[SPARK-50483][CORE][SQL][3.5] BlockMissingException should be thrown even if ignoreCorruptFiles is enabled #49105
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…if ignoreCorruptFiles is enabled ### What changes were proposed in this pull request? `BlockMissingException` extends from `IOException`. When `BlockMissingException` occurs and ignoreCorruptFiles is enabled, the current task may not get any data and will be marked as successful([code](https://github.com/apache/spark/blob/0d045db8d15d0aeb0f54a1557fd360363e77ed42/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/FileScanRDD.scala#L271-L273)). This will cause data quality issues. Generally speaking, `BlockMissingException` is a system issue, not a file corruption issue. Therefore, `BlockMissingException` should be thrown even if ignoreCorruptFiles is enabled. Related error message: ``` 24/11/29 01:56:00 WARN FileScanRDD: Skipped the rest of the content in the corrupted file: path: viewfs://hadoop-cluster/path/to/data/part-00320-7915e327-3214-4585-a44e-f9c58e362b43.c000.snappy.parquet, range: 191727616-281354675, partition values: [empty row] org.apache.hadoop.hdfs.BlockMissingException: Could not obtain block: BP-169998034-10.210.23.11-1507067630530:blk_83565156183_82548880660 file/path/to/data/part-00320-7915e327-3214-4585-a44e-f9c58e362b43.c000.snappy.parquet No live nodes contain current block Block locations: DatanodeInfoWithStorage[10.209.145.174:50010,DS-c7c0a172-5ffa-4f90-bfb5-717fb1e9ecf2,DISK] DatanodeInfoWithStorage[10.3.22.142:50010,DS-a1ba9ac9-dc92-4131-a2c2-9f7d03b97caf,DISK] DatanodeInfoWithStorage[10.209.146.156:50010,DS-71d8ae97-15d3-454e-a715-d9490e184989,DISK] Dead nodes: DatanodeInfoWithStorage[10.209.146.156:50010,DS-71d8ae97-15d3-454e-a715-d9490e184989,DISK] DatanodeInfoWithStorage[10.209.145.174:50010,DS-c7c0a172-5ffa-4f90-bfb5-717fb1e9ecf2,DISK] DatanodeInfoWithStorage[10.3.22.142:50010,DS-a1ba9ac9-dc92-4131-a2c2-9f7d03b97caf,DISK] ```  ### Why are the changes needed? Avoid data issue if ignoreCorruptFiles is enabled when `BlockMissingException` occurred. ### Does this PR introduce _any_ user-facing change? Yes. ### How was this patch tested? Manual test. ### Was this patch authored or co-authored using generative AI tooling? No. Closes apache#49089 from wangyum/SPARK-50483. Authored-by: Yuming Wang <yumwang@ebay.com> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org> (cherry picked from commit 934a387)
…if ignoreCorruptFiles is enabled ### What changes were proposed in this pull request? `BlockMissingException` extends from `IOException`. When `BlockMissingException` occurs and ignoreCorruptFiles is enabled, the current task may not get any data and will be marked as successful([code](https://github.com/apache/spark/blob/0d045db8d15d0aeb0f54a1557fd360363e77ed42/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/FileScanRDD.scala#L271-L273)). This will cause data quality issues. Generally speaking, `BlockMissingException` is a system issue, not a file corruption issue. Therefore, `BlockMissingException` should be thrown even if ignoreCorruptFiles is enabled. Related error message: ``` 24/11/29 01:56:00 WARN FileScanRDD: Skipped the rest of the content in the corrupted file: path: viewfs://hadoop-cluster/path/to/data/part-00320-7915e327-3214-4585-a44e-f9c58e362b43.c000.snappy.parquet, range: 191727616-281354675, partition values: [empty row] org.apache.hadoop.hdfs.BlockMissingException: Could not obtain block: BP-169998034-10.210.23.11-1507067630530:blk_83565156183_82548880660 file/path/to/data/part-00320-7915e327-3214-4585-a44e-f9c58e362b43.c000.snappy.parquet No live nodes contain current block Block locations: DatanodeInfoWithStorage[10.209.145.174:50010,DS-c7c0a172-5ffa-4f90-bfb5-717fb1e9ecf2,DISK] DatanodeInfoWithStorage[10.3.22.142:50010,DS-a1ba9ac9-dc92-4131-a2c2-9f7d03b97caf,DISK] DatanodeInfoWithStorage[10.209.146.156:50010,DS-71d8ae97-15d3-454e-a715-d9490e184989,DISK] Dead nodes: DatanodeInfoWithStorage[10.209.146.156:50010,DS-71d8ae97-15d3-454e-a715-d9490e184989,DISK] DatanodeInfoWithStorage[10.209.145.174:50010,DS-c7c0a172-5ffa-4f90-bfb5-717fb1e9ecf2,DISK] DatanodeInfoWithStorage[10.3.22.142:50010,DS-a1ba9ac9-dc92-4131-a2c2-9f7d03b97caf,DISK] ```  ### Why are the changes needed? Avoid data issue if ignoreCorruptFiles is enabled when `BlockMissingException` occurred. ### Does this PR introduce _any_ user-facing change? Yes. ### How was this patch tested? Manual test. ### Was this patch authored or co-authored using generative AI tooling? No. Closes apache#49089 from wangyum/SPARK-50483. Authored-by: Yuming Wang <yumwang@ebay.com> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org> (cherry picked from commit 934a387)
dongjoon-hyun
approved these changes
Dec 7, 2024
|
cc @LuciferYang |
|
Thank you for letting me know @dongjoon-hyun |
dongjoon-hyun
pushed a commit
that referenced
this pull request
Dec 8, 2024
…even if ignoreCorruptFiles is enabled ### What changes were proposed in this pull request? `BlockMissingException` extends from `IOException`. When `BlockMissingException` occurs and ignoreCorruptFiles is enabled, the current task may not get any data and will be marked as successful([code](https://github.com/apache/spark/blob/0d045db8d15d0aeb0f54a1557fd360363e77ed42/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/FileScanRDD.scala#L271-L273)). This will cause data quality issues. Generally speaking, `BlockMissingException` is a system issue, not a file corruption issue. Therefore, `BlockMissingException` should be thrown even if ignoreCorruptFiles is enabled. Related error message: ``` 24/11/29 01:56:00 WARN FileScanRDD: Skipped the rest of the content in the corrupted file: path: viewfs://hadoop-cluster/path/to/data/part-00320-7915e327-3214-4585-a44e-f9c58e362b43.c000.snappy.parquet, range: 191727616-281354675, partition values: [empty row] org.apache.hadoop.hdfs.BlockMissingException: Could not obtain block: BP-169998034-10.210.23.11-1507067630530:blk_83565156183_82548880660 file/path/to/data/part-00320-7915e327-3214-4585-a44e-f9c58e362b43.c000.snappy.parquet No live nodes contain current block Block locations: DatanodeInfoWithStorage[10.209.145.174:50010,DS-c7c0a172-5ffa-4f90-bfb5-717fb1e9ecf2,DISK] DatanodeInfoWithStorage[10.3.22.142:50010,DS-a1ba9ac9-dc92-4131-a2c2-9f7d03b97caf,DISK] DatanodeInfoWithStorage[10.209.146.156:50010,DS-71d8ae97-15d3-454e-a715-d9490e184989,DISK] Dead nodes: DatanodeInfoWithStorage[10.209.146.156:50010,DS-71d8ae97-15d3-454e-a715-d9490e184989,DISK] DatanodeInfoWithStorage[10.209.145.174:50010,DS-c7c0a172-5ffa-4f90-bfb5-717fb1e9ecf2,DISK] DatanodeInfoWithStorage[10.3.22.142:50010,DS-a1ba9ac9-dc92-4131-a2c2-9f7d03b97caf,DISK] ```  ### Why are the changes needed? Avoid data issue if ignoreCorruptFiles is enabled when `BlockMissingException` occurred. ### Does this PR introduce _any_ user-facing change? Yes. ### How was this patch tested? Manual test. ### Was this patch authored or co-authored using generative AI tooling? No. Closes #49105 from wangyum/SPARK-50483-branch-3.5. Authored-by: Yuming Wang <yumwang@ebay.com> Signed-off-by: Dongjoon Hyun <dongjoon@apache.org>
|
Merged to branch-3.5 for Apache Spark 3.5.4. |
|
To @LuciferYang , please let me know if you have any blocker or need my help during your Spark 3.5.4 RC1 preparation step. AFAIK, this is the last one we have been waiting for. |
Yes, I think this should be the last one. Thank you very much for helping resolve many issues over the past few days. @dongjoon-hyun |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
BlockMissingExceptionextends fromIOException. WhenBlockMissingExceptionoccurs and ignoreCorruptFiles is enabled, the current task may not get any data and will be marked as successful(code). This will cause data quality issues.Generally speaking,
BlockMissingExceptionis a system issue, not a file corruption issue. Therefore,BlockMissingExceptionshould be thrown even if ignoreCorruptFiles is enabled.Related error message:
Why are the changes needed?
Avoid data issue if ignoreCorruptFiles is enabled when
BlockMissingExceptionoccurred.Does this PR introduce any user-facing change?
Yes.
How was this patch tested?
Manual test.
Was this patch authored or co-authored using generative AI tooling?
No.