DanielLeens commented on issue #10508:
URL: https://github.com/apache/seatunnel/issues/10508#issuecomment-6415640292

   Thanks for checking the newer Debezium documentation. You are right that my 
earlier blanket statement about upstream standby support is outdated: current 
Debezium documents physical-standby capture as an incubating feature, with the 
primary used for snapshots/metadata and the standby used for LogMiner 
streaming. It requires a flat-file dictionary and `dictionary_from_file` (among 
other settings). Debezium 1.9's documentation, by contrast, says standby 
capture is unsupported. See 
https://debezium.io/documentation/reference/stable/connectors/oracle.html and 
https://debezium.io/documentation/reference/1.9/connectors/oracle.html.
   
   SeaTunnel 2.3.12 and current `dev` still pin `debezium-connector-oracle` to 
`1.9.8.Final`. SeaTunnel also carries Oracle LogMiner code: 
`LogMinerAdapter.startSession()` explicitly uses `DICT_FROM_ONLINE_CATALOG`, 
and `LogMinerStreamingChangeEventSource` calls `DBMS_LOGMNR_D.BUILD`. The 
existing `log.mining.read.only` setting only addresses the flush behavior; 
changing the transaction snapshot boundary or adding `dictionary_from_file` to 
the 2.3.12 config does not replace those paths. I found no open SeaTunnel PR 
implementing the newer physical-standby mode, so there is no verified 
upgrade/support date to give.
   
   For a scoped contribution, please confirm the Oracle version and whether 
this is a physical or logical ADG standby, whether a primary connection is 
available for snapshots/metadata, and whether the standby can access a 
dictionary file exported from the primary (redact hosts and credentials). The 
design/PR then needs to cover the Debezium upgrade plus SeaTunnel's overrides, 
primary/standby connection routing, schema-history and checkpoint behavior, and 
a real primary-to-ADG restore test. A dependency-only bump would not establish 
support. I marked this as `help wanted` for that focused design and validation 
work.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to