Examples of org.exist.storage.journal.Journal

org.exist.storage.journal.Journal

Manages the journalling log. The database uses one central journal for all data files. If the journal exceeds the predefined maximum size, a new file is created. Every journal file has a unique number, which keeps growing during the lifetime of the db. The name of the file corresponds to the file number. The file with the highest number will be used for recovery. A buffer is used to temporarily buffer journal entries. To guarantee consistency, the buffer will be flushed and the journal is synched after every commit or whenever a db page is written to disk. Each entry has the structure:

[byte: entryType, long: transactionId, short length, byte[] data, short backLink]

entryType is a unique id that identifies the log record. Entry types are registered via the {@link org.exist.storage.journal.LogEntryTypes} class.
transactionId: the id of the transaction that created the record.
length: the length of the log entry data.
data: the payload data provided by the {@link org.exist.storage.journal.Loggable} object.
backLink: offset to the start of the record. Used when scanning the log file backwards.

@author wolf

     */
    public TransactionManager(BrokerPool pool, File dataDir, boolean transactionsEnabled) throws EXistException {
      this.pool = pool;
        enabled = transactionsEnabled;
        if (enabled)
            {journal = new Journal(pool, dataDir);}
        final Boolean groupOpt = (Boolean) pool.getConfiguration().getProperty(PROPERTY_RECOVERY_GROUP_COMMIT);
        if (groupOpt != null) {
            groupCommit = groupOpt.booleanValue();
            if (LOG.isDebugEnabled())
                {LOG.debug("GroupCommits = " + groupCommit);}

Examples of org.exist.storage.journal.Journal

Related Classes of org.exist.storage.journal.Journal