nzw921rx commented on PR #12829:
URL: https://github.com/apache/seatunnel/pull/12829#issuecomment-6246584957
Reviewed the current head. One item should be addressed before merge:
**`query.timeout` bounds the whole streamed read, not just the start of the
query.** In Couchbase Java SDK 3.5.1, `QueryChunkResponseParser` notes that the
client-side request timeout is cancelled once streaming begins, but the server
still enforces the query timeout while streaming and returns a timeout (code
1080) if the full read exceeds it. With the default `query.timeout = 300`, any
collection export longer than 5 minutes will fail; since the source checkpoints
no offset, the retry restarts from row 1 and can fail again at the same point.
Please document this explicitly in the option description and in both EN/ZH
docs, and recommend a larger value for large collections (or reconsider the
default). A small-timeout + slow-collector test against the real fixture would
confirm the behavior.
The remaining items are Low and can be follow-ups: whole-document `SELECT
RAW d` without projection, the fixture-only `CouchbaseSourceConfigTest`, and
the shared option text ("to write to") / missing factory Javadoc. The overall
design looks good.
--
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]
