[SPARK-25240][SQL] Fix for a deadlock in RECOVER PARTITIONS #22233
There are no files selected for viewing
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
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
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -52,23 +52,24 @@ class InMemoryCatalogedDDLSuite extends DDLSuite with SharedSQLContext with Befo | |
| protected override def generateTable( | ||
| catalog: SessionCatalog, | ||
| name: TableIdentifier, | ||
| isDataSource: Boolean = true, | ||
| partitionCols: Seq[String] = Seq("a", "b")): CatalogTable = { | ||
| val storage = | ||
| CatalogStorageFormat.empty.copy(locationUri = Some(catalog.defaultTablePath(name))) | ||
| val metadata = new MetadataBuilder() | ||
| .putString("key", "value") | ||
| .build() | ||
| val schema = new StructType() | ||
| .add("col1", "int", nullable = true, metadata = metadata) | ||
| .add("col2", "string") | ||
| CatalogTable( | ||
| identifier = name, | ||
| tableType = CatalogTableType.EXTERNAL, | ||
| storage = storage, | ||
| schema = schema.copy( | ||
| fields = schema.fields ++ partitionCols.map(StructField(_, IntegerType))), | ||
| provider = Some("parquet"), | ||
| partitionColumnNames = partitionCols, | ||
| createTime = 0L, | ||
| createVersion = org.apache.spark.SPARK_VERSION, | ||
| tracksPartitionsInCatalog = true) | ||
|
|
@@ -176,7 +177,8 @@ abstract class DDLSuite extends QueryTest with SQLTestUtils { | |
| protected def generateTable( | ||
| catalog: SessionCatalog, | ||
| name: TableIdentifier, | ||
| isDataSource: Boolean = true, | ||
| partitionCols: Seq[String] = Seq("a", "b")): CatalogTable | ||
|
|
||
| private val escapedIdentifier = "`(.+)`".r | ||
|
|
||
|
|
@@ -228,8 +230,10 @@ abstract class DDLSuite extends QueryTest with SQLTestUtils { | |
| private def createTable( | ||
| catalog: SessionCatalog, | ||
| name: TableIdentifier, | ||
| isDataSource: Boolean = true, | ||
| partitionCols: Seq[String] = Seq("a", "b")): Unit = { | ||
| catalog.createTable( | ||
| generateTable(catalog, name, isDataSource, partitionCols), ignoreIfExists = false) | ||
| } | ||
|
|
||
| private def createTablePartition( | ||
|
|
@@ -1131,7 +1135,7 @@ abstract class DDLSuite extends QueryTest with SQLTestUtils { | |
| } | ||
|
|
||
| test("alter table: recover partition (parallel)") { | ||
| withSQLConf("spark.rdd.parallelListingThreshold" -> "0") { | ||
|
Member
There was a problem hiding this comment. @MaxGekk, out of curiosity, why does this have to be 0?
Member
Author
There was a problem hiding this comment. On the recursive calls this condition statuses.length is 1 and threshold is 1. So, it leads to sequential listening of files. I just enforce parallel scanning even for 1 file/folder.
|
||
| testRecoverPartitions() | ||
| } | ||
| } | ||
|
|
@@ -1144,23 +1148,32 @@ abstract class DDLSuite extends QueryTest with SQLTestUtils { | |
| } | ||
|
|
||
| val tableIdent = TableIdentifier("tab1") | ||
| createTable(catalog, tableIdent, partitionCols = Seq("a", "b", "c")) | ||
| val part1 = Map("a" -> "1", "b" -> "5", "c" -> "19") | ||
| createTablePartition(catalog, part1, tableIdent) | ||
| assert(catalog.listPartitions(tableIdent).map(_.spec).toSet == Set(part1)) | ||
|
|
||
| val part2 = Map("a" -> "2", "b" -> "6", "c" -> "31") | ||
| val root = new Path(catalog.getTableMetadata(tableIdent).location) | ||
| val fs = root.getFileSystem(spark.sessionState.newHadoopConf()) | ||
| // valid | ||
| fs.mkdirs(new Path(new Path(new Path(root, "a=1"), "b=5"), "c=19")) | ||
| fs.createNewFile(new Path(new Path(root, "a=1/b=5/c=19"), "a.csv")) // file | ||
| fs.createNewFile(new Path(new Path(root, "a=1/b=5/c=19"), "_SUCCESS")) // file | ||
|
|
||
| fs.mkdirs(new Path(new Path(new Path(root, "A=2"), "B=6"), "C=31")) | ||
| fs.createNewFile(new Path(new Path(root, "A=2/B=6/C=31"), "b.csv")) // file | ||
| fs.createNewFile(new Path(new Path(root, "A=2/B=6/C=31"), "c.csv")) // file | ||
| fs.createNewFile(new Path(new Path(root, "A=2/B=6/C=31"), ".hiddenFile")) // file | ||
| fs.mkdirs(new Path(new Path(root, "A=2/B=6/C=31"), "_temporary")) | ||
|
|
||
| val parts = (10 to 100).map { a => | ||
| val part = Map("a" -> a.toString, "b" -> "5", "c" -> "42") | ||
| fs.mkdirs(new Path(new Path(new Path(root, s"a=$a"), "b=5"), "c=42")) | ||
| fs.createNewFile(new Path(new Path(root, s"a=$a/b=5/c=42"), "a.csv")) // file | ||
| createTablePartition(catalog, part, tableIdent) | ||
| part | ||
| } | ||
|
|
||
| // invalid | ||
| fs.mkdirs(new Path(new Path(root, "a"), "b")) // bad name | ||
|
|
@@ -1174,7 +1187,7 @@ abstract class DDLSuite extends QueryTest with SQLTestUtils { | |
| try { | ||
| sql("ALTER TABLE tab1 RECOVER PARTITIONS") | ||
| assert(catalog.listPartitions(tableIdent).map(_.spec).toSet == | ||
| Set(part1, part2) ++ parts) | ||
| if (!isUsingHiveMetastore) { | ||
| assert(catalog.getPartition(tableIdent, part1).parameters("numFiles") == "1") | ||
| assert(catalog.getPartition(tableIdent, part2).parameters("numFiles") == "2") | ||
|
|
||
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
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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
A question about the changes in this file. Are they related to the work of this PR?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@gatorsmile Yes, the changes are related to an existing test which was modified to reproduce the issue. In particular, this line is related to support of any number of partition columns.