server not issue shared locks while reading data. So,
you
can read an uncommitted transaction that might get rolled back later. This
isolation level is also called dirty read. This is the lowest isolation
level.
It ensures only that a physically corrupt data will not be read.
READ COMMITTED
This is the default isolation level in SQL Server. When it's used, SQL
Server
will use shared locks while reading data. It ensures that a physically
corrupt
data will not be read and will never read data that another application has
changed and not yet committed, but it not ensures that the data will not be
changed before the end of the transaction.
REPEATABLE READ
When it's used, then dirty reads and nonrepeatable reads cannot occur.
It means that locks will be placed on all data that is used in a query,
and another transactions cannot update the data.
This is the definition of nonrepeatable read from SQL Server Books Online:
nonrepeatable readWhen a transaction reads the same row more than one time, and between the
two (or more) reads, a separate transaction modifies that row. Because the
row was modified between reads within the same transaction, each read
produces different values, which introduces inconsistency.
SERIALIZABLE
Most restrictive isolation level. When it's used, then phantom values cannot
occur. It prevents other users from updating or inserting rows into the data
set until the transaction is complete.
This is the definition of phantom from SQL Server Books Online:
phantomPhantom behavior occurs when a transaction attempts to select a row that
does not exist and a second transaction inserts the row before the first
transaction finishes. If the row is inserted, the row appears as a phantom
to the first transaction, inconsistently appearing and disappearing.
You can set the appropriate isolation level for an entire SQL Server session
with the SET TRANSACTION ISOLATION LEVEL statement.
This is the syntax from SQL Server Books Online:
SET TRANSACTION ISOLATION LEVEL{
READ COMMITTED
| READ UNCOMMITTED
| REPEATABLE READ
| SERIALIZABLE
}
You can use DBCC USEROPTIONS command to determine the Transaction Isolation
Level currently set. This command returns the set options that are active
for the current connection. This is the example:
SET TRANSACTION ISOLATION LEVEL READ UNCOMMITTEDGO
DBCC USEROPTIONS
GO
Lock types
There are three main types of locks that SQL Server 7.0 uses:
- Shared locks
- Update locks
- Exclusive locks
Shared locks are used for operations that do not change or update
data, such as a SELECT statement.
Update locks are used when SQL Server intends to modify a
page, and later promotes the update page lock to an exclusive page
lock before actually making the changes.
Exclusive locks are used for the data modification operations,
such as UPDATE, INSERT, or DELETE.
Shared locks are compatible with other Shared locks or
Update locks.
Update locks are compatible with Shared locks only.
Exclusive locks are not compatible with other lock types.
Let me to describe it on the real example. There are four processes,
which attempt to lock the same page of the same table. These processes
start one after the other, so Process1 is the first process, Process2
is the second process and so on.
Process1 : SELECT Process2 : SELECT Process3 : UPDATE Process4 : SELECT
Process1 sets the Shared lock on the page, because there are
no another locks on this page.
Process2 sets the Shared lock on the page, because Shared
locks are compatible with other Shared locks.
Process3 wants to modify data and wants to set Exclusive lock, but it
cannot make it before Process1 and Process2 will be finished, because
Exclusive lock is not compatible with other lock types. So,
Process3 sets Update lock.
Process4 cannot set Shared lock on the page before Process3 will be
finished. So, there are no Lock starvation. Lock starvation
occurs when read transactions can monopolize a table or page, forcing
a write transaction to wait indefinitely. So, Process4 waits before
Process3 will be finished.
After Process1 and Process2 were finished, Process3 transfer Update
lock into Exclusive lock to modify data. After Process3 was finished,
Process4 sets the Shared lock on the page to select data.
Locking optimizer hints
There are twelve Locking optimizer hints in SQL Server 7.0:
- NOLOCK
- HOLDLOCK
- UPDLOCK
- TABLOCK
- PAGLOCK
- TABLOCKX
- READCOMMITTED
- READUNCOMMITTED
- REPEATABLEREAD
- SERIALIZABLE
- READPAST
- ROWLOCK
The first six Locking optimizer hints are used in SQL Server 6.5 and
was described in my previous article about "Locking in SQL Server 6.5".
See this article for more information:
SQL Server 6.5
The last six Locking optimizer hints are the new SQL Server 7.0 hints.
Let me to describe them.
NOLOCK is also known as "dirty reads". This option directs SQL Server
not to issue shared locks and not to honor exclusive locks. So, if this
option is specified, it is possible to read an uncommitted transaction.
This results in higher concurrency and in lower consistency.
HOLDLOCK directs SQL Server to hold a shared lock until completion
of the transaction in which HOLDLOCK is used. You cannot use HOLDLOCK
in a SELECT statement that includes the FOR BROWSE option. HOLDLOCK is
equivalent to SERIALIZABLE.
UPDLOCK instructs SQL Server to use update locks instead of shared
locks while reading a table and holds them until the end of the command
or transaction.
TABLOCK takes a shared lock on the table that is held until the
end of the command. if you also specify HOLDLOCK, the lock is held until
the end of the transaction.
PAGLOCK is used by default. Directs SQL Server to use shared page
locks.
TABLOCKX takes an exclusive lock on the table that is held until the
end of the command or transaction.
READCOMMITTED performs a scan with the same locking semantics as a
transaction running at the READ COMMITTED isolation level. By default,
SQL Server operates at this isolation level.
READUNCOMMITTED is equivalent to NOLOCK.
REPEATABLEREAD performs a scan with the same locking semantics as a
transaction running at the REPEATABLE READ isolation level.
SERIALIZABLE performs a scan with the same locking semantics as a
transaction running at the SERIALIZABLE isolation level. Equivalent to
HOLDLOCK.
READPAST skips locked rows. This option causes a transaction to skip
over rows locked by other transactions that would ordinarily appear in
the result set. The READPAST lock hint applies only to the SELECT statement.
Note You can only specify the READPAST lock in the READ COMMITTED
or REPEATABLE READ isolation levels.
ROWLOCK uses row-level locks rather than use the page or table-level
locks.
You can specify one of these locking options in a SELECT statement.
This is the example:
SELECT au_fname FROM pubs..authors (holdlock)
Lock Escalation
This is the description of the Lock Escalation process from the
SQL Server Books Online:
Lock escalation is the process of converting many fine-grain locksinto fewer coarse-grain locks, reducing system overhead.
You cannot customize locking by setting Lock Escalation level in
SQL Server 7.0. There is no such option now. SQL Server 7.0
automatically escalates row locks and page locks into table locks
when a transaction exceeds its escalation threshold.
Lock escalation thresholds are determined dynamically by SQL Server 7.0
and cannot be configured manually.
Deadlocks
Deadlock occurs when two users have locks on separate objects and
each user wants a lock on the other's object. For example, User1
has a lock on object "A" and wants a lock on object "B", and User2
has a lock on object "B" and wants a lock on object "A". In this
case, SQL Server 7.0 ends a deadlock by choosing the user, who will
be a deadlock victim. After that, SQL Server rolls back the breaking
user's transaction, sends message number 1205 to notify the user's
application about breaking, and then allows the nonbreaking user's
process to continue.
You can decide which connection will be the candidate for deadlock
victim by using SET DEADLOCK_PRIORITY. In other case, SQL Server
selects the deadlock victim by choosing the process that completes
the circular chain of locks.
So, in a multiuser situation, your application should check for
message 1205 to indicate that the transaction was rolled back and
if so, to restart the transaction.
Note To reduce the chance of a deadlock, you should minimize the
size of transactions and transaction times.
Viewing locks (sp_lock)
Sometimes you need a reference to information about locks.
Microsoft recommends to use sp_lock system stored procedure
to report locks information. This very useful procedure returns the
information about SQL Server process ID, which lock the data, about
locked database ID, about locked object ID, about locked index ID
and about type of locking (type, resource, mode and status columns).
This is the results set of sp_lock stored procedure:
spid dbid ObjId IndId Type Resource Mode Status------ ------ ----------- ------ ---- ---------------- -------- ------
1 1 0 0 DB S GRANT
6 1 0 0 DB S GRANT
7 1 0 0 DB S GRANT
8 1 0 0 DB S GRANT
9 1 0 0 DB S GRANT
9 2 0 0 DB S GRANT
9 1 117575457 0 TAB IS GRANT
10 1 0 0 DB S GRANT
11 5 0 0 DB S GRANT
The information, returned by sp_lock stored procedure, is needed in
some clarification, because it's difficult to understand database
name, object name and index name by their ID numbers.
Microsoft provides an enhanced version of the sp_lock system stored
procedure, which returns user name, host name, database name and
object name also.
You can find this stored procedure at here:
INF: sp_lock2 Returns Additional Locking Details
http://support.microsoft.com/support/kb/articles/Q255/5/96.ASP
This enhanced stored procedure works under SQL Server 7.0 as well, but
has syntax error under SQL Server 2000. It does not return the name of
an index also.
Here you can find the new version of the sp_lock2 stored procedure
for SQL Server 7.0 and SQL Server 2000. This version returns index
name and object owner also:
locking view: sp_lock2
Literature
- SQL Server Books Online
- Transaction Isolation Level
- INF: sp_lock2 Returns Additional Locking Details
http://support.microsoft.com/support/kb/articles/Q255/5/96.ASP
- Detailed locking view: sp_lock2
- INF: Analyzing and Avoiding Deadlocks in SQL Server
http://support.microsoft.com/support/kb/articles/Q169/9/60.ASP