corda/docs/source/api-persistence.rst

API: Persistence

Corda offers developers the option to expose all or some part of a contract state to an Object Relational Mapping (ORM) tool to be persisted in a RDBMS. The purpose of this is to assist vault development by effectively indexing persisted contract states held in the vault for the purpose of running queries over them and to allow relational joins between Corda data and private data local to the organisation owning a node.

The ORM mapping is specified using the Java Persistence API (JPA) as annotations and is converted to database table rows by the node automatically every time a state is recorded in the node's local vault as part of a transaction.

Note

Presently the node includes an instance of the H2 database. H2 database is supported for development purposes, and we have certified R3 Corda to work against SQL Server 2017 and Azure SQL. PostgreSQL 9.6 is supported preliminarily. Other databases will be officially supported very soon. Much of the node internal state is also persisted there. You can access the internal H2 database via JDBC, please see the info in "node-administration" for details.

Schemas

Every ContractState can implement the QueryableState interface if it wishes to be inserted into the node's local database and accessible using SQL.

../../core/src/main/kotlin/net/corda/core/schemas/PersistentTypes.kt

The QueryableState interface requires the state to enumerate the different relational schemas it supports, for instance in cases where the schema has evolved, with each one being represented by a MappedSchema object return by the supportedSchemas() method. Once a schema is selected it must generate that representation when requested via the generateMappedObject() method which is then passed to the ORM.

Nodes have an internal SchemaService which decides what to persist and what not by selecting the MappedSchema to use.

../../node/src/main/kotlin/net/corda/node/services/api/SchemaService.kt

../../core/src/main/kotlin/net/corda/core/schemas/PersistentTypes.kt

The SchemaService can be configured by a node administrator to select the schemas used by each app. In this way the relational view of ledger states can evolve in a controlled fashion in lock-step with internal systems or other integration points and not necessarily with every upgrade to the contract code. It can select from the MappedSchema offered by a QueryableState, automatically upgrade to a later version of a schema or even provide a MappedSchema not originally offered by the QueryableState.

It is expected that multiple different contract state implementations might provide mappings to some common schema. For example an Interest Rate Swap contract and an Equity OTC Option contract might both provide a mapping to a common Derivative schema. The schemas should typically not be part of the contract itself and should exist independently of it to encourage re-use of a common set within a particular business area or Cordapp.

MappedSchema offer a family name that is disambiguated using Java package style name-spacing derived from the class name of a schema family class that is constant across versions, allowing the SchemaService to select a preferred version of a schema.

The SchemaService is also responsible for the SchemaOptions that can be configured for a particular MappedSchema which allow the configuration of a database schema or table name prefixes to avoid any clash with other MappedSchema.

Note

It is intended that there should be plugin support for the SchemaService to offer the version upgrading and additional schemas as part of Cordapps, and that the active schemas be configurable. However the present implementation offers none of this and simply results in all versions of all schemas supported by a QueryableState being persisted. This will change in due course. Similarly, it does not currently support configuring SchemaOptions but will do so in the future.

Custom schema registration

Custom contract schemas are automatically registered at startup time for CorDapps. The node bootstrap process will scan for schemas (any class that extends the MappedSchema interface) in the plugins configuration directory in your CorDapp jar.

For testing purposes it is necessary to manually register the packages containing custom schemas as follows:

• Tests using MockNetwork and MockNode must explicitly register packages using the cordappPackages parameter of MockNetwork
• Tests using MockServices must explicitly register packages using the cordappPackages parameter of the MockServices makeTestDatabaseAndMockServices() helper method.

Note

Tests using the DriverDSL will automatically register your custom schemas if they are in the same project structure as the driver call.

Object relational mapping

The persisted representation of a QueryableState should be an instance of a PersistentState subclass, constructed either by the state itself or a plugin to the SchemaService. This allows the ORM layer to always associate a StateRef with a persisted representation of a ContractState and allows joining with the set of unconsumed states in the vault.

The PersistentState subclass should be marked up as a JPA 2.1 Entity with a defined table name and having properties (in Kotlin, getters/setters in Java) annotated to map to the appropriate columns and SQL types. Additional entities can be included to model these properties where they are more complex, for example collections, so the mapping does not have to be flat. The MappedSchema must provide a list of all of the JPA entity classes for that schema in order to initialise the ORM layer.

Several examples of entities and mappings are provided in the codebase, including Cash.State and CommercialPaper.State. For example, here's the first version of the cash schema.

../../finance/src/main/kotlin/net/corda/finance/schemas/CashSchemaV1.kt

Identity mapping

Schema entity attributes defined by identity types (AbstractParty, Party, AnonymousParty) are automatically processed to ensure only the X500Name of the identity is persisted where an identity is well known, otherwise a null value is stored in the associated column. To preserve privacy, identity keys are never persisted. Developers should use the IdentityService to resolve keys from well know X500 identity names.

JDBC session

Apps may also interact directly with the underlying Node's database by using a standard JDBC connection (session) as described by the Java SQL Connection API

Use the ServiceHub jdbcSession function to obtain a JDBC connection as illustrated in the following example:

../../node/src/test/kotlin/net/corda/node/services/database/HibernateConfigurationTest.kt

JDBC sessions can be used in Flows and Service Plugins (see "flow-state-machines")

The following example illustrates the creation of a custom corda service using a jdbcSession:

../../docs/source/example-code/src/main/kotlin/net/corda/docs/CustomVaultQuery.kt

which is then referenced within a custom flow:

../../docs/source/example-code/src/main/kotlin/net/corda/docs/CustomVaultQuery.kt

Database Migration

As a database migration tool, we use the open source library liquibase <http://www.liquibase.org/>.

Migration is enabled by specifying true in the database.runMigration node configuration setting (default behaviour is false). When enabled, the database state is checked, and updated during node startup.

The default behaviour (database.runMigration=false) is to just check the database state, and fail if it is not up to date. To bring the database to the correct state we provide an advanced migration tool. See below for details.

For example, if migration is enabled, after deploying a new version of the code that contains database migrations (see example below for a possible scenario), they are executed at that point (during startup).

Possible database changes range from schema changes to data changes. (The database changes are grouped together in changesets. See the example below.).

About Liquibase

Liquibase will create a table called DATABASECHANGELOG, that will store information about each executed change (like timestamp, description, user, md5 hash so it can't be changed, etc). This table will be used every time a migration command is run to determine what changesets need to be applied. Changesets should never be modified once they were executed. Any correction should be applied in a new changeset. We can also "tag" the database at each release to make rollback easier.

Database changes are maintained in several xml files per MappedSchema, so that only migrations corresponding to the node’s configured schemas are run. The migration file(s) for all MappedSchemas are dynamically included in the global changelog, as long as they are present on the classpath and are either explicitly declared in the MappedSchema implementation, or follow a naming convention based on the MappedSchema name. (The migration tool that we provide can generate liquibase files with the correct name for a schema)

Our convention is to maintain a "master" changelog file per MappedSchema which will include "version" changelogs. By following our versioning convention, and using the node-info schema as an example, if there are any database changes for release 12, the changes will be added to a new file called: node-info.changelog-v12.xml which has to be included in node-info.changelog-master.xml.

Example:

Let's suppose that at some point, at version 12, there is a need to add a new column: contentSize to the DBAttachment entity.

This means we have to:

• In the source code, add the contentSize property and map it to a new column.
• create the column in the node_attachments table.
• Run an update to set the size of all existing attachments, to not break the code that uses the new property

class DBAttachment(
    ...
    @Column(name = "content")
    @Lob
    var content: ByteArray,

    //newly added column
    @Column(name = "content_size")
    var contentSize: Int,
    ...
)

The DBAttachment entity is included in the NodeServicesV1 schema, so we create a file node-services.changelog-v12.xml with this changeset:

<changeSet author="R3.Corda" id="add content_size column">
    <addColumn tableName="node_attachments">
        <column name="content_size" type="INT"/>
    </addColumn>
    <update tableName="node_attachments">
        <column name="content_size" valueComputed="length(content)"/>
    </update>
    <rollback>
        <dropColumn tableName="node_attachments" columnName="content_size"/>
    </rollback>
</changeSet>

And include it in `node-services.changelog-master.xml`:

<databaseChangeLog>
    <!--the original schema-->
    <include file="migration/node-services.changelog-init.xml"/>

    <!--migrations from previous releases-->
    <include file="migration/node-services.changelog-v4.xml"/>
    <include file="migration/node-services.changelog-v7.xml"/>

    <!--added now-->
    <include file="migration/node-services.changelog-v12.xml"/>
</databaseChangeLog>

By adding the rollback script, we give users the option to revert to an older version of the software.

An easy way to manage the db version is to tag it on every release (or on every release that has migrations) <http://www.liquibase.org/documentation/changes/tag_database.html>

Usage:

Configurations:

• To enable migration at startup, set:

  • database.runMigration = true // false by default.

Migration tool:

The Migration tool will be distributed as a standalone jar file, with the following options:

Option	Description
--help	Print help message
--mode	Either 'NODE' or 'DOORMAN'. By default 'NODE'
--base-directory	The node or doorman directory
--config-file	The name of the config file. By default 'node.conf' for a simple node and 'network-management.conf' for a doorman.
--doorman-jar-path	The path to the doorman fat jar
--create-migration-sql-for-cordapp	Create migration files for a CorDapp. You can specify the fully qualified of the MappedSchema class. If not specified it will generate foll all schemas that don't have migrations. The output directory is the base-directory, where a migration folder is created.
--dry-run	Output the database migration to the specified output file. The output directory is the base-directory. You can specify a file name or 'CONSOLE' if you want to send the output to the console.
--execute-migration	This option will run the db migration on the configured database
--release-lock	Releases whatever locks are on the database change log table, in case shutdown failed.

It is intended to be used by R3 Corda node administrators. Currently it has these features : - it allows running the migration on the database (--execute-migration ) - offers the option to inspect the actual sql statements that will be run as part of the current migration (--dry-run ) - can be used to release the migration lock (--release-lock) - when a CorDapp released by the open source community is ready to be deployed on a production node, using this tool it can be "upgraded" (--create-migration-sql-for-cordapp). See below for details.

CorDapps:

CorDapp developers who decide to store contract state in custom entities can create migration files for the MappedSchema they define.

There are 2 ways of associating a migration file with a schema:

1. By overriding val migrationResource: String and pointing to a file that needs to be in the classpath
2. By putting a file on the classpath in a migration package whose name is the hyphenated name of the schema. (All supported file extensions will be appended to the name)

CorDapp developers can use any of the supported formats (xml, sql, json, yaml) for the migration files they create.

In case CorDapp developers distribute their CorDapps with migration files, these will be automatically applied when the CorDapp is deployed on an R3 Corda node. If they are deployed on a standard ("Open source") Corda node, then the migration will be ignored, and the database tables will be generated by Hibernate.

In case CorDapp developers don't distribute a CorDapp with migration files, then the organisation that decides to deploy this CordApp on an R3 Corda ("Enterprise Blockchain") node has the responsibility to manage the database.

The following options are available:

1. In case the organisation is running a demo or trial node on the default H2 database, then the CorDapp will just work when deployed by relying on the migration tool provided by hibernate, which is not intended for production.
2. In case the organisation is running a production node (with live data) on an enterprise database, then they will have to manage the database migration for the CorDapp.

   These are the steps to do this:

   • deploy the CorDapp on your node (copy the jar in the cordapps folder)
   • find out the name of the MappedSchema containing the new contract state entities and hyphenate it. For example:net.corda.finance.schemasCommercialPaperSchemaV1
   • call the migration tool java -jar migration-tool.jar --base-directory path_to_node --create-migration-sql-for-cordapp net.corda.finance.schemasCommercialPaperSchemaV1
   • this will generate a file called commercial-paper-schema-v1.changelog-master.sql in a folder called migration in the base-directory
   • in case you don't specify the actual MappedSchema name, the tool will generate one sql file for each schema defined in the CorDapp
   • inspect the file(s) to make sure it is correct
   • create a jar with the migration folder (by convention it could be named: originalCorDappName-migration.jar), and deploy this jar together with the CorDapp
   • To make sure that the new migration will be used, the migration tool can be run in a dry-run mode and inspect the output file