[CARBONDATA-2702] Fix bugs in clear bloom datamap

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.5.0, 1.4.1
• Assignee: Chuanyin Xu
• Reporter: Chuanyin Xu
• Created: 06/Jul/18 06:22
• Updated: 11/Jul/18 07:13
• Resolved: 11/Jul/18 07:13

Description

In test output, sometimes there are error logs like:
```
org.apache.spark.SparkException: Job aborted due to stage failure: Task 3 in stage 805.0 failed 1 times, most recent failure: Lost task 3.0 in stage 805.0 (TID 5751, localhost, executor driver): java.lang.NegativeArraySizeException
at java.util.AbstractCollection.toArray(AbstractCollection.java:136)
at java.util.ArrayList.<init>(ArrayList.java:177)
at org.apache.carbondata.datamap.bloom.BloomCoarseGrainDataMapFactory.clear(BloomCoarseGrainDataMapFactory.java:340)
at org.apache.carbondata.core.datamap.TableDataMap.clear(TableDataMap.java:206)
at org.apache.carbondata.core.datamap.DataMapStoreManager.clearDataMaps(DataMapStoreManager.java:430)
at org.apache.carbondata.core.datamap.DistributableDataMapFormat$1.initialize(DistributableDataMapFormat.java:125)
at org.apache.carbondata.spark.rdd.DataMapPruneRDD.internalCompute(SparkDataMapJob.scala:73)
```

This error is caused by concurrent clear for datamaps, can refer to BlockletDataMapFactory at issue 2496 (PR2324)

Issue Links

• links to: GitHub Pull Request #2457