def test_Configuration(self):
emr.Configuration(
Classification="hadoop-env",
Configurations=[
emr.Configuration(
Classification="export",
ConfigurationProperties={
"HADOOP_DATANODE_HEAPSIZE": "2048",
"HADOOP_NAMENODE_OPTS": "opts",
},
),
],
).to_dict()
with self.assertRaises(TypeError):
emr.Configuration(
Classification="hadoop-env",
Configurations=[
"illegalvalue",
emr.Configuration(
Classification="export",
ConfigurationProperties={
"HADOOP_DATANODE_HEAPSIZE": "2048",
"HADOOP_NAMENODE_OPTS": "opts",
},
),
],
).to_dict()
with self.assertRaises(TypeError):
emr.Configuration(
Classification="hadoop-env",
Configurations="invalid",
).to_dict()
def test_allow_string_cluster(self):
spot = "2"
withSpotPrice = "WithSpotPrice"
cluster = emr.Cluster(
'Cluster',
# AdditionalInfo="Additional Info",
Applications=[
emr.Application(Name="Hadoop"),
emr.Application(Name="Hive"),
emr.Application(Name="Mahout"),
emr.Application(Name="Pig"),
emr.Application(Name="Spark")
],
BootstrapActions=[
emr.BootstrapActionConfig(
Name='Dummy bootstrap action',
ScriptBootstrapAction=emr.ScriptBootstrapActionConfig(
Path='file:/usr/share/aws/emr/scripts/install-hue',
Args=["dummy", "parameter"]))
],
Configurations=[
emr.Configuration(Classification="core-site",
ConfigurationProperties={
'hadoop.security.groups.cache.secs':
'250'
})
],
Instances=emr.JobFlowInstancesConfig(
Ec2KeyName="KeyName",
Ec2SubnetId="SubnetId",
MasterInstanceGroup=emr.InstanceGroupConfigProperty(
InstanceCount="1",
InstanceType=M4_LARGE,
AutoScalingPolicy=emr.AutoScalingPolicy(
Constraints=emr.ScalingConstraints(MinCapacity="1",
MaxCapacity="3"),
Rules=self.generate_rules("MasterAutoScalingPolicy")),
),
CoreInstanceGroup=emr.InstanceGroupConfigProperty(
Name="Core Instance",
BidPrice=If(withSpotPrice, Ref(spot), Ref("AWS::NoValue")),
Market=If(withSpotPrice, "SPOT", "ON_DEMAND"),
InstanceCount="1",
InstanceType=M4_LARGE,
AutoScalingPolicy=emr.AutoScalingPolicy(
Constraints=emr.ScalingConstraints(MinCapacity="1",
MaxCapacity="3"),
Rules=self.generate_rules("CoreAutoScalingPolicy"),
)),
),
JobFlowRole="EMRJobFlowRole",
LogUri="s3://cluster-logs",
Name="EMR Cluster",
ReleaseLabel="emr-5.5.0",
ServiceRole="EMRServiceRole",
AutoScalingRole="EMR_AutoScaling_DefaultRole",
VisibleToAllUsers="true",
Tags=Tags(Name="EMR Sample Cluster"))
cluster.to_dict()
# EMR Cluster Resource
cluster = template.add_resource(
emr.Cluster("EMRSampleCluster",
Name="EMR Sample Cluster",
ReleaseLabel='emr-4.4.0',
BootstrapActions=[
emr.BootstrapActionConfig(
Name='Dummy bootstrap action',
ScriptBootstrapAction=emr.ScriptBootstrapActionConfig(
Path='/bin/sh', Args=['echo', 'Hello World']))
],
Configurations=[
emr.Configuration(Classification="core-site",
ConfigurationProperties={
'hadoop.security.groups.cache.secs':
'250'
}),
emr.Configuration(
Classification="mapred-site",
ConfigurationProperties={
'mapred.tasktracker.map.tasks.maximum': '2',
'mapreduce.map.sort.spill.percent': '90',
'mapreduce.tasktracker.reduce.tasks.maximum': '5'
}),
emr.Configuration(Classification="hadoop-env",
Configurations=[
emr.Configuration(
Classification="export",
ConfigurationProperties={
"HADOOP_DATANODE_HEAPSIZE":
def test_allow_string_cluster(self):
cluster_security_configuration = emr.SecurityConfiguration(
'emrsecurityconfiguration',
Name="EMRSecurityConfiguration",
SecurityConfiguration=security_configuration)
spot = "2"
withSpotPrice = "WithSpotPrice"
cluster = emr.Cluster(
'Cluster',
# AdditionalInfo="Additional Info",
Applications=[
emr.Application(Name="Hadoop"),
emr.Application(Name="Hive"),
emr.Application(Name="Mahout"),
emr.Application(Name="Pig"),
emr.Application(Name="Spark")
],
BootstrapActions=[
emr.BootstrapActionConfig(
Name='Dummy bootstrap action',
ScriptBootstrapAction=emr.ScriptBootstrapActionConfig(
Path='file:/usr/share/aws/emr/scripts/install-hue',
Args=["dummy", "parameter"]))
],
Configurations=[
emr.Configuration(Classification="core-site",
ConfigurationProperties={
'hadoop.security.groups.cache.secs':
'250'
})
],
Instances=emr.JobFlowInstancesConfig(
Ec2KeyName="KeyName",
Ec2SubnetId="SubnetId",
MasterInstanceGroup=emr.InstanceGroupConfigProperty(
InstanceCount="1",
InstanceType=M4_LARGE,
AutoScalingPolicy=emr.AutoScalingPolicy(
Constraints=emr.ScalingConstraints(MinCapacity="1",
MaxCapacity="3"),
Rules=self.generate_rules("MasterAutoScalingPolicy")),
),
CoreInstanceGroup=emr.InstanceGroupConfigProperty(
Name="Core Instance",
BidPrice=If(withSpotPrice, Ref(spot), Ref("AWS::NoValue")),
Market=If(withSpotPrice, "SPOT", "ON_DEMAND"),
InstanceCount="1",
InstanceType=M4_LARGE,
AutoScalingPolicy=emr.AutoScalingPolicy(
Constraints=emr.ScalingConstraints(MinCapacity="1",
MaxCapacity="3"),
Rules=self.generate_rules("CoreAutoScalingPolicy"),
)),
),
JobFlowRole="EMRJobFlowRole",
LogUri="s3://cluster-logs",
Name="EMR Cluster",
ReleaseLabel="emr-5.5.0",
SecurityConfiguration=Ref(cluster_security_configuration),
ServiceRole="EMRServiceRole",
AutoScalingRole="EMR_AutoScaling_DefaultRole",
VisibleToAllUsers="true",
Tags=Tags(Name="EMR Sample Cluster"))
cluster.to_dict()
autoscale_policy = emr.AutoScalingPolicy(
Constraints=emr.ScalingConstraints(MinCapacity=0, MaxCapacity=5),
Rules=[
emr.ScalingRule(
Name='ScaleUpContainerPending',
Description='Scale up on over-provisioned '
'containers',
Action=emr.ScalingAction(
SimpleScalingPolicyConfiguration=emr.
SimpleScalingPolicyConfiguration(
AdjustmentType=emr.CHANGE_IN_CAPACITY,
CoolDown=300,
ScalingAdjustment=1)),
Trigger=emr.ScalingTrigger(
CloudWatchAlarmDefinition=emr.
CloudWatchAlarmDefinition(
ComparisonOperator='GREATER_THAN',
MetricName='ContainerPendingRatio',
Period=300,
Threshold=0.75,
Dimensions=[
emr.MetricDimension(Key='JobFlowId',
Value='${emr.clusterId}')
]))),
emr.ScalingRule(
Name='ScaleUpMemory',
Description='Scale up on low memory',
Action=emr.ScalingAction(
SimpleScalingPolicyConfiguration=emr.
SimpleScalingPolicyConfiguration(
AdjustmentType='CHANGE_IN_CAPACITY',
CoolDown=300,
ScalingAdjustment=1)),
Trigger=emr.ScalingTrigger(
CloudWatchAlarmDefinition=emr.
CloudWatchAlarmDefinition(
ComparisonOperator='LESS_THAN',
MetricName='YARNMemoryAvailablePercentage',
Period=300,
Threshold=15,
Dimensions=[
emr.MetricDimension(Key='JobFlowId',
Value='${emr.clusterId}')
]))),
emr.ScalingRule(
Name='ScaleDownMemory',
Description='Scale down on high memory',
Action=emr.ScalingAction(
SimpleScalingPolicyConfiguration=emr.
SimpleScalingPolicyConfiguration(
AdjustmentType=emr.CHANGE_IN_CAPACITY,
CoolDown=300,
ScalingAdjustment=-1)),
Trigger=emr.ScalingTrigger(
CloudWatchAlarmDefinition=emr.
CloudWatchAlarmDefinition(
ComparisonOperator='GREATER_THAN',
MetricName='YARNMemoryAvailablePercentage',
Period=300,
Threshold=75,
Dimensions=[
emr.MetricDimension(Key='JobFlowId',
Value='${emr.clusterId}')
])))
])
emr.InstanceGroupConfig('TaskInstanceGroup',
AutoScalingPolicy=autoscale_policy,
InstanceCount=0,
InstanceType=M4_LARGE,
InstanceRole='TASK',
Market='ON_DEMAND',
Name='Task Instance',
JobFlowId=Ref(cluster))
Path=
's3://nicor-dev/deployments/emr/bootstrap_actions/bootstrap_jupyter.sh',
Args=['testemr', 's3://nicor-dev/jupyter-notebooks/']))
],
Configurations=[
emr.Configuration(
Classification="spark-env",
Configurations=[
emr.Configuration(
Classification="export",
ConfigurationProperties={
"PYSPARK_PYTHON":
os.path.join('/home/hadoop/miniconda',
'bin/python'),
"PYTHONPATH":
os.path.join('/home/hadoop/miniconda',
'bin/python') +
":/usr/lib/spark/python/:$PYTHONPATH",
"PYSPARK_DRIVER_PYTHON":
os.path.join('/home/hadoop/miniconda',
'bin/python'),
"SPARK_HOME":
"/usr/lib/spark",
"PYTHONHASHSEED":
"123"
})
]),
],
Applications=[
emr.Application(Name=app) for app in cfg['applications']
],
VisibleToAllUsers='true',
Name="EMR Sample Cluster",
ReleaseLabel="emr-4.4.0",
SecurityConfiguration=Ref(security_config),
BootstrapActions=[
emr.BootstrapActionConfig(
Name="Dummy bootstrap action",
ScriptBootstrapAction=emr.ScriptBootstrapActionConfig(
Path="file:/usr/share/aws/emr/scripts/install-hue",
Args=["dummy", "parameter"],
),
)
],
Configurations=[
emr.Configuration(
Classification="core-site",
ConfigurationProperties={
"hadoop.security.groups.cache.secs": "250"
},
),
emr.Configuration(
Classification="mapred-site",
ConfigurationProperties={
"mapred.tasktracker.map.tasks.maximum": "2",
"mapreduce.map.sort.spill.percent": "90",
"mapreduce.tasktracker.reduce.tasks.maximum": "5",
},
),
emr.Configuration(
Classification="hadoop-env",
Configurations=[
emr.Configuration(
Classification="export",
