github-actions[bot] commented on code in PR #60887:
URL: https://github.com/apache/doris/pull/60887#discussion_r3321849622
##########
fe/fe-core/src/main/java/org/apache/doris/tablefunction/MetadataGenerator.java:
##########
@@ -1913,11 +1914,19 @@ private static void
partitionsForInternalCatalog(UserIdentity currentUserIdentit
trow.addToColumnValue(new TCell().setStringVal("")); //
NODEGROUP (not available)
trow.addToColumnValue(new TCell().setStringVal("")); //
TABLESPACE_NAME (not available)
- Pair<Double, String> sizePair =
DebugUtil.getByteUint(partition.getDataSize(false));
+ long localDataSize = partition.getDataSize(false);
+ long remoteDataSize = partition.getRemoteDataSize();
+ // In cloud mode, FE replica dataSize may hold remote
bytes while remoteDataSize is 0.
+ // Keep the output semantic consistent with
backend_tablets.
Review Comment:
This leaves `information_schema.partitions` inconsistent for the same
fallback case handled in `TabletsProcDir`: when cloud tablet stats have
`data_size == 0` but `index_size`/`segment_size` have already been populated.
In cloud, `partition.getRemoteDataSize()` still sums
`Tablet#getRemoteDataSize()`, and `CloudTablet` does not override that method,
so it remains 0; if `partition.getDataSize(false)` is also 0, this branch does
nothing and the view returns `LOCAL_DATA_SIZE=0` and `REMOTE_DATA_SIZE=0` even
though `SHOW TABLETS` now derives the remote size from
`replica.getLocalInvertedIndexSize() + replica.getLocalSegmentSize()`. Please
apply the same fallback for partitions, e.g. by summing the visible indexes'
replica index/segment sizes (or centralizing the cloud size mapping), and add
coverage for this `data_size == 0` path.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
