Bewerken

Delen via


MSSQLSERVER_1205

Applies to: SQL Server Azure SQL Database Azure SQL Managed Instance

Details

Attribute Value
Product Name SQL Server
Event ID 1205
Event Source MSSQLSERVER
Component SQLEngine
Symbolic Name LK_VICTIM
Message Text Transaction (Process ID %d) was deadlocked on %.*ls resources with another process and has been chosen as the deadlock victim. Rerun the transaction.

Explanation

Resources are accessed in conflicting order on separate transactions, causing a deadlock. For example:

  • Transaction1 updates Table1.Row1, while Transaction2 updates Table2.Row2
  • Transaction1 tries to update Table2.Row2 but is blocked because Transaction2 hasn't yet committed and hasn't released its locks
  • Transaction2 now tries to update Table1.Row1 but is blocked because Transaction1 hasn't committed and hasn't released its locks
  • A deadlock occurs because Transaction1 is waiting for Transaction2 to complete, but Transaction2 is waiting for Transaction1 to complete.

The system will detect this deadlock and will choose one of the transactions involved as a 'victim'. It will then issue this error message, rolling back the victim's transaction. For detailed information, see Deadlocks.

User Action

Deadlocks are in most cases application-related issues and require application developers to make code changes. One approach when you receive error 1205 is to execute the queries again. See this blog for an example of how to retry - handle the deadlock and re-execute the query: Deadlock Simulator app for Developers: How to Handle a SQL Deadlock issue in Your App

You can also revise the application to avoid deadlocks. The transaction that was chosen as a victim can be retried and will likely succeed, depending on what operations are being executed simultaneously.

To prevent or avoid deadlocks from occurring, consider having all transactions access rows in the same order (Table1, then Table2). This way, although blocking might occur, a deadlock will be avoided.

For more information, see Handling Deadlocks and Minimizing Deadlocks.