Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[SPARKNLP-940] Adding changes to correctly copy cluster index storage… #14167

Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion build.sbt
Original file line number Diff line number Diff line change
Expand Up @@ -144,13 +144,17 @@ lazy val utilDependencies = Seq(
exclude ("com.fasterxml.jackson.core", "jackson-annotations")
exclude ("com.fasterxml.jackson.core", "jackson-databind")
exclude ("com.fasterxml.jackson.core", "jackson-core")
exclude ("com.fasterxml.jackson.dataformat", "jackson-dataformat-cbor")
exclude ("commons-configuration", "commons-configuration"),
liblevenshtein
exclude ("com.google.guava", "guava")
exclude ("org.apache.commons", "commons-lang3")
exclude ("com.google.code.findbugs", "annotations")
exclude ("org.slf4j", "slf4j-api"),
gcpStorage,
gcpStorage
exclude ("com.fasterxml.jackson.core", "jackson-core")
exclude ("com.fasterxml.jackson.dataformat", "jackson-dataformat-cbor")
,
greex,
azureIdentity,
azureStorage)
Expand Down
4 changes: 3 additions & 1 deletion examples/util/Load_Model_from_GCP_Storage.ipynb
Original file line number Diff line number Diff line change
Expand Up @@ -80,7 +80,8 @@
"1. GCP connector: You need to identify your hadoop version and set the required dependency in `spark.jars.packages`\n",
"2. ADC credentials: After following the instructions to setup ADC, you will have a JSON file that holds your authenticiation information. This file is setup in `spark.hadoop.google.cloud.auth.service.account.json.keyfile`\n",
"3. Hadoop File System: You also need to setup the Hadoop implementation to work with GCP Storage as file system. This is define in `spark.hadoop.fs.gs.impl`\n",
"3. Finally, to mitigate conflicts between Spark's dependencies and user dependencies. You must define `spark.driver.userClassPathFirst` as true. You may also need to define `spark.executor.userClassPathFirst` as true.\n",
"4. To mitigate conflicts between Spark's dependencies and user dependencies. You must define `spark.driver.userClassPathFirst` as true. You may also need to define `spark.executor.userClassPathFirst` as true.\n",
"5. Additonaly, to avoid conflict errors whe need to exclude the following dependency: `com.fasterxml.jackson.core:jackson-core`\n",
"\n"
]
},
Expand Down Expand Up @@ -128,6 +129,7 @@
" \"spark.jars.packages\": \"com.google.cloud.bigdataoss:gcs-connector:hadoop3-2.2.8\",\n",
" \"spark.hadoop.fs.gs.impl\": \"com.google.cloud.hadoop.fs.gcs.GoogleHadoopFileSystem\",\n",
" \"spark.driver.userClassPathFirst\": \"true\",\n",
" \"spark.jars.excludes\": \"com.fasterxml.jackson.core:jackson-core\",\n",
" \"spark.hadoop.google.cloud.auth.service.account.json.keyfile\": json_keyfile,\n",
" \"spark.jsl.settings.gcp.project_id\": PROJECT_ID,\n",
" \"spark.jsl.settings.pretrained.cache_folder\": CACHE_FOLDER\n",
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -42,9 +42,9 @@ final class RocksDBConnection private (path: String) extends AutoCloseable {
}

def findLocalIndex: String = {
val localPath = RocksDBConnection.getLocalPath(path)
if (new File(localPath).exists()) {
localPath
val tmpIndexStorageLocalPath = RocksDBConnection.getTmpIndexStorageLocalPath(path)
if (new File(tmpIndexStorageLocalPath).exists()) {
tmpIndexStorageLocalPath
} else if (new File(path).exists()) {
path
} else {
Expand Down Expand Up @@ -135,7 +135,7 @@ object RocksDBConnection {
def getOrCreate(database: Database.Name, refName: String): RocksDBConnection =
getOrCreate(database.toString, refName)

def getLocalPath(fileName: String): String = {
def getTmpIndexStorageLocalPath(fileName: String): String = {
Path
.mergePaths(new Path(SparkFiles.getRootDirectory()), new Path("/storage/" + fileName))
.toString
Expand Down
49 changes: 35 additions & 14 deletions src/main/scala/com/johnsnowlabs/storage/StorageHelper.scala
Original file line number Diff line number Diff line change
Expand Up @@ -84,13 +84,38 @@ object StorageHelper {
sparkContext: SparkContext): Unit = {
destinationScheme match {
case "file" => {
val destination = new Path(RocksDBConnection.getLocalPath(clusterFileName))
copyIndexToLocal(source, destination, sparkContext)
val sourceFileSystemScheme = source.getFileSystem(sparkContext.hadoopConfiguration)
val tmpIndexStorageLocalPath =
RocksDBConnection.getTmpIndexStorageLocalPath(clusterFileName)
sourceFileSystemScheme.getScheme match {
case "file" => {
if (!doesDirectoryExistJava(tmpIndexStorageLocalPath) ||
!doesDirectoryExistHadoop(tmpIndexStorageLocalPath, sparkContext)) {
copyIndexToLocal(source, new Path(tmpIndexStorageLocalPath), sparkContext)
}
}
case "s3a" =>
copyIndexToLocal(source, new Path(tmpIndexStorageLocalPath), sparkContext)
case _ => copyIndexToCluster(source, clusterFilePath, sparkContext)
}
}
case _ => {
copyIndexToCluster(source, clusterFilePath, sparkContext)
}
case _ => copyIndexToCluster(source, clusterFilePath, sparkContext)
}
}

private def doesDirectoryExistJava(path: String): Boolean = {
val directory = new File(path)
directory.exists && directory.isDirectory
}

private def doesDirectoryExistHadoop(path: String, sparkContext: SparkContext): Boolean = {
val localPath = new Path(path)
val fileSystem = localPath.getFileSystem(sparkContext.hadoopConfiguration)
fileSystem.exists(localPath)
}

private def copyIndexToCluster(
sourcePath: Path,
dst: Path,
Expand Down Expand Up @@ -129,21 +154,17 @@ object StorageHelper {
val fileSystemDestination = destination.getFileSystem(sparkContext.hadoopConfiguration)
val fileSystemSource = source.getFileSystem(sparkContext.hadoopConfiguration)

if (fileSystemDestination.exists(destination)) {
return
}

if (fileSystemSource.getScheme == "s3a" && fileSystemDestination.getScheme == "file") {
if (fileSystemSource.getScheme == "file") {
fileSystemDestination.copyFromLocalFile(false, true, source, destination)
} else {
CloudResources.downloadBucketToLocalTmp(
source.toString,
destination.toString,
isIndex = true)
sparkContext.addFile(destination.toString, recursive = true)
return
}

if (fileSystemDestination.getScheme != "s3a") {
fileSystemDestination.copyFromLocalFile(false, true, source, destination)
val isLocalMode = sparkContext.master.startsWith("local")
if (isLocalMode) {
sparkContext.addFile(destination.toString, recursive = true)
}
}
}

Expand Down
Loading