Home » Dynamic Management View

Category Archives: Dynamic Management View

Always on\Database Mirroring Automatic Page Repair

A special thanks to Daniel Jones [daniel.jones.0543@gmail.com] for showing up interest and writing informative article for our blog (http://mssqlfun.com/).

 

Always on\Database Mirroring Automatic Page Repair

One of the most useful as well as hottest feature in SQL Server is mirroring of database. It is helpful for numerous of organizations that implements the successful as well as relatively high-availability of inexpensive strategies. However, in all the latest versions of SQL Server its service is being updated in numerous of ways such as database mirroring automatic page repair. It repairs the corrupt pages of the database. The automatic page repair is maintained by the database mirroring and Always on Availability Groups. After some specific types of errors, which had occurred in the database further make it in non-readable form. An availability replica tries to automatically repair the pages. The replica, which is unable to tread the page from another replica. If the request is implemented then, all the non-readable pages are replaced by the readable copies. It mainly resolves the occurrence of error.

Note: The automatic repair is completely differing from DBCC repair as all the database is stored by an automatic recovery of page.

Errors Occurred by Automatic Page-Repair

Database Mirroring of Automatic page repair efforts to repair in the data file through which an operation is failed as mentioned below:

  • Error 883

Description: Action is executed only if the OS implement CRC (cyclic redundancy check), which fails on the data.

Instances: ERROR_CRC. The operating system value for this error is 23

  • Error 824

Description: Logical

Instances: Logical data errors like checksum of torn write or bad page.

 

Techniques to Resolve Errors

  • Primary Database

The primary database of automatic page repair is tried when the state of database is SYNCHRONIZED. Moreover, the principal is still transferring log records for the database to the secondary. The elementary sequence of actions in an automatic attempt of page-repair are as mentioned:

  1. When an error of read comes up on a data page in the primary database then, the primary inserts row in the suspect_pages table with the correct status of an error. For database mirroring, the principal then, requests a copy of page from the mirror. For the Always On Availability Groups, the primary broadcasts the appeal to all the secondary’s and acquires the page from the first to react. The request identifies the ID and LSN, which is presently at end of flushed log. The page is marked as pending restore. This marks it inaccessible throughout the automatic page-repair effort. Attempts to utilize this page in the course of the repair attempt will fail with an error code 829.
  2. After getting page request, the secondary gaps till it has recreated the log up to the LSN stated in the request. After that, the mirror efforts to utilize the page in its database copy. If the page can be used, the mirror directs the copy of page to the principal. Otherwise, the secondary proceeds an error to the principal and the automatic page-repair effort fails.
  3. The primary methods the response, which holds the fresh copy of page.
  4. After the automatic page-repair effort fixes a suspect page, the page is noticeable in the suspect_pages table as reinstated.

 

  • Secondary Database

The input and output errors on pages, which happen on the secondary database are held in generally the similar way by mirroring and Always On Availability Groups.

  1. With mirroring of database, if the mirror comes upon one or more page errors of I/O when it rebuilds a log record, the session of mirroring arrives with the SUSPENDED state. Through Always On Availability Groups, if a replica of secondary meets one or more page errors of I/O when it recreates a log record, the secondary database enters the state of SUSPENDED. At that point, the mirror insets a row in the suspect_pages table with the suitable error status. The secondary then, requests a copy of page from the primary.
  2. The principal attempts to use the page in its database copy. If the page can be edited, the primary directs the copy of page to the secondary.
  3. If the mirror obtains copies of each page it has demanded, the secondary efforts to restart the mirroring session. If an automatic page-repair go fixes a suspect page, the page is clear in the suspect_pages table as reinstated.

If secondary does not obtain a page, which is requested from the principal, the automatic page-repair effort fails. With mirroring of database, the mirroring session leftovers suspended. With Always On Availability Groups, the secondary database leftovers suspended. If the mirroring session or secondary database is restarted manually, the corrupted pages will be again hit at the time of synchronization phase.

View Attempts of Automatic Page-Repair

The mentioned dynamic management previews return rows for the up-to-date automatic page-repair efforts on a given availability database, with maximum of 100 rows per database.

  • Always On Availability Groups:

sys.dm_hadr_auto_page_repair

Gives a row for each automatic page-repair effort on any database availability on availability replica, which is hosted for any group availability by server instance.

  • Database mirroring:

sys.dm_db_mirroring_auto_page_repair

Returns row for all automatic page-repair try on any database mirrored on the server instance.

Conclusion

It is important to repair the page to maintain the continuity of the work. Therefore, we have discussed Database mirroring automatic page repair in the above discussion that makes easy for users to remove the hurdle and understand it clearly.

Reference: Rohit Garg (http://mssqlfun.com/)

You can find and follow MSSQLFUN:-

http://www.facebook.com/mssqlfun

http://mssqlfun.com/feed

https://twitter.com/Rgarg86

Other Linked Profiles :-

http://www.sqlservercentral.com/blogs/mssqlfun/

http://social.msdn.microsoft.com/Profile/rohitgarg

http://www.toadworld.com/members/rohit-garg/blogs/default.aspx

http://beyondrelational.com/members/RohitGarg/default.asp

How Row versioning impact tempDB ?

Scenario: DBA found that tempdDBdatabase usage is getting high and most of the size is consumed by row versioning. DBA raised the issue with application team running that query. Once Application team close the session, tempDB usage comes normal.

Question: Application tea raise concern “How can a select statement on table consume huge amount of resources in tempDB?”

Answer:

As per BOL (https://technet.microsoft.com/en-us/library/ms175492%28v=sql.105%29.aspx) : Row versioning-based isolation levels (RCSI – Read Committed Snapshot Isolation) reduce the number of locks acquired by transaction by eliminating the use of shared locks on read operations. This increases system performance by reducing the resources used to manage locks. Performance is also increased by reducing the number of times a transaction is blocked by locks acquired by other transactions. Row versioning-based isolation levels increase the resources needed by data modifications.

While updating the row in a table or index, the new row is marked with a value called the transaction sequence number (XSN), XSN is unique number and keeps updating with upcoming new command \ transactions. When updating a row, New version is generated. If DB is already having previous version of this row, previous version of the row is stored in the version store and the new version of the row contains a pointer to the old version of the row in the version store. New row version stores corresponding XSN value for reflecting the time the row was modified.

All version stores are linked in a chain. SQL Server keep running clean up task to remove old versions which are not in use. As long as a transaction is open, all versions of rows that have been modified by that transaction must be kept in the version store. This is the reason due to which long running open transaction can cause huge tempDB row versioning space.

Here, Row X is getting updated by multiple transactions (Tv,Tx,Ty,Tz) and each time new version store is created. When DB engine tries to clear old unused version stores, it found open transaction using version X. DB engine not able to clear any row version from version store after version X. Due to which TempDB size keeps increasing due to row versioning.

–Check Size of tempDB consumed by Version Store

SELECT SUM(VERSION_STORE_RESERVED_PAGE_COUNT) AS [VERSION STORE PAGES USED],

(SUM(VERSION_STORE_RESERVED_PAGE_COUNT)*1.0/128) AS [VERSION STORE SPACE IN MB],

SUM(INTERNAL_OBJECT_RESERVED_PAGE_COUNT) AS [INTERNAL OBJECT PAGES USED],

(SUM(INTERNAL_OBJECT_RESERVED_PAGE_COUNT)*1.0/128) AS [INTERNAL OBJECT SPACE IN MB],

SUM(USER_OBJECT_RESERVED_PAGE_COUNT) AS [USER OBJECT PAGES USED],

(SUM(USER_OBJECT_RESERVED_PAGE_COUNT)*1.0/128) AS [USER OBJECT SPACE IN MB],

SUM(UNALLOCATED_EXTENT_PAGE_COUNT) AS [FREE PAGES],

(SUM(UNALLOCATED_EXTENT_PAGE_COUNT)*1.0/128) AS [FREE SPACE IN MB]

FROM SYS.DM_DB_FILE_SPACE_USAGE;

–Find session using version store

SELECT A.*,B.KPID,B.BLOCKED,B.LASTWAITTYPE,B.WAITRESOURCE,B.DBID,B.CPU,B.PHYSICAL_IO,B.MEMUSAGE,B.LOGIN_TIME,B.LAST_BATCH,

B.OPEN_TRAN,B.STATUS,B.HOSTNAME,B.PROGRAM_NAME,B.CMD,B.LOGINAME,REQUEST_ID

FROM SYS.DM_TRAN_ACTIVE_SNAPSHOT_DATABASE_TRANSACTIONS A

INNER JOIN SYS.SYSPROCESSES B

ON A.SESSION_ID = B.SPID

Reference: Rohit Garg (http://mssqlfun.com/)

You can find and follow MSSQLFUN:-

http://www.facebook.com/mssqlfun

http://mssqlfun.com/feed

https://twitter.com/Rgarg86

Other Linked Profiles :-

http://www.sqlservercentral.com/blogs/mssqlfun/

http://social.msdn.microsoft.com/Profile/rohitgarg

http://www.toadworld.com/members/rohit-garg/blogs/default.aspx

http://beyondrelational.com/members/RohitGarg/default.asp

Restrict user to login from single Host

Restricting user to login from single host may be sometimes required from security point and other business requirements. We can achieve it through SERVER LEVEL LOGON TRIGGER.

ALTER TRIGGER TR_CHECK_LOGIN_TEST_HOST
ON ALL SERVER
FOR LOGON
AS
BEGIN

DECLARE @HOSTNAME VARCHAR(48)
DECLARE @PROGNAME VARCHAR(100)

SELECT @HOSTNAME = HOST_NAME FROM SYS.DM_EXEC_SESSIONS
WHERE SESSION_ID = @@SPID

IF ORIGINAL_LOGIN() = ‘TEST’ AND @HOSTNAME =’HostName’
ROLLBACK;

END

Once you create above trigger, Login “TEST” will be able to connect from specified host only. Connection from any other host will be failed and face below error: –

You can also find related errors in SQL Server Error log as well.

Error: 17892, Severity: 20, State: 1.

Logon failed for login ‘TEST’ due to trigger execution. [CLIENT: <local machine>]

Error: 3609, Severity: 16, State: 2.

The transaction ended in the trigger. The batch has been aborted.

 

Reference: Rohit Garg (http://mssqlfun.com/)

You can find and follow MSSQLFUN:-

http://www.facebook.com/mssqlfun

http://mssqlfun.com/feed

https://twitter.com/Rgarg86

Other Linked Profiles :-

http://www.sqlservercentral.com/blogs/mssqlfun/

http://social.msdn.microsoft.com/Profile/rohitgarg

http://www.toadworld.com/members/rohit-garg/blogs/default.aspx

http://beyondrelational.com/members/RohitGarg/default.asp

Thanks to Toadworld.com team for recognition !

Thanks to Toadworld.com team for recognition !

Toadworld.com Profile : http://www.toadworld.com/members/rohit-garg/blogs/default.aspx

If you liked this post, do like on Facebook at http://www.facebook.com/mssqlfun

Reference : Rohit Garg (http://mssqlfun.com/)

2013 with You & SQL Server

The WordPress.com stats helper monkeys prepared a 2013 annual report for this blog.

Here’s an excerpt:

The concert hall at the Sydney Opera House holds 2,700 people. This blog was viewed about 41,000 times in 2013. If it were a concert at Sydney Opera House, it would take about 15 sold-out performances for that many people to see it.

Click here to see the complete report.

DMV-15 : Pending I/O requests……..sys.dm_io_pending_io_requests

sys.dm_io_pending_io_requests DMV (Dynamic Management View), described by BOL as follows: http://msdn.microsoft.com/en-us/library/ms188762.aspx

Returns a row for each pending I/O request in SQL Server.

It’s a very simple DMV used to see all pending I/O requests & there description.

You can combine this DMV with DMF – sys.dm_io_virtual_file_stats to see I/O pending requests with database files. You should run this query multiple times to check if the same files or drive letters consistently coming up on the top. If this is the case that means you facing I/O bottlenecks for that file or drive letter.

Query 1 : Details of I/O pending requests against each DB file

SELECT

DB_NAME(MF.DATABASE_ID) AS [DATABASE],

MF.PHYSICAL_NAME,

IPIR.IO_TYPE,

SUM(IPIR.IO_PENDING) TOTAL_PENDING_IO,

SUM(IPIR.IO_PENDING_MS_TICKS) TOTAL_PENDING_MS_TICKS,

SUM(VFS.NUM_OF_READS) TOTAL_READS,

SUM(VFS.NUM_OF_WRITES) TOTAL_WRITES

FROM

SYS.DM_IO_PENDING_IO_REQUESTS AS IPIR

INNER JOIN

SYS.DM_IO_VIRTUAL_FILE_STATS(NULL,NULL) AS VFS

ON IPIR.IO_HANDLE = VFS.FILE_HANDLE

INNER JOIN

SYS.MASTER_FILES AS MF

ON VFS.DATABASE_ID = MF.DATABASE_ID

AND VFS.FILE_ID = MF.FILE_ID

GROUP BY MF.DATABASE_ID, MF.PHYSICAL_NAME, IPIR.IO_TYPE

ORDER BY SUM(IPIR.IO_PENDING)

Remarks

1. To use this DMV, User required VIEW SERVER STATE permission on the server.

If you liked this post, do like on Facebook at http://www.facebook.com/mssqlfun

Reference : Rohit Garg (http://mssqlfun.com/)

DMV-14 : Which Table has Max. rows & Size……..sys.dm_db_partition_stats

sys.dm_db_partition_stats DMV (Dynamic Management View), described by BOL as follows: http://msdn.microsoft.com/en-us/library/ms187737.aspx

Returns page and row-count information for every partition in the current database.

This DMV gives useful information about rowcount, free & reserved space of each table in database. It can be used for analysis purpose while adding new indexes, doing index Maintenance or checking the biggest table of database.

Query 1 : Details of table with row count & space used

SELECT

SS.NAME AS SCHEMANAME,

SO.NAME AS OBJECTNAME,

SI.NAME AS INDEXNAME,

PS.RESERVED_PAGE_COUNT*8 TOTAL_SPACE_CONSUMED_KB,

PS.USED_PAGE_COUNT*8 USED_SPACE_KB,

(PS.RESERVED_PAGE_COUNT – PS.USED_PAGE_COUNT)*8 FREE_SPACE_KB,

CASE

WHEN PS.INDEX_ID IN (0,1) THEN PS.ROW_COUNT

ELSE NULL

END AS ROW_COUNT

FROM

SYS.DM_DB_PARTITION_STATS PS

INNER JOIN SYS.OBJECTS SO ON SO.OBJECT_ID = PS.OBJECT_ID

INNER JOIN SYS.SCHEMAS SS ON SS.SCHEMA_ID = SO.SCHEMA_ID

LEFT JOIN SYS.INDEXES SI ON SI.OBJECT_ID = PS.OBJECT_ID

AND SI.INDEX_ID = PS.INDEX_ID

WHERE

SO.IS_MS_SHIPPED = 0

Sample Result

Remarks

1. To use this DMV, User required VIEW DATABASE STATE permission on the server.

If you liked this post, do like on Facebook at http://www.facebook.com/mssqlfun

Reference : Rohit Garg (http://mssqlfun.com/)

DMV-13 : Finding locking & blocking……..sys.dm_tran_locks

sys.dm_tran_locks DMV (Dynamic Management View), described by BOL as follows: http://msdn.microsoft.com/en-us/library/ms190345.aspx

Returns information about currently active lock manager resources. Each row represents a currently active request to the lock manager for a lock that has been granted or is waiting to be granted.

The columns in the result set are divided into two main groups: resource and request. The resource group describes the resource on which the lock request is being made, and the request group describes the lock request.

This DMV is very useful in helping to identify locking and blocking issues on your data­base instances.

Query 1 : Details of currently active locks

SELECT

CASE DTL.REQUEST_SESSION_ID

WHEN -2 THEN ‘ORPHANED DISTRIBUTED TRANSACTION’

WHEN -3 THEN ‘DEFERRED RECOVERY TRANSACTION’

ELSE DTL.REQUEST_SESSION_ID END AS SPID,

DB_NAME(DTL.RESOURCE_DATABASE_ID) AS DATABASENAME,

SO.NAME AS LOCKEDOBJECTNAME,

DTL.RESOURCE_TYPE AS LOCKEDRESOURCE,

DTL.REQUEST_MODE AS LOCKTYPE,

ST.TEXT AS SQLSTATEMENTTEXT,

ES.LOGIN_NAME AS LOGINNAME,

ES.HOST_NAME AS HOSTNAME,

CASE TST.IS_USER_TRANSACTION

WHEN 0 THEN ‘SYSTEM TRANSACTION’

WHEN 1 THEN ‘USER TRANSACTION’ END AS USER_OR_SYSTEM_TRANSACTION,

AT.NAME AS TRANSACTIONNAME,

DTL.REQUEST_STATUS

FROM

SYS.DM_TRAN_LOCKS DTL

JOIN SYS.PARTITIONS SP ON SP.HOBT_ID = DTL.RESOURCE_ASSOCIATED_ENTITY_ID

JOIN SYS.OBJECTS SO ON SO.OBJECT_ID = SP.OBJECT_ID

JOIN SYS.DM_EXEC_SESSIONS ES ON ES.SESSION_ID = DTL.REQUEST_SESSION_ID

JOIN SYS.DM_TRAN_SESSION_TRANSACTIONS TST ON ES.SESSION_ID = TST.SESSION_ID

JOIN SYS.DM_TRAN_ACTIVE_TRANSACTIONS AT ON TST.TRANSACTION_ID = AT.TRANSACTION_ID

JOIN SYS.DM_EXEC_CONNECTIONS EC ON EC.SESSION_ID = ES.SESSION_ID

CROSS APPLY SYS.DM_EXEC_SQL_TEXT(EC.MOST_RECENT_SQL_HANDLE) AS ST

WHERE

RESOURCE_DATABASE_ID = DB_ID()

ORDER BY DTL.REQUEST_SESSION_ID

Sample Result

Remarks

1. Column request_session_id display the session ID owns the lock or raise request.

If value display is -2 indicates that the request belongs to an orphaned distributed transaction.

If value display is -3 indicates that the request belongs to a deferred recovery transaction, such as, a transaction for which a rollback has been deferred at recovery because the rollback could not be completed successfully.

2. Column request_owner_type details :-

· TRANSACTION = The request is owned by a transaction.

· CURSOR = The request is owned by a cursor.

· SESSION = The request is owned by a user session.

· SHARED_TRANSACTION_WORKSPACE = The request is owned by the shared part of the transaction workspace.

· EXCLUSIVE_TRANSACTION_WORKSPACE = The request is owned by the exclusive part of the transaction workspace.

· NOTIFICATION_OBJECT = The request is owned by an internal SQL Server component.

3. To use this DMV, User required VIEW SERVER STATE permission on the server.

If you liked this post, do like on Facebook at http://www.facebook.com/mssqlfun

Reference : Rohit Garg (http://mssqlfun.com/)

DMV-12 : Retaion in Session & Transaction……..sys.dm_tran_session_transactions

sys.dm_tran_session_transactions DMV (Dynamic Management View), described by BOL as follows: http://msdn.microsoft.com/en-us/library/ms188739.aspx

Returns correlation information for associated transactions and sessions.

This DMV helps in correlating the relation between session & transaction. We can get several details about transaction by correcting it will other DMVs & system catalogs.

Query 1 : Query to check Transaction & Session details

SELECT

ST.SESSION_ID,

ST.TRANSACTION_ID,

DB_NAME(SP.DBID) DB_NAME,

CASE IS_USER_TRANSACTION

WHEN 0 THEN ‘SYSTEM TRANSACTION’

WHEN 1 THEN ‘USER TRANSACTION’ END AS USER_OR_SYSTEM_TRANSACTION,

CASE IS_LOCAL

WHEN 0 THEN ‘DISTRIBUTED TRANSACTION’

WHEN 1 THEN ‘LOCAL TRANSACTION’ END AS TRANSACTION_ORIGIN,

SP.HOSTNAME,

SP.LOGINAME,

SP.STATUS,

SP.LASTWAITTYPE,

SQLT.TEXT

FROM

SYS.DM_TRAN_SESSION_TRANSACTIONS ST

JOIN

SYS.SYSPROCESSES SP

ON SP.SPID = ST.SESSION_ID

CROSS APPLY

SYS.DM_EXEC_SQL_TEXT(SP.SQL_HANDLE) SQLT

Sample Result

Remarks

1. Meaning of int value for is_user_transaction column

a. 1 = The transaction was initiated by a user request.

b. 0 = System transaction.

2. All possible state of transactions for is_local coulmn

a. 1 = Local transaction.

b. 0 = Distributed transaction or an enlisted bound session transaction.

3. To use this DMV, User required VIEW SERVER STATE permission on the server.

4. TRANSACTION_ID > 1000 & SESSION_ID > 50 is used for user transactions & sessions.

If you liked this post, do like on Facebook at http://www.facebook.com/mssqlfun

Reference : Rohit Garg (http://mssqlfun.com/)

DMV-11 : T-Log space used by transaction……..sys.dm_tran_database_transactions

sys.dm_tran_database_transactions DMV (Dynamic Management View), described by BOL as follows: http://msdn.microsoft.com/en-us/library/ms186957.aspx

Returns information about transactions at the database level.

Some time , T-log space is hugely consumed by database & we are not aware that which transaction is culprit for this. This is useful DMV in such conditions to check T-log space used by each transaction on database.

Query 1 : Query to check Transaction log space used by each transaction will complete detail

SELECT

ST.SESSION_ID,

DT.TRANSACTION_ID,

DB_NAME(DT.DATABASE_ID) DB_NAME,

DATABASE_TRANSACTION_BEGIN_TIME TRANSACTION_BEGIN_TIME,

CASE DATABASE_TRANSACTION_TYPE

WHEN 1 THEN ‘READ/WRITE’

WHEN 2 THEN ‘READ ONLY’

WHEN 3 THEN ‘SYSTEM’ END AS TRANSACTION_TYPE,

CASE DATABASE_TRANSACTION_STATE

WHEN 1 THEN ‘NOT INITIALIZED’

WHEN 3 THEN ‘TRANSACTION NO LOG’

WHEN 4 THEN ‘TRANSACTION WITH LOG’

WHEN 5 THEN ‘TRANSACTION PREPARED’

WHEN 10 THEN ‘COMMITED’

WHEN 11 THEN ‘ROLLED BACK’

WHEN 12 THEN ‘COMMITED AND LOG GENERATED’ END AS TRANSACTION_STATE,

SP.HOSTNAME,

SP.LOGINAME,

SP.STATUS,

SP.LASTWAITTYPE,

SQLT.TEXT,

DATABASE_TRANSACTION_LOG_RECORD_COUNT LOG_RECORD_COUNT,

(DATABASE_TRANSACTION_LOG_BYTES_USED + DATABASE_TRANSACTION_LOG_BYTES_RESERVED )/1024 TOTAL_LOG_SPACE_USED_KB,

DATABASE_TRANSACTION_LOG_BYTES_USED LOG_BYTES_USED,

DATABASE_TRANSACTION_LOG_BYTES_RESERVED LOG_BYTES_RESERVED

FROM

SYS.DM_TRAN_DATABASE_TRANSACTIONS DT JOIN

SYS.DM_TRAN_SESSION_TRANSACTIONS ST

ON DT.TRANSACTION_ID=ST.TRANSACTION_ID

JOIN

SYS.SYSPROCESSES SP

ON SP.SPID = ST.SESSION_ID

CROSS APPLY

SYS.DM_EXEC_SQL_TEXT(SP.SQL_HANDLE) SQLT

WHERE

DT.TRANSACTION_ID > 1000 AND ST.SESSION_ID >50

Sample Result

Remarks

1. Meaning of int value for database_transaction_type column

· 1 = Read/write transaction

· 2 = Read-only transaction

· 3 = System transaction

2. All possible state of transactions for database_transaction_state coulmn

· 1 = The transaction has not been initialized.

· 3 = The transaction has been initialized but has not generated any log records.

· 4 = The transaction has generated log records.

· 5 = The transaction has been prepared.

· 10 = The transaction has been committed.

· 11 = The transaction has been rolled back.

· 12 = The transaction is being committed. In this state the log record is being generated, but it has not been materialized or persisted.

3. To use this DMV, User required VIEW SERVER STATE permission on the server.

4. If column database_transaction_begin_time has NULL then transaction is read only otherwise read write.

5. TRANSACTION_ID > 1000 & SESSION_ID > 50 is used for user transactions & sessions.

 

If you liked this post, do like on Facebook at http://www.facebook.com/mssqlfun

Reference : Rohit Garg (http://mssqlfun.com/)

LinkedIn Auto Publish Powered By : XYZScripts.com
%d bloggers like this: