Hyperledger Fabric databases use multi-version concurrency control (MVCC) to avoid double-spending and data inconsistency.

When the same state is updated, a new version of the record overwrites the old version. If there are concurrent requests to update the same key in a block, an MVCC_READ_CONFLICT error might be generated.

To reduce MVCC errors for transfer, mint, burn, and hold operations, you can enable the MVCC optimization for token chaincode. This optimization works on Oracle Blockchain Platform only. By default, the optimization is disabled. To enable the optimization, complete the applicable following step.

• CLI: Specify the Boolean -m or --enable_mvcc_optimization parameter with the ochain init command. By default, the parameter is set to false. To enable the optimization, add -m true to the ochain init command line.
• Visual Studio Code: When you create a chaincode, select Enable MVCC optimization on the Create Chaincode window.

To use the optimization with chaincode created in previous versions of Blockchain App Builder, complete the following steps:

1. After you install the latest version of Blockchain App Builder, upgrade the chaincode as described in Upgrading Chaincode Projects in the CLI and Upgrading Chaincode Projects in Visual Studio Code.
2. Edit the .ochain.json file in the root folder of the chaincode to set enableMvccOptimization to true.
3. Synchronize the chaincode, which adds the optimization and creates two new folders in the root folder of the the chaincode: statedb and tokens. For more information about synchronization, see Synchronize Specification File Changes With Generated Source Code and Synchronize Specification File Changes With Generated Source Code.

Other methods to work around MVCC_READ_CONFLICT errors including having the client application retry requests when this error is generated, or using a queue to capture concurrent requests before they are sent to the blockchain network. For more information, see Read-Write set semantics in the Hyperledger Fabric documentation.