BsoBird commented on code in PR #9546:
URL: https://github.com/apache/iceberg/pull/9546#discussion_r1506038699
##########
core/src/main/java/org/apache/iceberg/hadoop/HadoopTableOperations.java:
##########
@@ -149,26 +183,71 @@ public void commit(TableMetadata base, TableMetadata
metadata) {
String codecName =
metadata.property(
TableProperties.METADATA_COMPRESSION,
TableProperties.METADATA_COMPRESSION_DEFAULT);
+ // TODO:This is not compatible with the scenario where the user modifies
the metadata file
+ // compression codec arbitrarily.
+ // We can inform the user about this bug first, and fix it later.(Do not
modify the compressed
+ // format after the table is created.)
TableMetadataParser.Codec codec =
TableMetadataParser.Codec.fromName(codecName);
String fileExtension = TableMetadataParser.getFileExtension(codec);
- Path tempMetadataFile = metadataPath(UUID.randomUUID().toString() +
fileExtension);
+ Path tempMetadataFile = metadataPath(UUID.randomUUID() + fileExtension);
TableMetadataParser.write(metadata,
io().newOutputFile(tempMetadataFile.toString()));
int nextVersion = (current.first() != null ? current.first() : 0) + 1;
Path finalMetadataFile = metadataFilePath(nextVersion, codec);
FileSystem fs = getFileSystem(tempMetadataFile, conf);
-
- // this rename operation is the atomic commit operation
- renameToFinal(fs, tempMetadataFile, finalMetadataFile, nextVersion);
-
- LOG.info("Committed a new metadata file {}", finalMetadataFile);
-
- // update the best-effort version pointer
- writeVersionHint(nextVersion);
-
- deleteRemovedMetadataFiles(base, metadata);
-
- this.shouldRefresh = true;
+ boolean versionCommitSuccess = false;
+ try {
+ deleteOldVersionHint(fs, versionHintFile(), nextVersion);
Review Comment:
Yes, I've basically given up on concurrency checks. It's hard to guarantee
atomicity.
However, if the contents of the versionHint are already wrong, it's
basically pointless to continue down the commit path, and there's a high
probability that we'll get an exception in the commitNewVersion method.
Since we're bound to get the wrong result with the wrong data, why don't we
just stop everything at the beginning?If the check fails(FileNotFound), it's a
good way to stop execution.
--
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: issues-unsubscr...@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org
For additional commands, e-mail: issues-h...@iceberg.apache.org
