前言:storm是三大计算框架中相对easy的一个计算框架,原理也比mapreduce的排序合并原理更容易理解

实践之构建storm集群,并测试:

[实践前需要完成:安装docker,docker-compose,maven,jdk,python等工具]

需要有一定的python基础,大数据基础和容器基础

 

1基于docker-compose服务编排工具构建 

git clone https://github.com/ziyunhx/storm-mono-docker

 

使用docker-compose启动storm集群

docker-compose up -d 【docker compose文件路径下执行】

 

2下载storm windows: apache-storm-1.2.2.zip(storm网站)

 

3 使用maven构建jar 【idea执行clean package 】

4 执行python storm.py jar app.jar com.App

笔者很久没接触大数据了,所以记录这次的采坑记

1 storm.cmd  app.jar过期【最终提交到远程集群方案如下】

python storm.py jar app.jar com.xxx.storm.test.App

2 storm.psl脚本无权限问题

set-executionpolicy remotesigned[管理员打开power-shell]

 

3 maven配置:集合了项目的数据模块
<dependency>
<groupId>org.apache.storm</groupId>
<artifactId>storm-core</artifactId>
<version>1.0.4</version> <!-- 设施版本1.2.2. -->
<exclusions>

<exclusion> <groupId>org.apache.logging.log4j</groupId>
<artifactId>log4j-slf4j-impl</artifactId>
</exclusion>

<exclusion>
<groupId>javax.servlet</groupId>
<artifactId>servlet-api</artifactId>
</exclusion>

</exclusions>
</dependency>

 

4 spring-boot 打包问题:spring-boot由于携带tomcat容器默认打出的jar ,使用storm无法执行计算拓扑,会报错无法找到主类

 

5 storm.yaml配置

由于docker映射出来很多端口发生改变,

zk如下

 storm.zookeeper.servers:
     - "localhost"
 storm.zookeeper.port: 2181

storm个组件如下

nimbus.seeds: ["storm.com"] // hosts文件配置

nimbus.thrift.port: 6627

supervisor.thrift.port: 6628

ui.port: 8080

 

 

远程发布的坑

在windows上配置zk zkPort nimbus.seeds 和nimbus.thrift.port
执行python storm.py jar app.jar Main 一直卡住
后转到linux执行:报错【报错就好办了,哈哈,卡住耗了我一天时间】
主要就是说根据配置的nimbus.seeds 无法找到nimbus这个老大

1 我的storm.yaml配置如下
storm.zookeeper.servers:
- "mystorm"
storm.zookeeper.port: 49181
# nimbus.seeds: ["host1", "host2", "host3"]
nimbus.seeds: ["mystorm"]
nimbus.thrift.port: 49627


备注:【mystorm->192.168.10.131】

2 但是zk上是节点如下172.17.0.3:6627也就是说这个ip是docker容器内部的ip
导致上述异常:NimbusLeaderNotFoundException: Could not find leader nimbus from seed hosts

[zk: 192.168.10.131:49181(CONNECTED) 0] ls /storm/nimbuses
[172.17.0.3:6627]


3 解决方式:
3.1我的机器和docker所在机器不在一个网段
3.2我在docker所在网段的机器上下了个tomcat的镜像并启动,在把拓扑jar和apache-storm-1.2.2上传到容器内部
3.3 在容器内部执行./storm jar app.jar Main
执行成功!
哈哈
容器内部的storm.yaml配置如下:zk的地址配置docker映射出来的端口以及映射机器的ip
storm.zookeeper.servers:
- "192.168.10.131"
storm.zookeeper.port: 49181
#
# nimbus.seeds: ["host1", "host2", "host3"]
nimbus.seeds: ["dockerIp172.17.0.3"]
nimbus.thrift.port: 6627




错误如下


5421 [main] INFO o.a.s.u.NimbusClient - Found leader nimbus : 172.17.0.3:6627
8428 [main] WARN o.a.s.u.StormBoundedExponentialBackoffRetry - WILL SLEEP FOR 2001ms (NOT MAX)
11433 [main] WARN o.a.s.u.StormBoundedExponentialBackoffRetry - WILL SLEEP FOR 2002ms (NOT MAX)
14439 [main] WARN o.a.s.u.StormBoundedExponentialBackoffRetry - WILL SLEEP FOR 2006ms (NOT MAX)
17445 [main] WARN o.a.s.u.StormBoundedExponentialBackoffRetry - WILL SLEEP FOR 2009ms (NOT MAX)
22461 [main] WARN o.a.s.u.StormBoundedExponentialBackoffRetry - WILL SLEEP FOR 2028ms (NOT MAX)
25467 [main] WARN o.a.s.u.NimbusClient - Ignoring exception while trying to get leader nimbus info from 192.168.10.131. will retry with a different seed host.
java.lang.RuntimeException: java.lang.RuntimeException: org.apache.storm.thrift.transport.TTransportException: java.net.NoRouteToHostException: No route to host (Host unreachable)





至此:以后要发布实时计算,只要把jar传递到tomcat所在容器,后执行即可

 

另复yaml配置参数表如下:摘自​​https://github.com/apache/storm/blob/master/conf/defaults.yaml​

 

java.library.path: "/usr/local/lib:/opt/local/lib:/usr/lib:/usr/lib64"

 

 

 

### storm.* configs are general configurations

 

# the local dir is where jars are kept

 

storm.local.dir: "storm-local"

 

storm.log4j2.conf.dir: "log4j2"

 

storm.zookeeper.servers:

 

- "localhost"

 

storm.zookeeper.port: 2181

 

storm.zookeeper.root: "/storm"

 

storm.zookeeper.session.timeout: 20000

 

storm.zookeeper.connection.timeout: 15000

 

storm.zookeeper.retry.times: 5

 

storm.zookeeper.retry.interval: 1000

 

storm.zookeeper.retry.intervalceiling.millis: 30000

 

storm.zookeeper.auth.user: null

 

storm.zookeeper.auth.password: null

 

storm.exhibitor.port: 8080

 

storm.exhibitor.poll.uripath: "/exhibitor/v1/cluster/list"

 

storm.cluster.mode: "distributed" # can be distributed or local

 

storm.local.mode.zmq: false

 

storm.thrift.transport: "org.apache.storm.security.auth.SimpleTransportPlugin"

 

storm.thrift.socket.timeout.ms: 600000

 

storm.principal.tolocal: "org.apache.storm.security.auth.DefaultPrincipalToLocal"

 

storm.group.mapping.service: "org.apache.storm.security.auth.ShellBasedGroupsMapping"

 

storm.group.mapping.service.params: null

 

storm.messaging.transport: "org.apache.storm.messaging.netty.Context"

 

storm.nimbus.retry.times: 5

 

storm.nimbus.retry.interval.millis: 2000

 

storm.nimbus.retry.intervalceiling.millis: 60000

 

storm.nimbus.zookeeper.acls.check: true

 

storm.nimbus.zookeeper.acls.fixup: true

 

 

 

storm.auth.simple-white-list.users: []

 

storm.cluster.state.store: "org.apache.storm.cluster.ZKStateStorageFactory"

 

storm.meta.serialization.delegate: "org.apache.storm.serialization.GzipThriftSerializationDelegate"

 

storm.codedistributor.class: "org.apache.storm.codedistributor.LocalFileSystemCodeDistributor"

 

storm.workers.artifacts.dir: "workers-artifacts"

 

storm.health.check.dir: "healthchecks"

 

storm.health.check.timeout.ms: 5000

 

storm.disable.symlinks: false

 

 

 

### nimbus.* configs are for the master

 

nimbus.seeds : ["localhost"]

 

nimbus.thrift.port: 6627

 

nimbus.thrift.threads: 64

 

nimbus.thrift.max_buffer_size: 1048576

 

nimbus.childopts: "-Xmx1024m"

 

nimbus.task.timeout.secs: 30

 

nimbus.supervisor.timeout.secs: 60

 

nimbus.monitor.freq.secs: 10

 

nimbus.cleanup.inbox.freq.secs: 600

 

nimbus.inbox.jar.expiration.secs: 3600

 

nimbus.code.sync.freq.secs: 120

 

nimbus.task.launch.secs: 120

 

nimbus.file.copy.expiration.secs: 600

 

nimbus.topology.validator: "org.apache.storm.nimbus.DefaultTopologyValidator"

 

topology.min.replication.count: 1

 

topology.max.replication.wait.time.sec: 60

 

nimbus.credential.renewers.freq.secs: 600

 

nimbus.queue.size: 100000

 

scheduler.display.resource: false

 

nimbus.local.assignments.backend.class: "org.apache.storm.assignments.InMemoryAssignmentBackend"

 

nimbus.assignments.service.threads: 10

 

nimbus.assignments.service.thread.queue.size: 100

 

nimbus.worker.heartbeats.recovery.strategy.class: "org.apache.storm.nimbus.TimeOutWorkerHeartbeatsRecoveryStrategy"

 

nimbus.topology.blobstore.deletion.delay.ms: 300000

 

 

 

### ui.* configs are for the master

 

ui.host: 0.0.0.0

 

ui.port: 8080

 

ui.childopts: "-Xmx768m"

 

ui.actions.enabled: true

 

ui.filter: null

 

ui.filter.params: null

 

ui.users: null

 

ui.header.buffer.bytes: 4096

 

ui.http.creds.plugin: org.apache.storm.security.auth.DefaultHttpCredentialsPlugin

 

ui.pagination: 20

 

ui.disable.http.binding: true

 

ui.disable.spout.lag.monitoring: true

 

 

 

logviewer.port: 8000

 

logviewer.childopts: "-Xmx128m"

 

logviewer.cleanup.age.mins: 10080

 

logviewer.appender.name: "A1"

 

logviewer.max.sum.worker.logs.size.mb: 4096

 

logviewer.max.per.worker.logs.size.mb: 2048

 

logviewer.disable.http.binding: true

 

logviewer.filter: null

 

logviewer.filter.params: null

 

 

 

logs.users: null

 

 

 

drpc.port: 3772

 

drpc.worker.threads: 64

 

drpc.max_buffer_size: 1048576

 

drpc.queue.size: 128

 

drpc.invocations.port: 3773

 

drpc.invocations.threads: 64

 

drpc.request.timeout.secs: 600

 

drpc.childopts: "-Xmx768m"

 

drpc.http.port: 3774

 

drpc.https.port: -1

 

drpc.https.keystore.password: ""

 

drpc.https.keystore.type: "JKS"

 

drpc.http.creds.plugin: org.apache.storm.security.auth.DefaultHttpCredentialsPlugin

 

drpc.authorizer.acl.filename: "drpc-auth-acl.yaml"

 

drpc.authorizer.acl.strict: false

 

drpc.disable.http.binding: true

 

 

 

transactional.zookeeper.root: "/transactional"

 

transactional.zookeeper.servers: null

 

transactional.zookeeper.port: null

 

 

 

## blobstore configs

 

supervisor.blobstore.class: "org.apache.storm.blobstore.NimbusBlobStore"

 

supervisor.blobstore.download.thread.count: 5

 

supervisor.blobstore.download.max_retries: 3

 

supervisor.localizer.cache.target.size.mb: 10240

 

supervisor.localizer.cleanup.interval.ms: 30000

 

 

 

nimbus.blobstore.class: "org.apache.storm.blobstore.LocalFsBlobStore"

 

nimbus.blobstore.expiration.secs: 600

 

 

 

storm.blobstore.inputstream.buffer.size.bytes: 65536

 

storm.blobstore.dependency.jar.upload.chunk.size.bytes: 1048576

 

client.blobstore.class: "org.apache.storm.blobstore.NimbusBlobStore"

 

storm.blobstore.replication.factor: 3

 

# For secure mode we would want to change this config to true

 

storm.blobstore.acl.validation.enabled: false

 

 

 

### supervisor.* configs are for node supervisors

 

# Define the amount of workers that can be run on this machine. Each worker is assigned a port to use for communication

 

supervisor.slots.ports:

 

- 6700

 

- 6701

 

- 6702

 

- 6703

 

supervisor.childopts: "-Xmx256m"

 

supervisor.run.worker.as.user: false

 

#how long supervisor will wait to ensure that a worker process is started

 

supervisor.worker.start.timeout.secs: 120

 

#how long between heartbeats until supervisor considers that worker dead and tries to restart it

 

supervisor.worker.timeout.secs: 30

 

#How many seconds to allow for graceful worker shutdown when killing workers before resorting to force kill

 

supervisor.worker.shutdown.sleep.secs: 3

 

#how frequently the supervisor checks on the status of the processes it's monitoring and restarts if necessary

 

supervisor.monitor.frequency.secs: 3

 

#how frequently the supervisor heartbeats to the cluster state (for nimbus)

 

supervisor.heartbeat.frequency.secs: 5

 

#max timeout for a node worker heartbeats when master gains leadership

 

supervisor.worker.heartbeats.max.timeout.secs: 600

 

#For topology configurable heartbeat timeout, maximum allowed heartbeat timeout.

 

worker.max.timeout.secs: 600

 

supervisor.enable: true

 

supervisor.supervisors: []

 

supervisor.supervisors.commands: []

 

supervisor.memory.capacity.mb: 4096.0

 

#By convention 1 cpu core should be about 100, but this can be adjusted if needed

 

# using 100 makes it simple to set the desired value to the capacity measurement

 

# for single threaded bolts

 

supervisor.cpu.capacity: 400.0

 

 

 

#Supervisor thrift config

 

supervisor.thrift.port: 6628

 

supervisor.queue.size: 128

 

supervisor.thrift.threads: 16

 

supervisor.thrift.max_buffer_size: 1048576

 

supervisor.thrift.socket.timeout.ms: 5000

 

 

 

### worker.* configs are for task workers

 

worker.heap.memory.mb: 768

 

worker.childopts: "-Xmx%HEAP-MEM%m -XX:+PrintGCDetails -Xloggc:artifacts/gc.log -XX:+PrintGCDateStamps -XX:+PrintGCTimeStamps -XX:+UseGCLogFileRotation -XX:NumberOfGCLogFiles=10 -XX:GCLogFileSize=1M -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=artifacts/heapdump"

 

worker.gc.childopts: ""

 

 

 

# Unlocking commercial features requires a special license from Oracle.

 

# See http://www.oracle.com/technetwork/java/javase/terms/products/index.html

 

# For this reason, profiler features are disabled by default.

 

worker.profiler.enabled: false

 

worker.profiler.childopts: "-XX:+UnlockCommercialFeatures -XX:+FlightRecorder"

 

worker.profiler.command: "flight.bash"

 

worker.heartbeat.frequency.secs: 1

 

 

 

# check whether dynamic log levels can be reset from DEBUG to INFO in workers

 

worker.log.level.reset.poll.secs: 30

 

 

 

# control how many worker receiver threads we need per worker

 

topology.worker.receiver.thread.count: 1

 

 

 

# Executor metrics reporting interval.

 

# Cause the ui only show built in metrics, we should keep sync with the built in metrics interval,

 

# also the metrics consumer's collecting interval.

 

# See topology.builtin.metrics.bucket.size.secs and storm.cluster.metrics.consumer.publish.interval.secs.

 

executor.metrics.frequency.secs: 60

 

 

 

task.heartbeat.frequency.secs: 3

 

task.refresh.poll.secs: 10

 

task.credentials.poll.secs: 30

 

 

 

# Used by workers to communicate

 

storm.messaging.netty.server_worker_threads: 1

 

storm.messaging.netty.client_worker_threads: 1

 

storm.messaging.netty.buffer_size: 5242880 #5MB buffer

 

 

 

# The netty write buffer high watermark in bytes.

 

# If the number of bytes queued in the netty's write buffer exceeds this value, the netty client will block

 

# until the value falls below the low water mark.

 

storm.messaging.netty.buffer.high.watermark: 16777216 # 16 MB

 

# The netty write buffer low watermark in bytes.

 

# Once the number of bytes queued in the write buffer exceeded the high water mark and then

 

# dropped down below this value, any blocked clients will unblock and start processing further messages.

 

storm.messaging.netty.buffer.low.watermark: 8388608 # 8 MB

 

# Since nimbus.task.launch.secs and supervisor.worker.start.timeout.secs are 120, other workers should also wait at least that long before giving up on connecting to the other worker. The reconnection period need also be bigger than storm.zookeeper.session.timeout(default is 20s), so that we can abort the reconnection when the target worker is dead.

 

storm.messaging.netty.max_wait_ms: 1000

 

storm.messaging.netty.min_wait_ms: 100

 

 

 

# If the Netty messaging layer is busy(netty internal buffer not writable), the Netty client will try to batch message as more as possible up to the size of storm.messaging.netty.transfer.batch.size bytes, otherwise it will try to flush message as soon as possible to reduce latency.

 

storm.messaging.netty.transfer.batch.size: 262144

 

# Sets the backlog value to specify when the channel binds to a local address

 

storm.messaging.netty.socket.backlog: 500

 

 

 

# By default, the Netty SASL authentication is set to false. Users can override and set it true for a specific topology.

 

storm.messaging.netty.authentication: false

 

 

 

# Default plugin to use for automatic network topology discovery

 

storm.network.topography.plugin: org.apache.storm.networktopography.DefaultRackDNSToSwitchMapping

 

 

 

# default number of seconds group mapping service will cache user group

 

storm.group.mapping.service.cache.duration.secs: 120

 

 

 

### topology.* configs are for specific executing storms

 

topology.enable.message.timeouts: true

 

topology.debug: false

 

topology.workers: 1

 

topology.acker.executors: null

 

topology.eventlogger.executors: 0

 

topology.tasks: null

 

# maximum amount of time a message has to complete before it's considered failed

 

topology.message.timeout.secs: 30

 

topology.multilang.serializer: "org.apache.storm.multilang.JsonSerializer"

 

topology.shellbolt.max.pending: 100

 

topology.skip.missing.kryo.registrations: false

 

topology.max.task.parallelism: null

 

topology.max.spout.pending: null # ideally should be larger than topology.producer.batch.size. (esp. if topology.batch.flush.interval.millis=0)

 

topology.state.synchronization.timeout.secs: 60

 

topology.stats.sample.rate: 0.05

 

topology.builtin.metrics.bucket.size.secs: 60

 

topology.fall.back.on.java.serialization: true

 

topology.worker.childopts: null

 

topology.worker.logwriter.childopts: "-Xmx64m"

 

topology.tick.tuple.freq.secs: null

 

topology.worker.shared.thread.pool.size: 4

 

 

 

# Spout Wait Strategy - employed when there is no data to produce

 

topology.spout.wait.strategy: "org.apache.storm.policy.WaitStrategyProgressive"

 

topology.spout.wait.park.microsec : 100 # park time for org.apache.storm.policy.WaitStrategyPark. Busy spins if set to 0.

 

 

 

topology.spout.wait.progressive.level1.count: 0 # number of iterations to spend in level 1 [no sleep] of WaitStrategyProgressive, before progressing to level 2

 

topology.spout.wait.progressive.level2.count: 0 # number of iterations to spend in level 2 [parkNanos(1)] of WaitStrategyProgressive, before progressing to level 3

 

topology.spout.wait.progressive.level3.sleep.millis: 1 # sleep duration for idling iterations in level 3 of WaitStrategyProgressive

 

 

 

# Bolt Wait Strategy - employed when there is no data in its receive buffer to process

 

topology.bolt.wait.strategy : "org.apache.storm.policy.WaitStrategyProgressive"

 

 

 

topology.bolt.wait.park.microsec : 100 # park time for org.apache.storm.policy.WaitStrategyPark. Busy spins if set to 0.

 

 

 

topology.bolt.wait.progressive.level1.count: 1 # number of iterations to spend in level 1 [no sleep] of WaitStrategyProgressive, before progressing to level 2

 

topology.bolt.wait.progressive.level2.count: 1000 # number of iterations to spend in level 2 [parkNanos(1)] of WaitStrategyProgressive, before progressing to level 3

 

topology.bolt.wait.progressive.level3.sleep.millis: 1 # sleep duration for idling iterations in level 3 of WaitStrategyProgressive

 

 

 

# BackPressure Wait Strategy - for any producer (spout/bolt/transfer thread) when the downstream Q is full

 

topology.backpressure.wait.strategy: "org.apache.storm.policy.WaitStrategyProgressive"

 

 

 

topology.backpressure.wait.park.microsec: 100 # park time for org.apache.storm.policy.WaitStrategyPark. Busy spins if set to 0.

 

 

 

topology.backpressure.wait.progressive.level1.count: 1 # number of iterations to spend in level 1 [no sleep] of WaitStrategyProgressive, before progressing to level 2

 

topology.backpressure.wait.progressive.level2.count: 1000 # number of iterations to spend in level 2 [parkNanos(1)] of WaitStrategyProgressive, before progressing to level 3

 

topology.backpressure.wait.progressive.level3.sleep.millis: 1 # sleep duration for idling iterations in level 3 of WaitStrategyProgressive

 

 

 

 

 

topology.backpressure.check.millis: 50 # how often to check if backpressure has relieved on executors under BP, for informing other workers to resume sending msgs to them. Must be > 0

 

topology.executor.overflow.limit: 0 # max items in overflowQ of any bolt/spout. When exceeded, worker will drop incoming messages (from the workers) destined to that overflowing spout/bolt. Set to 0 to disable overflow limiting. Enabling this may degrade perf slightly.

 

 

 

topology.error.throttle.interval.secs: 10

 

topology.max.error.report.per.interval: 5

 

topology.kryo.factory: "org.apache.storm.serialization.DefaultKryoFactory"

 

topology.tuple.serializer: "org.apache.storm.serialization.types.ListDelegateSerializer"

 

topology.trident.batch.emit.interval.millis: 500

 

topology.testing.always.try.serialize: false

 

topology.classpath: null

 

topology.environment: null

 

 

 

topology.transfer.buffer.size: 1000 # size of recv queue for transfer worker thread

 

topology.transfer.batch.size: 1 # can be no larger than half of `topology.transfer.buffer.size`

 

 

 

topology.executor.receive.buffer.size: 32768 # size of recv queue for spouts & bolts. Will be internally rounded up to next power of 2 (if not already a power of 2)

 

topology.producer.batch.size: 1 # can be no larger than half of `topology.executor.receive.buffer.size`

 

 

 

topology.batch.flush.interval.millis: 1 # Flush tuples are disabled if this is set to 0 or if (topology.producer.batch.size=1 and topology.transfer.batch.size=1).

 

topology.spout.recvq.skips: 3 # Check recvQ once every N invocations of Spout's nextTuple() [when ACKs disabled]

 

 

 

topology.disable.loadaware.messaging: false

 

topology.state.checkpoint.interval.ms: 1000

 

topology.localityaware.higher.bound: 0.8

 

topology.localityaware.lower.bound: 0.2

 

topology.serialized.message.size.metrics: false

 

 

 

# Configs for Resource Aware Scheduler

 

# topology priority describing the importance of the topology in decreasing importance starting from 0 (i.e. 0 is the highest priority and the priority importance decreases as the priority number increases).

 

# Recommended range of 0-29 but no hard limit set.

 

topology.priority: 29

 

topology.component.resources.onheap.memory.mb: 128.0

 

topology.component.resources.offheap.memory.mb: 0.0

 

topology.component.cpu.pcore.percent: 10.0

 

topology.worker.max.heap.size.mb: 768.0

 

topology.scheduler.strategy: "org.apache.storm.scheduler.resource.strategies.scheduling.DefaultResourceAwareStrategy"

 

resource.aware.scheduler.priority.strategy: "org.apache.storm.scheduler.resource.strategies.priority.DefaultSchedulingPriorityStrategy"

 

topology.ras.constraint.max.state.search: 10_000 # The maximum number of states that will be searched looking for a solution in the constraint solver strategy

 

resource.aware.scheduler.constraint.max.state.search: 100_000 # Daemon limit on maximum number of states that will be searched looking for a solution in the constraint solver strategy

 

topology.ras.one.executor.per.worker: false

 

topology.ras.one.component.per.worker: false

 

 

 

blacklist.scheduler.tolerance.time.secs: 300

 

blacklist.scheduler.tolerance.count: 3

 

blacklist.scheduler.resume.time.secs: 1800

 

blacklist.scheduler.reporter: "org.apache.storm.scheduler.blacklist.reporters.LogReporter"

 

blacklist.scheduler.strategy: "org.apache.storm.scheduler.blacklist.strategies.DefaultBlacklistStrategy"

 

blacklist.scheduler.assume.supervisor.bad.based.on.bad.slot: true

 

 

 

dev.zookeeper.path: "/tmp/dev-storm-zookeeper"

 

 

 

pacemaker.servers: []

 

pacemaker.port: 6699

 

pacemaker.base.threads: 10

 

pacemaker.max.threads: 50

 

pacemaker.client.max.threads: 2

 

pacemaker.thread.timeout: 10

 

pacemaker.childopts: "-Xmx1024m"

 

pacemaker.auth.method: "NONE"

 

pacemaker.kerberos.users: []

 

pacemaker.thrift.message.size.max: 10485760

 

 

 

#default storm daemon metrics reporter plugins

 

storm.daemon.metrics.reporter.plugins:

 

- "org.apache.storm.daemon.metrics.reporters.JmxPreparableReporter"

 

 

 

storm.metricstore.class: "org.apache.storm.metricstore.rocksdb.RocksDbStore"

 

storm.metricprocessor.class: "org.apache.storm.metricstore.NimbusMetricProcessor"

 

storm.metricstore.rocksdb.location: "storm_rocks"

 

storm.metricstore.rocksdb.create_if_missing: true

 

storm.metricstore.rocksdb.metadata_string_cache_capacity: 4000

 

storm.metricstore.rocksdb.retention_hours: 240

 

 

 

# configuration of cluster metrics consumer

 

storm.cluster.metrics.consumer.publish.interval.secs: 60

 

 

 

storm.resource.isolation.plugin: "org.apache.storm.container.cgroup.CgroupManager"

 

# Also determines whether the unit tests for cgroup runs.

 

# If storm.resource.isolation.plugin.enable is set to false the unit tests for cgroups will not run

 

storm.resource.isolation.plugin.enable: false

 

storm.cgroup.memory.enforcement.enable: false

 

storm.cgroup.inherit.cpuset.configs: false

 

 

 

# Configs for CGroup support

 

storm.cgroup.hierarchy.dir: "/cgroup/storm_resources"

 

storm.cgroup.resources:

 

- "cpu"

 

- "memory"

 

storm.cgroup.hierarchy.name: "storm"

 

storm.supervisor.cgroup.rootdir: "storm"

 

storm.cgroup.cgexec.cmd: "/bin/cgexec"

 

storm.cgroup.memory.limit.tolerance.margin.mb: 0.0

 

storm.supervisor.memory.limit.tolerance.margin.mb: 128.0

 

storm.supervisor.hard.memory.limit.multiplier: 2.0

 

storm.supervisor.hard.memory.limit.overage.mb: 2024

 

storm.supervisor.low.memory.threshold.mb: 1024

 

storm.supervisor.medium.memory.threshold.mb: 1536

 

storm.supervisor.medium.memory.grace.period.ms: 30000

 

 

 

storm.worker.min.cpu.pcore.percent: 0.0

 

 

 

storm.topology.classpath.beginning.enabled: false

 

worker.metrics:

 

"CGroupMemory": "org.apache.storm.metric.cgroup.CGroupMemoryUsage"

 

"CGroupMemoryLimit": "org.apache.storm.metric.cgroup.CGroupMemoryLimit"

 

"CGroupCpu": "org.apache.storm.metric.cgroup.CGroupCpu"

 

"CGroupCpuGuarantee": "org.apache.storm.metric.cgroup.CGroupCpuGuarantee"

 

 

 

# The number of buckets for running statistics

 

num.stat.buckets: 20

 

 

 

# Metrics v2 configuration (optional)

 

#storm.metrics.reporters:

 

# # Graphite Reporter

 

# - class: "org.apache.storm.metrics2.reporters.GraphiteStormReporter"

 

# daemons:

 

# - "supervisor"

 

# - "nimbus"

 

# - "worker"

 

# report.period: 60

 

# report.period.units: "SECONDS"

 

# graphite.host: "localhost"

 

# graphite.port: 2003

 

#

 

# # Console Reporter

 

# - class: "org.apache.storm.metrics2.reporters.ConsoleStormReporter"

 

# daemons:

 

# - "worker"

 

# report.period: 10

 

# report.period.units: "SECONDS"

 

# filter:

 

# class: "org.apache.storm.metrics2.filters.RegexFilter"

 

# expression: ".*my_component.*emitted.*"