Oracle LogMiner Agent Changelog
Release notes
2.3.2.1
Released: November 11, 2026
-
A region of the cache whose flush to the disk fails stays dirty and is flushed again, instead of being forgotten after that one failed attempt;
2.3.2.0
Released: November 10, 2026
-
GSSD-1483: Target:
CHAR,VARCHARandVARCHAR2columns are created in characters (n CHAR) in target tables, CDC tables and the bulk staging table, instead of followingNLS_LENGTH_SEMANTICS(bytes by default), which failed a snapshot in a loop withORA-12899once a line held accented letters; existing CDC tables are left alone; -
GSSD-1483: Target: the blank padding of a key no longer reaches the values written, instead of three accented letters padded to
CHAR(10 CHAR)making 13 bytes and failingORA-12899on aCHAR(10 BYTE)column; -
GSSD-1346: Target: the bulk flush matches staged keys one NULL pattern at a time (
IS NULLor=), so a flush of 5,000 staged keys on a 970K-row table deletes in 0.04 to 0.11 s instead of 136 s;
2.3.1.4
Released: November 2, 2026
-
PARALLEL is written as a real optimizer hint where it belongs and dropped where it does not, instead of a plain comment that always ran serially while the text suggested otherwise;
2.3.1.3
Released: October 29, 2026
-
CDC: an
IS NULLcondition in a redoWHEREclause is matched whole and anchored, so aHEXTORAWliteral no longer costs quadratic time and an "is null" inside a string literal is no longer read as a null column with the real condition lost;
2.3.1.1
Released: October 27, 2026
-
CDC: the local cache is drained up to its head in one poll instead of one message per poll, so a backlog reaches the target faster;
-
GSSD-1456: resetting the entities deletes the shared read checkpoint, so the reader no longer resumes from the old position;
-
Metrics: a page’s read time no longer includes the wait for the target to consume the previous page;
-
Target: an object field from a NoSQL source is written as its JSON text instead of
{name=Ada, age=36}, and an array that holds a map or a nested array is written as a JSON array instead of[Ljava.lang.Object;@…;
2.3.0.3
Released: October 21, 2026
-
GSSD-1245: Rows written after another session’s commit with a higher SCN are no longer skipped. LogMiner returns rows in redo order, which under concurrency is not SCN order, and the reader dropped every later row below the last commit it had read in the same pass: committed inserts, updates, and deletes never reached the target;
-
GSSD-1245: The rows a
ROLLBACK TO SAVEPOINTwrites are kept when LogMiner returns them with an unresolved transaction id (…FFFFFFFF): they are matched to the only open transaction in their undo slot. Before, they were dropped, so the target kept the rows the savepoint undid and lost the ones it restored; -
GSSD-1245: New
logminerMaxWindowScnssetting in the agent’s specific configuration: it caps the span of a mining window, so a source that writes redo faster than a window is mined no longer gets wider and slower windows each time (1 to 2 million SCNs and over an hour each were seen). The redo above the cap is mined by the next polls, and a capped window logs how many SCNs of written redo are left.0, the default, leaves windows uncapped; -
GSSD-1245: An insert on a table with LOB columns, also one written through a LOB locator, is cancelled against the
DELETEaROLLBACK TO SAVEPOINTwrites to undo it. Before, thatDELETElogged "Skipping DELETE … will remain on the target" although nothing was lost; -
GSSD-1245: A
DELETEwithout a usable before-image is no longer reported as a lost delete while its transaction may still roll back: the sample is a warning, rows generated by a rollback are not sampled, and the line that says a row stayed on the target, "Skipping DELETE on table", is the error;
2.3.0.1
Released: October 15, 2026
-
Bulk operations: an insert and a delete, or two updates, of a row keyed by binary or array values in one batch are merged instead of leaving the row on the target or failing with a duplicate key;
-
The agent hands its pooled database back when its pipeline is deleted or detached and when its initialization fails. Before, the next pipeline writing to the same database could fail in a loop with
NoClassDefFoundError;
2.3.0
Released: October 13, 2026
-
Agent compatibility upgrade to Gluesync 2.3;
-
The source cache saves the corrected read position after a recovery scan, so a restart no longer starts reading mid-record or discards records written after a torn record;
-
The type matrix no longer lists INT, LONG and SHORT, which Oracle does not have, as target types;
-
Columns are discovered in their table order instead of the order the query plan happened to produce;
-
Discovery returns schemas and tables in alphabetical order, so Query Studio pages no longer skip or repeat objects;