Log files in Hudi MOR table are not getting deleted

**_Tips before filing an issue_**

- Have you gone through our [FAQs](https://hudi.apache.org/learn/faq/)?

- Join the mailing list to engage in conversations and get faster support at dev-subscribe@hudi.apache.org.

- If you have triaged this as a bug, then file an [issue](https://issues.apache.org/jira/projects/HUDI/issues) directly.

We use MOR Hudi table , We read kinesis stream and process in AWS EMR using spark streaming.
We use inline compaction and cleanup based on the commits,
Though compaction and cleanup is running we see some of hudi log files that are generated while ingestion are not getting cleared fully.
We process with batch interval of 5 mints for every 5 mints we do one hudi commit.

Below are our hudi configs

DataSourceWriteOptions.TABLE_TYPE.key() -> DataSourceWriteOptions.MOR_TABLE_TYPE_OPT_VAL,

DataSourceWriteOptions.RECORDKEY_FIELD.key() -> "guid",

DataSourceWriteOptions.PARTITIONPATH_FIELD.key() -> "collectionName",

DataSourceWriteOptions.PRECOMBINE_FIELD.key() -> "operationTime",

DataSourceWriteOptions.HIVE_PARTITION_FIELDS.key() -> "collectionName",

DataSourceWriteOptions.HIVE_PARTITION_EXTRACTOR_CLASS.key() -> classOf[MultiPartKeysValueExtractor].getName,

DataSourceWriteOptions.HIVE_SYNC_MODE.key() -> "hms",

DataSourceWriteOptions.HIVE_USE_JDBC.key() -> "false",*/

HoodieCompactionConfig.INLINE_COMPACT_TRIGGER_STRATEGY.key() -> CompactionTriggerStrategy.TIME_ELAPSED.name,

HoodieCompactionConfig.INLINE_COMPACT_TIME_DELTA_SECONDS.key() -> String.valueOf(60 * 6 * 60),

HoodieCompactionConfig.CLEANER_POLICY.key() -> HoodieCleaningPolicy.KEEP_LATEST_COMMITS.name(),

HoodieCompactionConfig.CLEANER_COMMITS_RETAINED.key() -> "228",

HoodieCompactionConfig.MIN_COMMITS_TO_KEEP.key() -> "229",

HoodieCompactionConfig.MAX_COMMITS_TO_KEEP.key() -> "252",

HoodieCompactionConfig.ASYNC_CLEAN.key() -> "false",

HoodieCompactionConfig.INLINE_COMPACT.key() -> "true",

HoodieMetricsConfig.TURN_METRICS_ON.key() -> "true",

HoodieMetricsConfig.METRICS_REPORTER_TYPE_VALUE.key() -> MetricsReporterType.DATADOG.name(),

HoodieMetricsDatadogConfig.API_SITE_VALUE.key() -> "US",

HoodieMetricsDatadogConfig.METRIC_PREFIX_VALUE.key() -> "XXXX.hudi",

HoodieMetricsDatadogConfig.API_KEY_SUPPLIER.key() -> "XXXXX.XXX.DatadogKeySupplier",

HoodieMetadataConfig.ENABLE.key() -> "false",

HoodieWriteConfig.ROLLBACK_USING_MARKERS_ENABLE.key() -> "false",

Snapshots attached :



If you want the .hoodi folder, Please let us know we will share it here.

* Hudi version : 0.9

* Spark version : 3.2.1

* EMR Version : 6.7

* Storage (HDFS/S3/GCS..) : S3

* Running on Docker? (yes/no) : no

Start by reviewing the MOR compaction and cleaner settings in the issue, especially the retained-commit values, then inspect the attached snapshots and reproduce the behavior with Hudi 0.9, Spark 3.2.1, EMR 6.7, and S3. Done means determining why generated log files remain after compaction and cleanup, with the relevant configuration or reproduction narrowed down.

Written by the indexing model from the issue text.