Configure Tessera transaction manager
The only mandatory parameter for running a minimal Transaction Manager is the location of the configuration file to use.
-configfile <path> argument to specify the location of the configuration file.
Other CLI arguments can be passed, and details of these commands can be found in their respective pages - particularly around key vaults and key generation.
By default, Tessera uses an H2 file-based database, but any JDBC compatible database can be used.
To do this, add the necessary drivers to the classpath, and run the
com.quorum.tessera.Launcher class, like the following:
java -cp some-jdbc-driver.jar:/path/to/tessera-app.jar:. com.quorum.tessera.Launcher
For example, to use Oracle database:
java -cp ojdbc7.jar:tessera-app.jar:. com.quorum.tessera.Launcher -configfile config.json
Some DDL scripts have been provided for more popular databases, but feel free to adapt these to whichever database you wish to use.
Flavours of transaction manager
For advanced users, you may decide on certain options for the transaction manager, or to disable other parts.
The default transaction manager comes with the standard options most setups will use, but other versions are as follows:
- gRPC communication (experimental)
- Non-remote only enclaves (named “tessera-simple”)
These must be built from source and can be found inside the
Data recovery (Legacy)
Tessera contains functionality to request transactions from other nodes in the network; this is useful if the database is lost or corrupted somehow.
However, depending on the size of the network and the number of transactions made between peers, this can put heavy strain on the network resending all the data.
How to enable
The data recovery mechanism is intended to be a “switch-on” feature as a startup command. The times when you will need this will be known prior to starting the application (usually after a disaster event). When starting Tessera, simply add the following property to the startup command:
-Dspring.profiles.active=enable-sync-poller. This should go before any jar or class definitions, for example
java -Dspring.profiles.active=enable-sync-poller -jar tessera.jar -configfile config.json.
How it works
The data recovery procedure works by invoking a “resend request” to each new node it sees in the network. This request will cause the target node to resend each of its transactions to the intended recipient, meaning they will again save the transaction in their database.
The target node will not send back transactions as a response the request in order to ensure that a malicious node cannot get access to the transactions. example: anyone can send a request for a particular key, but it will mean that the node that holds that key will receive the transactions, not the node making the request. In normal usage, the node making the request and the node holding the public key are the same.
Data recovery (Enhanced from
Privacy Enhancement release)
Due to the interdependence between Party Protection and PSV transactions (the existence and validation of ACOTHs), transactions cannot just be accepted but need to be recovered in the appropriate manner so that they don’t get wrongly rejected.
To do this we introduce a separate persistence unit called
tessera-recover which consists a number of tables so that incoming history transactions received can be sorted before synchronising to the main database
The recovery process will include these steps:
- Request - the transaction manager that runs in recovery mode will send resend requests to other nodes in the network and wait for requests to be completed. The requested nodes will attempt to resend the transactions they have for the recovery node in batches (rather than singles compared to the legacy resend process). Transactions received are persisted to a separate staging database unit, and the batch request will be considered successful once the requested node finishes sending the transactions.
- Stage - All transactions in the staging area will be sorted by dependency. This is done by executing a special staging query multiple times, until all transactions in the staging area are sorted and validated.
- Sync - Once the staging process is done, the transactions are copied to the main database - by utilising the normal /push. During the sync, enhanced-privacy transactions are checked and validated the same way they were before.
Tessera recovery process will stop and shutdown once the above steps are executed. Each stage result will be reported as SUCCESS(0), PARTIAL_SUCCESS(1), or FAILURE(2). The result code would be useful for scripting purpose (for example automatically start Tessera in normal mode if recovery successfully completed)
To trigger the recovery process, Tessera will need to be started in recovery mode by using the command line
tessera -r or tessera --recover
During the recovery process, Tessera won’t accept any new enhanced-privacy transactions but will continue to accept ‘standard’ private transactions.