Copilot commented on code in PR #2859:
URL: https://github.com/apache/fluss/pull/2859#discussion_r2929413067
##########
fluss-flink/fluss-flink-common/src/main/java/org/apache/fluss/flink/tiering/source/TieringSource.java:
##########
@@ -110,6 +112,9 @@ public SourceReader<TableBucketWriteResult<WriteResult>,
TieringSplit> createRea
SourceReaderContext sourceReaderContext) {
FutureCompletingBlockingQueue<RecordsWithSplitIds<TableBucketWriteResult<WriteResult>>>
elementsQueue = new FutureCompletingBlockingQueue<>();
+ flussConf.set(
+ CLIENT_SCANNER_IO_TMP_DIR,
+ getClientScannerIoTmpDir(flussConf,
sourceReaderContext.getConfiguration()));
Review Comment:
The new fallback relies on `getClientScannerIoTmpDir(...)`, but that helper
currently builds the Flink-based path via `new File(flinkTmpDirs, "/fluss")`.
In Java, a child path starting with `/` is treated as absolute, so the parent
(`io.tmp.dirs`) is ignored and this will resolve to `/fluss` (and may fail due
to permissions), defeating the purpose of falling back to Flink’s tmp dir. Fix
`getClientScannerIoTmpDir` to join with a relative child (e.g., `"fluss"`) and
consider handling multi-value `io.tmp.dirs` (comma-separated) deterministically.
--
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]
