Advertisements
Home > Informatica, SQL Errors > Informatica | Thread: WRITER_1_*_1, Message Code: WRT_8229, Native Error: 3621, Violation of PRIMARY KEY constraint

Informatica | Thread: WRITER_1_*_1, Message Code: WRT_8229, Native Error: 3621, Violation of PRIMARY KEY constraint


Today while working with an Informatica mapping, I faced a strange issue. The image below shows the design of the INFA mapping:

INFA_Mapping

Here I’m Acquiring a new table with ~500k records and the table is also very fat with PK as a GUID column. Its just a plain data pull with a simple SELECT query, no JOINs, UNIONs, etc. But while running the Workflow it gave me PK Violation error, as shown below:

Severity: ERROR
Timestamp: 5/30/2013 7:22:54 PM
Node: INFA_NODE_SERVERNAME
Thread: WRITER_1_*_1
Process ID: 8216
Message Code: WRT_8229
Message: Database errors occurred: 
Microsoft OLE DB Provider for SQL Server: The statement has been terminated.
SQL State: 01000	Native Error: 3621
State: 1	Severity: 0
SQL Server Message: The statement has been terminated.

Microsoft OLE DB Provider for SQL Server:
Violation of PRIMARY KEY constraint 'PK_tblTableName'. 
Cannot insert duplicate key in object 'dbo.tblTableName'.
The duplicate key value is (2ea8b6b9-e505-4ef1-a385-0cf9143d2cfd).

SQL State: 23000	Native Error: 2627
State: 1	Severity: 14
SQL Server Message: Violation of PRIMARY KEY constraint 'PK_tblTableName'. 
Cannot insert duplicate key in object 'dbo.tblTableName'. 
The duplicate key value is (2ea8b6b9-e505-4ef1-a385-0cf9143d2cfd).

Database driver error...
Function Name : Execute Multiple
SQL Stmt : INSERT INTO tblTableName

This was strange because the error was for the GUID PK column, and there is no reason for duplicate values here. As I’ve already taken care of INSERTs & UPDATEs by ROUTER Transformation, which can be seen in the image above. I tried to find the root cause of the error, BINGed/GOOGLEd a lot, but no luck.

Then I checked with an experienced person in my team about this error. He immediately asked me if I’ve added any “WITH (NOLOCK)” option in my query with the Source Table. I said yes, so he asked me to remove it and try, as they had faced similar issues with NOLOCK option before. I removed it and the Workflow ran successfully 🙂 .

I knew that with NOLOCK option I’m doing Dirty Reads, and had added it just to avoid locking/blocking at the Source end. But I was not aware that these Dirty Reads means any kind of data, which can also be duplicate. Thus by adding NOLOCK option with the Table, which is also going through lot of changes, we may allow data to be read more than once. This may be due to Data Movement, Uncommitted Data or Page Splits on the Source table during our reads, where we may be reading the data before and after the Page-Splits.

So, be careful while using NOLOCK option while designing your queries.

For more information check following blog: http://sqlmag.com/database-development/quaere-verum-clustered-index-scans-part-iii

Advertisements
  1. April 19, 2016 at 5:13 pm

    GUID generation is unique and thread safe, but when you read the rows by using nolock hint, then irrespective of datatype you might get duplicate rows, if at the same time the source table is going through any updates. Check the reason of Page Splits I’ve given above.

  2. Amit
    April 19, 2016 at 5:09 pm

    Hi Manoj,

    I face the same situation and the solution given by you worked, but still I cam not clear what was the reason, GUID generation is unique and thread safe. Can you please give some idea….

    Regards
    Amit Srivastava

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: