[CARBONDATA-3885] [CARBONDATA-3884] Fix for load failures with isSIenabled = false and fix for concurrent load failure

Why is this PR needed?
Issue 1:
In some concurrent scenarios it was found that certain segment folders were getting deleted for the SI table. When
the segment was being inserted again back to the SI table during a new load to the main table by doing isSIenabled = false,
the load to the SI table failed.

Root Cause: It was found that when the segment folder gets deleted and it's entry from table status file is manually
deleted then, during new load to SI when table status is changed to InsertInProgress during the loading previously
deleted segments the stale .segmentfile still exists which was pointing to the mergeindexfile which does exist
anymore(since the segment folder is deleted). Since that mergeindexfile does not exist anymore the load to SI table failed

Issue 2:
In some concurrent load scenarios in main table with SI table with isSITableEnabled = false , load to SI table fails.

Root Cause: When concurrent loading is done with isSITableEnabled = false, maintable details are taken earlier in the
code and SItable details are taken later. It was found that there were some scenarios where another concurrent load to
main table finishes and can change the number of segments in the main table due to which there is mismatch in number
of segments in the SI table and in the main table. Due to this mismatch when the validmaintableload method is called,
there can be scenarios where the load fails.

What changes were proposed in this PR?
Solution for Issue 1: Whenever isSIenabled is set to false and it tries to load previously deleted SI segments to the
SI table we go and delete the stale .segment file from the metadata folder if it exists. Due to this, while loading a
new .segment file will be created thus having a new index file path.

Solution for Issue 2: Whenever isSIenabled is set to false and it tries to load previously deleted SI segments,
just before adding segments in SIFailedSegmentList checking again if the loadname is present in the main table. Only
adding in SIFailedSegmenList if the loadname is present in the main table and it's in SUCCESS state.

This closes #3802
2 files changed
tree: c850007044293a611b2a58f9343525729e88743f
  1. .github/
  2. assembly/
  3. bin/
  4. build/
  5. common/
  6. conf/
  7. core/
  8. dev/
  9. docs/
  10. examples/
  11. format/
  12. geo/
  13. hadoop/
  14. index/
  15. integration/
  16. licenses-binary/
  17. mv/
  18. processing/
  19. python/
  20. sdk/
  21. streaming/
  22. tools/
  23. .gitignore
  24. LICENSE
  25. NOTICE
  26. pom.xml
  27. README.md
README.md

Apache CarbonData is an indexed columnar data store solution for fast analytics on big data platform, e.g.Apache Hadoop, Apache Spark, etc.

You can find the latest CarbonData document and learn more at: http://carbondata.apache.org

CarbonData cwiki

Visit count: HitCount

Status

Spark2.4: Build Status Coverage Status

Features

CarbonData file format is a columnar store in HDFS, it has many features that a modern columnar format has, such as splittable, compression schema ,complex data type etc, and CarbonData has following unique features:

  • Stores data along with index: it can significantly accelerate query performance and reduces the I/O scans and CPU resources, where there are filters in the query. CarbonData index consists of multiple level of indices, a processing framework can leverage this index to reduce the task it needs to schedule and process, and it can also do skip scan in more finer grain unit (called blocklet) in task side scanning instead of scanning the whole file.
  • Operable encoded data :Through supporting efficient compression and global encoding schemes, can query on compressed/encoded data, the data can be converted just before returning the results to the users, which is “late materialized”.
  • Supports for various use cases with one single Data format : like interactive OLAP-style query, Sequential Access (big scan), Random Access (narrow scan).

Building CarbonData

CarbonData is built using Apache Maven, to build CarbonData

Online Documentation

Experimental Features

Some features are marked as experimental because the syntax/implementation might change in the future.

  1. Hybrid format table using Add Segment.
  2. Accelerating performance using MV on parquet/orc.
  3. Merge API for Spark DataFrame.
  4. Hive write for non-transactional table.

Integration

Other Technical Material

Fork and Contribute

This is an active open source project for everyone, and we are always open to people who want to use this system or contribute to it. This guide document introduce how to contribute to CarbonData.

Contact us

To get involved in CarbonData:

About

Apache CarbonData is an open source project of The Apache Software Foundation (ASF).