Showing posts with label DBCC. Show all posts
Showing posts with label DBCC. Show all posts

Monday 21 August 2017

Dealing with physical database corruptions

Beyond all doubt, not every single database in the world has a DBA dedicated to monitoring it 24x7 hours. Moreover, many database environments are unwittingly implemented/installed where there was no consideration to meet the basic software and hardware minimum requirements. This situation is compounded by the fact that many companies do not put much attention on databases from the beginning, maybe because at first the database are quite small and serving not many transactions per second. Unsurprisingly, all databases are becoming bigger and bigger with the passing of time so it is not rocket science to foresee that everything will get worse in terms of performance and physical integrity. In this new context, it is much more critical to have everything in place so as to prevent databases from getting damaged. However, many times it is too late when one realises that the database got damaged because of poor implementation. The overwhelmingly majority of physical corruption issues are not sparked by SQL Server on its own, but poor hardware implementation. Thus, if you find yourself working with databases struggling this problem, I wholeheartedly recommend reinstalling the whole server from scratch taking minimum requirements on board.
After having a good implementation of a database server, it is of paramount importance to carry out maintenance tasks at least once a month, and it should include executing full DBCC CHECKDB, and if possible DBCC CHECKDB WITH PHYSICAL_ONLY option once a week. It is also understandable that DBCC CHECKDB makes intensive use of resources, especially CPU and Disk, nevertheless it is possible to use MAXDOP option with DBCC CHECKDB to limit CPU usage and ease the pain. So, nowadays there is no feasible excuse to avoid executing it regularly.

Having said that, I would like to share with you some steps to follow in order to try to fix a physical database corruption:

1. To begin with, try to back your database up.
2. Set your database to SINGLE_USER mode, and then execute DBCC CHECKDB WITH PHYSICAL_ONLY to determine whether physical corruption exists or not.


If there is physical corruption, SQL Server will display some errors like these:

Msg 8909, Level 16, State 1, Line 1
Table error: Object ID 0, index ID -1, partition ID 0, alloc unit ID 0 (type Unknown),
page ID (1:219356) contains an incorrect page ID in its page header.
The PageId in the page header = (0:0).
Msg 8928, Level 16, State 1, Line 1
Object ID 501576825, index ID 1, partition ID 72057594040549376,
alloc unit ID 72057594044940288 (type In-row data): Page (1:1408252) could not be processed.

3. Sometimes only non-clustered indexes are damaged, and luckily in this case, you just need to recreate those indexes. If you want to figure out which indexes are damaged so that you can recreate them, you can have a look at this article I wrote many years ago.
4. However, if the clustered index or heap structure is damaged then you may need set the database to EMERGENCY mode and execute DBCC CHECKDB with REPAIR_ALLOW_DATA_LOSS option which may not always be the best option for bringing a database to a physically consistent state but when the clustered index or heap structure of a table is heavily damaged there is no other option, just do it at the cost of some data loss. It is also worth noting that if you want to find out which objects are damaged so that only execute DBCC CHECKDB repair process on them then you can also check out the same article. Here is this example, we are going to execute on the whole database.


5. Having executed DBCC CHECKDB with REPAIR_ALLOW_DATA_LOSS option does not always ensure that the database will be repaired, however, many times it is enough. So, after getting your database repaired, try to go to bottom of the problem, and fix the problem at hardware level because it is most likely that disk storage is not working properly. Bearing in mind that database corruption issue is just the tip of the iceberg so consider DBCC CHECKDB as a quick fix, not a final solution.
6. Finally, you should take a full database backup.

That is all for now, I hope you find this post helpful. Let me know any remarks you may have. Stay tuned.

Monday 25 January 2016

How to fix: The In-row data RSVD page count for object ‘SchemaName.ObjectName’, index ID ‘x’, partition ID ‘xxxxxxxxxxxxxx’

Undoubtedly, we all know that SQL Server 2000 is now out of compliance with Microsoft. Surprisingly, not all companies have their databases migrated, so in that sense, many of them are planning to upgrade to one of the latest version of SQL Server (SQL Server 2014/2016). This post pretends to give a recommendation that you may take it as a good practice in order to avoid inconsistency errors of object metadata for the upgraded databases. The inconsistency errors what I am talking about can be found when you run DBCC CHECKDB command to check the logical and physical integrity:

And here the following inconsistency error is likely to appear:
The In-row data RSVD page count for object ‘SchemaName.ObjectName’, index ID 0, partition ID 75863107960832, alloc unit ID 75863107960832 (type In-row data) is incorrect.


CHECKDB found 0 allocation errors and 1 consistency errors in table ‘SchemaName.ObjectName’ (object ID 1157579162). CHECKDB found 0 allocation errors and 1 consistency errors in database ‘MyDatabase’.
The output message above tells us that there is one LOGICAL consistency error inside the database "MyDatabase" because there are pages and row count inaccuracies for the table which ID is 1157579162. Luckily, errors of this sort can be fixed by running DBCC UPDATEUSAGE command which reports and corrects the inaccuracies. Only after successfully running DBCC UPDATEUSAGE(0) command in the database context will you be able to see the following messages indicating we got all inaccuracies fixed.
DBCC UPDATEUSAGE: Usage counts updated for table 'SchemaName.ObjectName' (index 'ObjectName' , partition 1):

        RSVD pages (In-row Data): changed from (-275) to (145) pages.

DBCC execution completed. If DBCC printed error messages, contact your system administrator.

Finally, we are now able to run again DBCC CHECKDB without any error. The recommendation is that so rapidly have you upgraded your databases (from SQL Server 2000 to superior version) that you must execute DBCC UPDATEUSAGE so as to avoid those errors (or more critical issues). Thanks for reading!

Wednesday 24 September 2014

How to use DBCC CHECKPRIMARYFILE to attach several SQL Server databases in seconds

A few hours ago my tip has been published at about “How to use DBCC CHECKPRIMARYFILE to attach several SQL Server databases in seconds”. You can read it at . Thanks for reading!

Thursday 21 August 2014

Error 601: Could not continue scan with NOLOCK due to SQL Server data movement

Just one hour ago one of my tips was published at related to “Error 601: Could not continue scan with NOLOCK due to SQL Server data movement”. You can read it at . Thanks for reading!
HELLO, I'M PERCY REYES! — a book lover, healthy lifestyle lover... I've been working as a senior SQL Server Database Administrator (DBA) for over 20 years; I'm a three-time awarded Microsoft Data Platform MVP. I'm currently doing a PhD in Computer Science (cryptography) at Loughborough University, England — working on cryptographic Boolean functions, algorithmic cryptanalysis, number theory, and other algebraic aspects of cryptography. READ MORE