Tungsten Replicator 4.0.1 is a bugfix release that contains critical fixes and improvements to the Tungsten Replicator 4.0.0 release.
The following issues may affect the operation of Tungsten Replicator and should be taken into account when deploying or updating to this release.
When running the trepctl reset command on a master, DDL statements could be placed into the binary log that would delete corresponding management tables within slaves. Binary logging for these operations is now suppressed for these operations.
When replicating from Oracle to datawarehouses, the heartbeat
table data would be replicated as a standard table row events,
leading to replicator failure when the corresponding heartbeat
not exist. HP Vertica and Amazon Redshift batch loaders have
been updated to create the appropriate staging table during
The filename for the CSV generated within the batch applier was URI-encoded, leading to odd filenames when the file was ultimately uploaded to HDFS causing a replication failure. The names are now correctly escaped and reformatted correctly so that the URI encoded format is not used.
The timezone information for the batch applier used by Hadoop,
HP Vertica, and Amazon Redshift deployments was forced to use
the UTC timezone. The code has been updated to support the
correct timezone as configured in
The timezone information for the
would be incorrect when using parallel replication with a server
timezone other than GMT.