CDC-8001 Snapshot too old

Code: CDC-8001 (SNAPSHOT_TOO_OLD). Raised while a snapshot reads a table, after the retries described below.

What the connector observed

A snapshot chunk's SELECT ... AS OF SCN failed with ORA-01555 or ORA-08181, and kept failing after the chunk was halved down to 1,000 rows with a fresh SCN each time. The message names the table.

Why it stopped rather than continued

Skipping the chunk would leave its rows out of the snapshot. Reading it without AS OF SCN would break the ordering with streamed changes that keeps consumers correct.

Confirm the cause

SELECT name, value FROM v$parameter WHERE name IN ('undo_retention', 'undo_management');
SELECT tuned_undoretention, maxquerylen, ssolderrcnt FROM v$undostat
WHERE ROWNUM <= 12 ORDER BY begin_time DESC;

SSOLDERRCNT above zero confirms snapshot-too-old errors in that period.

Recover

Raise UNDO_RETENTION and give the undo tablespace room for it, or lower cdc.snapshot.chunk.rows, then restart the task. The snapshot resumes at the chunk that failed; chunks already published are not read again.

Restart the task with:

curl -s -X POST "$CONNECT/connectors/$NAME/restart?includeTasks=true&onlyFailed=true"