public async Task TestSparkBatchJobCompletesWhenJobStarts()
{
SparkBatchClient client = CreateClient();
// Submit the Spark job
SparkBatchJobOptions createParams = SparkTestUtilities.CreateSparkJobRequestParameters(Recording, TestEnvironment);
SparkBatchOperation createOperation = await client.StartCreateSparkBatchJobAsync(createParams);
SparkBatchJob jobCreateResponse = await createOperation.WaitForCompletionAsync();
// Verify the Spark batch job submission starts successfully
Assert.True(LivyStates.Starting == jobCreateResponse.State || LivyStates.Running == jobCreateResponse.State || LivyStates.Success == jobCreateResponse.State,
string.Format(
"Job: {0} did not return success. Current job state: {1}. Error (if any): {2}",
jobCreateResponse.Id,
jobCreateResponse.State,
string.Join(", ", jobCreateResponse.Errors ?? new List <SparkServiceError>())
)
);
// Get the list of Spark batch jobs and check that the submitted job exists
List <SparkBatchJob> listJobResponse = await SparkTestUtilities.ListSparkBatchJobsAsync(client);
Assert.NotNull(listJobResponse);
Assert.IsTrue(listJobResponse.Any(job => job.Id == jobCreateResponse.Id));
}
public async Task TestSparkBatchJob()
{
SparkBatchClient client = CreateClient();
// Submit the Spark job
SparkBatchJobOptions createParams = SparkTestUtilities.CreateSparkJobRequestParameters(Recording, TestEnvironment);
SparkBatchOperation createOperation = await client.StartCreateSparkBatchJobAsync(createParams);
SparkBatchJob jobCreateResponse = await createOperation.WaitForCompletionAsync();
// Verify the Spark batch job completes successfully
Assert.True("success".Equals(jobCreateResponse.State, StringComparison.OrdinalIgnoreCase) && jobCreateResponse.Result == SparkBatchJobResultType.Succeeded,
string.Format(
"Job: {0} did not return success. Current job state: {1}. Actual result: {2}. Error (if any): {3}",
jobCreateResponse.Id,
jobCreateResponse.State,
jobCreateResponse.Result,
string.Join(", ", jobCreateResponse.Errors ?? new List <SparkServiceError>())
)
);
// Get the list of Spark batch jobs and check that the submitted job exists
List <SparkBatchJob> listJobResponse = await SparkTestUtilities.ListSparkBatchJobsAsync(client);
Assert.NotNull(listJobResponse);
Assert.IsTrue(listJobResponse.Any(job => job.Id == jobCreateResponse.Id));
}
public async Task TestSparkBatchJob()
{
// Submit the Spark job
SparkBatchJobOptions createParams = this.CreateSparkJobRequestParameters();
SparkBatchJob jobCreateResponse = (await SparkBatchClient.CreateSparkBatchJobAsync(createParams)).Value;
// Poll the Spark job until it finishes
SparkBatchJob getJobResponse = await this.PollSparkBatchJobSubmissionAsync(jobCreateResponse);
// Verify the Spark batch job completes successfully
Assert.True("success".Equals(getJobResponse.State, StringComparison.OrdinalIgnoreCase) && getJobResponse.Result == SparkBatchJobResultType.Succeeded,
string.Format(
"Job: {0} did not return success. Current job state: {1}. Actual result: {2}. Error (if any): {3}",
getJobResponse.Id,
getJobResponse.State,
getJobResponse.Result,
string.Join(", ", getJobResponse.Errors ?? new List <SparkServiceError>())
)
);
// Get the list of Spark batch jobs and check that the submitted job exists
List <SparkBatchJob> listJobResponse = await this.ListSparkBatchJobsAsync();
Assert.NotNull(listJobResponse);
Assert.IsTrue(listJobResponse.Any(job => job.Id == getJobResponse.Id));
}
public void SubmitSparkJobSync()
{
// Environment variable with the Synapse workspace endpoint.
string endpoint = TestEnvironment.EndpointUrl;
// Environment variable with the Synapse Spark pool name.
string sparkPoolName = TestEnvironment.SparkPoolName;
// Environment variable with the ADLS Gen2 storage account associated with the Synapse workspace.
string storageAccount = TestEnvironment.StorageAccountName;
// Environment variable with the file system of ADLS Gen2 storage account associated with the Synapse workspace.
string fileSystem = TestEnvironment.StorageFileSystemName;
#region Snippet:CreateSparkBatchClient
SparkBatchClient client = new SparkBatchClient(new Uri(endpoint), sparkPoolName, new DefaultAzureCredential());
#endregion
#region Snippet:SubmitSparkBatchJob
string name = $"batch-{Guid.NewGuid()}";
string file = string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/net/wordcount/wordcount.zip", fileSystem, storageAccount);
SparkBatchJobOptions request = new SparkBatchJobOptions(name, file)
{
ClassName = "WordCount",
Arguments =
{
string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/net/wordcount/shakespeare.txt", fileSystem, storageAccount),
string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/net/wordcount/result/", fileSystem, storageAccount),
},
DriverMemory = "28g",
DriverCores = 4,
ExecutorMemory = "28g",
ExecutorCores = 4,
ExecutorCount = 2
};
SparkBatchJob jobCreated = client.CreateSparkBatchJob(request);
#endregion
#region Snippet:ListSparkBatchJobs
Response <SparkBatchJobCollection> jobs = client.GetSparkBatchJobs();
foreach (SparkBatchJob job in jobs.Value.Sessions)
{
Console.WriteLine(job.Name);
}
#endregion
#region Snippet:GetSparkBatchJob
SparkBatchJob retrievedJob = client.GetSparkBatchJob(jobCreated.Id);
Debug.WriteLine($"Job is returned with name {retrievedJob.Name} and state {retrievedJob.State}");
#endregion
#region Snippet:DeleteSparkBatchJob
Response operation = client.CancelSparkBatchJob(jobCreated.Id);
#endregion
}
public async Task TestGetSparkBatchJob()
{
SparkBatchJobCollection sparkJobs = (await SparkBatchClient.GetSparkBatchJobsAsync()).Value;
foreach (SparkBatchJob expectedSparkJob in sparkJobs.Sessions)
{
SparkBatchJob actualSparkJob = await SparkBatchClient.GetSparkBatchJobAsync(expectedSparkJob.Id);
ValidateSparkBatchJob(expectedSparkJob, actualSparkJob);
}
}
public void SparkSample()
{
#region Snippet:CreateBatchClient
// Replace the string below with your actual endpoint url.
string endpoint = "<my-endpoint-url>";
/*@@*/ endpoint = TestEnvironment.EndpointUrl;
string sparkPoolName = TestEnvironment.SparkPoolName;
SparkBatchClient client = new SparkBatchClient(endpoint: new Uri(endpoint), sparkPoolName: sparkPoolName, credential: new DefaultAzureCredential());
#endregion
// Environment variable with the storage account associated with the Synapse workspace endpoint.
string storageAccount = TestEnvironment.StorageAccountName;
// Environment variable with the file system of the storage account.
string fileSystem = TestEnvironment.StorageFileSystemName;
#region Snippet:CreateBatchJob
string name = $"batchSample";
string file = string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/java/wordcount/wordcount.jar", fileSystem, storageAccount);
SparkBatchJobOptions options = new SparkBatchJobOptions(name: name, file: file)
{
ClassName = "WordCount",
Arguments =
{
string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/java/wordcount/shakespeare.txt", fileSystem, storageAccount),
string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/java/wordcount/result/", fileSystem, storageAccount),
},
DriverMemory = "28g",
DriverCores = 4,
ExecutorMemory = "28g",
ExecutorCores = 4,
ExecutorCount = 2
};
SparkBatchJob jobCreated = client.CreateSparkBatchJob(options);
#endregion
#region Snippet:ListSparkBatchJobs
Response <SparkBatchJobCollection> jobs = client.GetSparkBatchJobs();
foreach (SparkBatchJob job in jobs.Value.Sessions)
{
Console.WriteLine(job.Name);
}
#endregion
#region Snippet:DeleteSparkBatchJob
/*@@*/ int jobId = jobs.Value.Sessions.First().Id;
// Replace the integer below with your actual job ID.
//@@ string jobId = 0;
Response operation = client.CancelSparkBatchJob(jobId);
#endregion
}
public SynapseAnalyticsSparkClient(string workspaceName, string sparkPoolName, IAzureContext context)
{
if (context == null)
{
throw new AzPSInvalidOperationException(Resources.InvalidDefaultSubscription);
}
string suffix = context.Environment.GetEndpoint(AzureEnvironment.ExtendedEndpoint.AzureSynapseAnalyticsEndpointSuffix);
Uri uri = new Uri("https://" + workspaceName + "." + suffix);
_sparkBatchClient = new SparkBatchClient(uri, sparkPoolName, new AzureSessionCredential(context));
_sparkSessionClient = new SparkSessionClient(uri, sparkPoolName, new AzureSessionCredential(context));
}
public void SubmitSparkJobSync()
{
// Environment variable with the Synapse workspace endpoint.
string workspaceUrl = TestEnvironment.WorkspaceUrl;
// Environment variable with the Synapse Spark pool name.
string sparkPoolName = TestEnvironment.SparkPoolName;
// Environment variable with the ADLS Gen2 storage account associated with the Synapse workspace.
string storageAccount = TestEnvironment.StorageAccountName;
// Environment variable with the file system of ADLS Gen2 storage account associated with the Synapse workspace.
string fileSystem = TestEnvironment.StorageFileSystemName;
#region Snippet:SparkBatchSample1SparkBatchClient
SparkBatchClient client = new SparkBatchClient(new Uri(workspaceUrl), sparkPoolName, new DefaultAzureCredential());
#endregion
#region Snippet:SparkBatchSample1SubmitSparkJob
string name = $"batch-{Guid.NewGuid()}";
string file = string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/java/wordcount/wordcount.jar", fileSystem, storageAccount);
SparkBatchJobOptions request = new SparkBatchJobOptions(name, file)
{
ClassName = "WordCount",
Arguments = new List <string>
{
string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/java/wordcount/shakespeare.txt", fileSystem, storageAccount),
string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/java/wordcount/result/", fileSystem, storageAccount),
},
DriverMemory = "28g",
DriverCores = 4,
ExecutorMemory = "28g",
ExecutorCores = 4,
ExecutorCount = 2
};
SparkBatchJob jobCreated = client.CreateSparkBatchJob(request);
#endregion
#region Snippet:SparkBatchSample1GetSparkJob
SparkBatchJob job = client.GetSparkBatchJob(jobCreated.Id);
Debug.WriteLine($"Job is returned with name {job.Name} and state {job.State}");
#endregion
#region Snippet:SparkBatchSample1CancelSparkJob
Response operation = client.CancelSparkBatchJob(jobCreated.Id);
#endregion
}
public void CreateClient()
{
// Environment variable with the Synapse workspace endpoint.
string workspaceUrl = TestEnvironment.WorkspaceUrl;
// Environment variable with the Synapse Spark pool name.
string sparkPoolName = TestEnvironment.SparkPoolName;
#region Snippet:CreateBatchClient
// Create a new access Spark batch client using the default credential from Azure.Identity using environment variables previously set,
// including AZURE_CLIENT_ID, AZURE_CLIENT_SECRET, and AZURE_TENANT_ID.
SparkBatchClient client = new SparkBatchClient(endpoint: new Uri(workspaceUrl), sparkPoolName: sparkPoolName, credential: new DefaultAzureCredential());
#endregion
this.batchClient = client;
}
public static async Task <List <SparkBatchJob> > ListSparkBatchJobsAsync(SparkBatchClient client, bool detailed = true)
{
List <SparkBatchJob> batches = new List <SparkBatchJob>();
int from = 0;
int currentPageSize;
int pageSize = 20;
do
{
SparkBatchJobCollection page = (await client.GetSparkBatchJobsAsync(detailed: detailed, from: from, size: pageSize)).Value;
currentPageSize = page.Total;
from += currentPageSize;
batches.AddRange(page.Sessions);
} while (currentPageSize == pageSize);
return(batches);
}
public async Task TestGetSparkBatchJob()
{
SparkBatchJobCollection sparkJobs = (await SparkBatchClient.GetSparkBatchJobsAsync()).Value;
foreach (SparkBatchJob expectedSparkJob in sparkJobs.Sessions)
{
try
{
SparkBatchJob actualSparkJob = await SparkBatchClient.GetSparkBatchJobAsync(expectedSparkJob.Id);
ValidateSparkBatchJob(expectedSparkJob, actualSparkJob);
}
catch (Azure.RequestFailedException)
{
}
}
}
public void SubmitSparkJobSync()
{
#region Snippet:CreateSparkBatchClient
#if SNIPPET
// Replace the strings below with the spark, endpoint, and file system information
string sparkPoolName = "<my-spark-pool-name>";
string endpoint = "<my-endpoint-url>";
string storageAccount = "<my-storage-account-name>";
string fileSystem = "<my-storage-filesystem-name>";
#else
string sparkPoolName = TestEnvironment.SparkPoolName;
string endpoint = TestEnvironment.EndpointUrl;
string storageAccount = TestEnvironment.StorageAccountName;
string fileSystem = TestEnvironment.StorageFileSystemName;
#endif
SparkBatchClient client = new SparkBatchClient(new Uri(endpoint), sparkPoolName, new DefaultAzureCredential());
#endregion
#region Snippet:SubmitSparkBatchJob
string name = $"batch-{Guid.NewGuid()}";
string file = string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/net/wordcount/wordcount.zip", fileSystem, storageAccount);
SparkBatchJobOptions request = new SparkBatchJobOptions(name, file)
{
ClassName = "WordCount",
Arguments =
{
string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/net/wordcount/shakespeare.txt", fileSystem, storageAccount),
string.Format("abfss://{0}@{1}.dfs.core.windows.net/samples/net/wordcount/result/", fileSystem, storageAccount),
},
DriverMemory = "28g",
DriverCores = 4,
ExecutorMemory = "28g",
ExecutorCores = 4,
ExecutorCount = 2
};
SparkBatchOperation createOperation = client.StartCreateSparkBatchJob(request);
while (!createOperation.HasCompleted)
{
System.Threading.Thread.Sleep(2000);
createOperation.UpdateStatus();
}
SparkBatchJob jobCreated = createOperation.Value;
#endregion
#region Snippet:ListSparkBatchJobs
Response <SparkBatchJobCollection> jobs = client.GetSparkBatchJobs();
foreach (SparkBatchJob job in jobs.Value.Sessions)
{
Console.WriteLine(job.Name);
}
#endregion
#region Snippet:GetSparkBatchJob
SparkBatchJob retrievedJob = client.GetSparkBatchJob(jobCreated.Id);
Debug.WriteLine($"Job is returned with name {retrievedJob.Name} and state {retrievedJob.State}");
#endregion
#region Snippet:CancelSparkBatchJob
Response operation = client.CancelSparkBatchJob(jobCreated.Id);
#endregion
}
