kylin-2.6.3使用Spark构建Cube

Apache Kylin™是一个开源的分布式分析引擎,提供Hadoop/Spark之上的SQL查询接口及多维分析(OLAP)能力以支持超大规模数据,最初由eBay Inc. 开发并贡献至开源社区。它能在亚秒内查询巨大的Hive表。
下面是单机安装采坑记,直接上配置和问题解决。
找一台干净的机器,把hadoop hive hbase从原有节点分别拷贝一份,主要目的是配置文件,可以不在kylin所在机器启动相关进程。
开源版本搭建,非整合HDP和CDH。
个别问题解决参考其他博客。
官网http://kylin.apache.org/cn/docs/
MapReduce构建Cube的问题也已解决,所以使用MapReduce构建Cube也是正常的。

版本

java-1.8.0-openjdk-1.8.0.191.b12-1.el7_6.x86_64  
hadoop-2.8.5【官网下载】  
hbase-1.4.10【官网下载】  
hive-2.3.5【官网下载】  
apache-kylin-2.6.3-bin-hbase1x【官网下载】  
spark-2.3.2【$KYLIN_HOME/spark 通过$KYLIN_HOME/bin/download-spark.sh下载】  
spark-2.3.2-yarn-shuffle.jar【https://github.com/apache/spark/releases/tag/v2.3.2下载Source code自行编译(Oracle JDK1.8.0_181 hadoop2.7.3)】

环境变量

JAVA_HOME等。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
export HADOOP_HOME=/home/admin/hadoop-2.8.5
export PATH="$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$PATH"
export HADOOP_CONF_DIR=$HADOOP_HOME/etc/hadoop

export HBASE_HOME=/home/admin/hbase-1.4.10
export PATH="$HBASE_HOME/bin:$HBASE_HOME/sbin:$PATH"

export HIVE_HOME=/home/admin/hive-2.3.5
export PATH="$HIVE_HOME/bin:$HIVE_HOME/sbin:$PATH"
export HCAT_HOME=$HIVE_HOME/hcatalog

export KYLIN_HOME=/home/admin/kylin-2.6.3
export KYLIN_CONF_HOME=$KYLIN_HOME/conf
export CATALINA_HOME=$KYLIN_HOME/tomcat

export PATH=:$PATH:$KYLIN_HOME/bin:$CATALINE_HOME/bin
export tomcat_root=$KYLIN_HOME/tomcat
export hive_dependency=$HIVE_HOME/conf:$HIVE_HOME/lib/*:$HCAT_HOME/share/hcatalog/hive-hcatalog-core-2.3.5.jar

下载spark和上传spark的依赖包

从v2.6.1开始, Kylin不再包含Spark二进制包;需要另外下载Spark,然后设置SPARK_HOME系统变量到Spark安装目录(可以不设置,详见$KYLIN_HOME/bin/find-spark-dependency.sh)

使用脚本下载Spark[下载后的目录位于$KYLIN_HOME/spark]:

1
$ $KYLIN_HOME/bin/download-spark.sh

把Spark依赖的jars打包成一个jar上传到HDFS上面,这里参照官网,另外打包成zip也是可以的:

1
2
3
$ jar cv0f spark-libs.jar -C $KYLIN_HOME/spark/jars/ .
$ hadoop fs -mkdir -p /kylin/spark/
$ hadoop fs -put spark-libs.jar /kylin/spark/

并在$KYLIN_HOME/conf/kylin.properties里面或者$KYLIN_HOME/spark/conf/spark-defaults.conf里面进行配置:
分别是kylin.engine.spark-conf.spark.yarn.archive和
spark.yarn.archive,配置其一即可。

主要配置

$KYLIN_HOME/conf/kylin.properties

有些虽然解除了注释#,但并非必须的,默认值也可以;
另外在环境变量都设置好的情况下,比如HADOOP_CONF_DIR,不需要再配置kylin.env.hadoop-conf-dir。

说明:
SPARK ENGINE CONFIGS下面kylin.engine.spark-conf.xxxx
后面xxxx这种配置,完全可以在$KYLIN_HOME/spark/conf下的spark-defaults.conf里面进行配置。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
#
#### METADATA | ENV ###
#
## The metadata store in hbase
kylin.metadata.url=kylin_metadata@hbase
#
## metadata cache sync retry times
kylin.metadata.sync-retries=3
#
## Working folder in HDFS, better be qualified absolute path, make sure user has the right permission to this directory
kylin.env.hdfs-working-dir=/kylin
#
## DEV|QA|PROD. DEV will turn on some dev features, QA and PROD has no difference in terms of functions.
kylin.env=QA
#
## kylin zk base path
kylin.env.zookeeper-base-path=/kylin
#
#### SERVER | WEB | RESTCLIENT ###
#
## Kylin server mode, valid value [all, query, job]
kylin.server.mode=all
#
## List of web servers in use, this enables one web server instance to sync up with other servers.
kylin.server.cluster-servers=192.168.100.2:7070
#
## Display timezone on UI,format like[GMT+N or GMT-N]
kylin.web.timezone=GMT+8
#
## Timeout value for the queries submitted through the Web UI, in milliseconds
kylin.web.query-timeout=300000
#
kylin.web.cross-domain-enabled=true
#
##allow user to export query result
kylin.web.export-allow-admin=true
kylin.web.export-allow-other=true
#
## Hide measures in measure list of cube designer, separate by comma
kylin.web.hide-measures=RAW
#
##max connections of one route
#kylin.restclient.connection.default-max-per-route=20
#
##max connections of one rest-client
#kylin.restclient.connection.max-total=200
#
#### PUBLIC CONFIG ###
#kylin.engine.default=2
#kylin.storage.default=2
#kylin.web.hive-limit=20
#kylin.web.help.length=4
#kylin.web.help.0=start|Getting Started|http://kylin.apache.org/docs/tutorial/kylin_sample.html
#kylin.web.help.1=odbc|ODBC Driver|http://kylin.apache.org/docs/tutorial/odbc.html
#kylin.web.help.2=tableau|Tableau Guide|http://kylin.apache.org/docs/tutorial/tableau_91.html
#kylin.web.help.3=onboard|Cube Design Tutorial|http://kylin.apache.org/docs/howto/howto_optimize_cubes.html
#kylin.web.link-streaming-guide=http://kylin.apache.org/
#kylin.htrace.show-gui-trace-toggle=false
#kylin.web.link-hadoop=
#kylin.web.link-diagnostic=
#kylin.web.contact-mail=
#kylin.server.external-acl-provider=
#
#### SOURCE ###
#
## Hive client, valid value [cli, beeline]
#kylin.source.hive.client=cli
#
## Absolute path to beeline shell, can be set to spark beeline instead of the default hive beeline on PATH
#kylin.source.hive.beeline-shell=beeline
#
## Parameters for beeline client, only necessary if hive client is beeline
##kylin.source.hive.beeline-params=-n root --hiveconf hive.security.authorization.sqlstd.confwhitelist.append='mapreduce.job.*|dfs.*' -u jdbc:hive2://localhost:10000
#
## While hive client uses above settings to read hive table metadata,
## table operations can go through a separate SparkSQL command line, given SparkSQL connects to the same Hive metastore.
#kylin.source.hive.enable-sparksql-for-table-ops=false
##kylin.source.hive.sparksql-beeline-shell=/path/to/spark-client/bin/beeline
##kylin.source.hive.sparksql-beeline-params=-n root --hiveconf hive.security.authorization.sqlstd.confwhitelist.append='mapreduce.job.*|dfs.*' -u jdbc:hive2://localhost:10000
#
#kylin.source.hive.keep-flat-table=false
#
## Hive database name for putting the intermediate flat tables
kylin.source.hive.database-for-flat-table=kylin_flat_db
#
## Whether redistribute the intermediate flat table before building
#kylin.source.hive.redistribute-flat-table=true
#
#
#### STORAGE ###
#
## The storage for final cube file in hbase
kylin.storage.url=hbase
#
## The prefix of hbase table
kylin.storage.hbase.table-name-prefix=KYLIN_
#
## The namespace for hbase storage
kylin.storage.hbase.namespace=default
#
## Compression codec for htable, valid value [none, snappy, lzo, gzip, lz4]
kylin.storage.hbase.compression-codec=none
#
## HBase Cluster FileSystem, which serving hbase, format as hdfs://hbase-cluster:8020
## Leave empty if hbase running on same cluster with hive and mapreduce
##kylin.storage.hbase.cluster-fs=
#
## The cut size for hbase region, in GB.
kylin.storage.hbase.region-cut-gb=5
#
## The hfile size of GB, smaller hfile leading to the converting hfile MR has more reducers and be faster.
## Set 0 to disable this optimization.
kylin.storage.hbase.hfile-size-gb=2
#
#kylin.storage.hbase.min-region-count=1
#kylin.storage.hbase.max-region-count=500
#
## Optional information for the owner of kylin platform, it can be your team's email
## Currently it will be attached to each kylin's htable attribute
#kylin.storage.hbase.owner-tag=whoami@kylin.apache.org
#
#kylin.storage.hbase.coprocessor-mem-gb=3
#
## By default kylin can spill query's intermediate results to disks when it's consuming too much memory.
## Set it to false if you want query to abort immediately in such condition.
#kylin.storage.partition.aggr-spill-enabled=true
#
## The maximum number of bytes each coprocessor is allowed to scan.
## To allow arbitrary large scan, you can set it to 0.
#kylin.storage.partition.max-scan-bytes=3221225472
#
## The default coprocessor timeout is (hbase.rpc.timeout * 0.9) / 1000 seconds,
## You can set it to a smaller value. 0 means use default.
## kylin.storage.hbase.coprocessor-timeout-seconds=0
#
## clean real storage after delete operation
## if you want to delete the real storage like htable of deleting segment, you can set it to true
#kylin.storage.clean-after-delete-operation=false
#
#### JOB ###
#
## Max job retry on error, default 0: no retry
kylin.job.retry=0
#
## Max count of concurrent jobs running
kylin.job.max-concurrent-jobs=10
#
## The percentage of the sampling, default 100%
#kylin.job.sampling-percentage=100
#
## If true, will send email notification on job complete
##kylin.job.notification-enabled=true
##kylin.job.notification-mail-enable-starttls=true
##kylin.job.notification-mail-host=smtp.office365.com
##kylin.job.notification-mail-port=587
##kylin.job.notification-mail-username=kylin@example.com
##kylin.job.notification-mail-password=mypassword
##kylin.job.notification-mail-sender=kylin@example.com
#
#
#### ENGINE ###
#
## Time interval to check hadoop job status
kylin.engine.mr.yarn-check-interval-seconds=10
#
#kylin.engine.mr.reduce-input-mb=500
#
#kylin.engine.mr.max-reducer-number=500
#
#kylin.engine.mr.mapper-input-rows=1000000
#
## Enable dictionary building in MR reducer
#kylin.engine.mr.build-dict-in-reducer=true
#
## Number of reducers for fetching UHC column distinct values
#kylin.engine.mr.uhc-reducer-count=3
#
## Whether using an additional step to build UHC dictionary
#kylin.engine.mr.build-uhc-dict-in-additional-step=false
#
#
#### CUBE | DICTIONARY ###
#
#kylin.cube.cuboid-scheduler=org.apache.kylin.cube.cuboid.DefaultCuboidScheduler
#kylin.cube.segment-advisor=org.apache.kylin.cube.CubeSegmentAdvisor
#
## 'auto', 'inmem', 'layer' or 'random' for testing
#kylin.cube.algorithm=layer
#
## A smaller threshold prefers layer, a larger threshold prefers in-mem
#kylin.cube.algorithm.layer-or-inmem-threshold=7
#
## auto use inmem algorithm:
## 1, cube planner optimize job
## 2, no source record
#kylin.cube.algorithm.inmem-auto-optimize=true
#
#kylin.cube.aggrgroup.max-combination=32768
#
#kylin.snapshot.max-mb=300
#
#kylin.cube.cubeplanner.enabled=true
#kylin.cube.cubeplanner.enabled-for-existing-cube=true
#kylin.cube.cubeplanner.expansion-threshold=15.0
#kylin.cube.cubeplanner.recommend-cache-max-size=200
#kylin.cube.cubeplanner.mandatory-rollup-threshold=1000
#kylin.cube.cubeplanner.algorithm-threshold-greedy=8
#kylin.cube.cubeplanner.algorithm-threshold-genetic=23
#
#
#### QUERY ###
#
## Controls the maximum number of bytes a query is allowed to scan storage.
## The default value 0 means no limit.
## The counterpart kylin.storage.partition.max-scan-bytes sets the maximum per coprocessor.
#kylin.query.max-scan-bytes=0
#
#kylin.query.cache-enabled=true
#
## Controls extras properties for Calcite jdbc driver
## all extras properties should undder prefix "kylin.query.calcite.extras-props."
## case sensitive, default: true, to enable case insensitive set it to false
## @see org.apache.calcite.config.CalciteConnectionProperty.CASE_SENSITIVE
#kylin.query.calcite.extras-props.caseSensitive=true
## how to handle unquoted identity, defualt: TO_UPPER, available options: UNCHANGED, TO_UPPER, TO_LOWER
## @see org.apache.calcite.config.CalciteConnectionProperty.UNQUOTED_CASING
#kylin.query.calcite.extras-props.unquotedCasing=TO_UPPER
## quoting method, default: DOUBLE_QUOTE, available options: DOUBLE_QUOTE, BACK_TICK, BRACKET
## @see org.apache.calcite.config.CalciteConnectionProperty.QUOTING
#kylin.query.calcite.extras-props.quoting=DOUBLE_QUOTE
## change SqlConformance from DEFAULT to LENIENT to enable group by ordinal
## @see org.apache.calcite.sql.validate.SqlConformance.SqlConformanceEnum
#kylin.query.calcite.extras-props.conformance=LENIENT
#
## TABLE ACL
#kylin.query.security.table-acl-enabled=true
#
## Usually should not modify this
#kylin.query.interceptors=org.apache.kylin.rest.security.TableInterceptor
#
#kylin.query.escape-default-keyword=false
#
## Usually should not modify this
#kylin.query.transformers=org.apache.kylin.query.util.DefaultQueryTransformer,org.apache.kylin.query.util.KeywordDefaultDirtyHack
#
#### SECURITY ###
#
## Spring security profile, options: testing, ldap, saml
## with "testing" profile, user can use pre-defined name/pwd like KYLIN/ADMIN to login
#kylin.security.profile=testing
#
## Admin roles in LDAP, for ldap and saml
#kylin.security.acl.admin-role=admin
#
## LDAP authentication configuration
#kylin.security.ldap.connection-server=ldap://ldap_server:389
#kylin.security.ldap.connection-username=
#kylin.security.ldap.connection-password=
#
## LDAP user account directory;
#kylin.security.ldap.user-search-base=
#kylin.security.ldap.user-search-pattern=
#kylin.security.ldap.user-group-search-base=
#kylin.security.ldap.user-group-search-filter=(|(member={0})(memberUid={1}))
#
## LDAP service account directory
#kylin.security.ldap.service-search-base=
#kylin.security.ldap.service-search-pattern=
#kylin.security.ldap.service-group-search-base=
#
### SAML configurations for SSO
## SAML IDP metadata file location
#kylin.security.saml.metadata-file=classpath:sso_metadata.xml
#kylin.security.saml.metadata-entity-base-url=https://hostname/kylin
#kylin.security.saml.keystore-file=classpath:samlKeystore.jks
#kylin.security.saml.context-scheme=https
#kylin.security.saml.context-server-name=hostname
#kylin.security.saml.context-server-port=443
#kylin.security.saml.context-path=/kylin
#
#### SPARK ENGINE CONFIGS ###
#
## Hadoop conf folder, will export this as "HADOOP_CONF_DIR" to run spark-submit
## This must contain site xmls of core, yarn, hive, and hbase in one folder
# kylin.env.hadoop-conf-dir=/home/admin/hadoop-2.8.5/etc/hadoop
#
## Estimate the RDD partition numbers
kylin.engine.spark.rdd-partition-cut-mb=100
#
## Minimal partition numbers of rdd
#kylin.engine.spark.min-partition=1
#
## Max partition numbers of rdd
#kylin.engine.spark.max-partition=5000
#
## Spark conf (default is in spark/conf/spark-defaults.conf)
kylin.engine.spark-conf.spark.master=yarn
kylin.engine.spark-conf.spark.submit.deployMode=cluster
kylin.engine.spark-conf.spark.yarn.queue=default
kylin.engine.spark-conf.spark.driver.memory=2G
kylin.engine.spark-conf.spark.executor.memory=1G
kylin.engine.spark-conf.spark.executor.instances=40
kylin.engine.spark-conf.spark.yarn.executor.memoryOverhead=1024
kylin.engine.spark-conf.spark.shuffle.service.enabled=true
kylin.engine.spark-conf.spark.eventLog.enabled=true
kylin.engine.spark-conf.spark.eventLog.dir=hdfs\:///kylin/spark-history
kylin.engine.spark-conf.spark.history.fs.logDirectory=hdfs\:///kylin/spark-history
kylin.engine.spark-conf.spark.hadoop.yarn.timeline-service.enabled=false
#
#### Spark conf for specific job
kylin.engine.spark-conf-mergedict.spark.executor.memory=2G
kylin.engine.spark-conf-mergedict.spark.memory.fraction=0.2
#
## manually upload spark-assembly jar to HDFS and then set this property will avoid repeatedly uploading jar at runtime
kylin.engine.spark-conf.spark.yarn.archive=hdfs://master:9000/kylin/spark/spark-libs.jar
##kylin.engine.spark-conf.spark.io.compression.codec=org.apache.spark.io.SnappyCompressionCodec
#
## uncomment for HDP
##kylin.engine.spark-conf.spark.driver.extraJavaOptions=-Dhdp.version=current
##kylin.engine.spark-conf.spark.yarn.am.extraJavaOptions=-Dhdp.version=current
##kylin.engine.spark-conf.spark.executor.extraJavaOptions=-Dhdp.version=current
#
#
#### QUERY PUSH DOWN ###
#
##kylin.query.pushdown.runner-class-name=org.apache.kylin.query.adhoc.PushDownRunnerJdbcImpl
#
##kylin.query.pushdown.update-enabled=false
##kylin.query.pushdown.jdbc.url=jdbc:hive2://sandbox:10000/default
##kylin.query.pushdown.jdbc.driver=org.apache.hive.jdbc.HiveDriver
##kylin.query.pushdown.jdbc.username=hive
##kylin.query.pushdown.jdbc.password=
#
##kylin.query.pushdown.jdbc.pool-max-total=8
##kylin.query.pushdown.jdbc.pool-max-idle=8
##kylin.query.pushdown.jdbc.pool-min-idle=0
#
#### JDBC Data Source
##kylin.source.jdbc.connection-url=
##kylin.source.jdbc.driver=
##kylin.source.jdbc.dialect=
##kylin.source.jdbc.user=
##kylin.source.jdbc.pass=
##kylin.source.jdbc.sqoop-home=
##kylin.source.jdbc.filed-delimiter=|


kylin.job.jar=$KYLIN_HOME/lib/kylin-job-2.6.3.jar
kylin.coprocessor.local.jar=$KYLIN_HOME/lib/kylin-coprocessor-2.6.3.jar

检查运行环境

1
$ $KYLIN_HOME/bin/check-env.sh

trouble shooting

日志目录:$KYLIN_HOME/logs下面有kylin.log、kylin.out 还有gc日志

问题0:默认使用hbase配置里的ZK-UnknownHostException

检查运行环境就会报。

1
Caused by: java.net.UnknownHostException: 192.168.100.5:12181: invalid IPv6 address

说明hbase里zk的配置是host:port;kylin读取zk配置的方式是单独读取host和port,修改为host和port分开配置。

修改hbase的配置文件,kylin所在机器的配置文件修改一下就行,为了保持一致,全部修改下再重启hbase。

$HBASE_HOME/conf/hbase-site.xml

1
2
3
4
5
6
7
8
<property>
<name>hbase.zookeeper.property.clientPort</name>
<value>12181</value>
</property>
<property>
<name>hbase.zookeeper.quorum</name>
<value>192.168.100.5,192.168.100.6,192.168.100.7</value>
</property>

环境正常后,加载示例数据并启动kylin server

1
2
3
4
5
6
$ $KYLIN_HOME/bin/sample.sh
$ $KYLIN_HOME/bin/kylin.sh start
...
A new Kylin instance is started by admin. To stop it, run 'kylin.sh stop'
Check the log at /home/admin/kylin-2.6.3/logs/kylin.log
Web UI is at http://slave1:7070/kylin

问题1:访问http://192.168.100.2:7070/kylin 404

1
2
3
4
5
SEVERE: Failed to load keystore type JKS with path conf/.keystore due to /home/admin/kylin-2.6.3/tomcat/conf/.keystore (No such file or directory)
java.io.FileNotFoundException: /home/admin/kylin-2.6.3/tomcat/conf/.keystore (No such file or directory)

SEVERE: Context [/kylin] startup failed due to previous errors
Invalid character found in method name. HTTP method names must be tokens

https CA证书问题,注释掉$KYLIN_HOME/tomcat/conf/server.xml里https的部分。

1
2
3
4
5
6
<!--
<Connector port="7443" protocol="org.apache.coyote.http11.Http11Protocol"
maxThreads="150" SSLEnabled="true" scheme="https" secure="true"
keystoreFile="conf/.keystore" keystorePass="changeit"
clientAuth="false" sslProtocol="TLS" />
-->

停止kylin server然后再启动,处理其他问题类似。

1
2
$ $KYLIN_HOME/bin/kylin.sh stop
$ $KYLIN_HOME/bin/kylin.sh start

问题2:jackson jar包冲突

1
2
2019-08-13T15:44:18,486 ERROR [localhost-startStop-1] org.springframework.web.context.ContextLoader - Context initialization failed
org.springframework.beans.factory.BeanCreationException: Error creating bean with name 'org.springframework.web.servlet.mvc.method.annotation.RequestMappingHandlerAdapter': Instantiation of bean failed; nested exception is org.springframework.beans.BeanInstantiationException: Failed to instantiate [org.springframework.web.servlet.mvc.method.annotation.RequestMappingHandlerAdapter]: Constructor threw exception; nested exception is java.lang.ClassCastException: com.fasterxml.jackson.datatype.joda.JodaModule cannot be cast to com.fasterxml.jackson.databind.Module

hive的依赖版本:

/home/admin/hive-2.3.5/lib/jackson-datatype-joda-2.4.6.jar

kylin的依赖版本:

/home/admin/kylin-2.6.3/tomcat/webapps/kylin/WEB-INF/lib/jackson-databind-2.9.5.jar 

将hive的删除[重命名]。

1
mv $HIVE_HOME/lib/jackson-datatype-joda-2.4.6.jar $HIVE_HOME/lib/jackson-datatype-joda-2.4.6.jarback

使用默认ADMIN/KYLIN[全大写]进行登录

用户名密码见$KYLIN_HOME/tomcat/webapps/kylin/WEB-INF/classes/kylinSecurity.xml。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
<bean class="org.springframework.security.core.userdetails.User" id="adminUser">
<constructor-arg value="ADMIN"/>
<constructor-arg
value="$2a$10$o3ktIWsGYxXNuUWQiYlZXOW5hWcqyNAFQsSSCSEWoC/BRVMAUjL32"/>
<constructor-arg ref="adminAuthorities"/>
</bean>
<bean class="org.springframework.security.core.userdetails.User" id="modelerUser">
<constructor-arg value="MODELER"/>
<constructor-arg
value="$2a$10$Le5ernTeGNIARwMJsY0WaOLioNQdb0QD11DwjeyNqqNRp5NaDo2FG"/>
<constructor-arg ref="modelerAuthorities"/>
</bean>
<bean class="org.springframework.security.core.userdetails.User" id="analystUser">
<constructor-arg value="ANALYST"/>
<constructor-arg
value="$2a$10$s4INO3XHjPP5Vm2xH027Ce9QeXWdrfq5pvzuGr9z/lQmHqi0rsbNi"/>
<constructor-arg ref="analystAuthorities"/>
</bean>

问题3:使用MapReduce构建示例Cube报 10020 failed

1
2
3
4
5
6
7
8
9
10
org.apache.kylin.engine.mr.exception.MapReduceException: Exception: java.net.ConnectException: Call From slave1/192.168.100.2 to 0.0.0.0:10020 failed on connection exception: java.net.ConnectException: Connection refused; For more details see:  http://wiki.apache.org/hadoop/ConnectionRefused
java.net.ConnectException: Call From slave1/192.168.100.2 to 0.0.0.0:10020 failed on connection exception: java.net.ConnectException: Connection refused; For more details see: http://wiki.apache.org/hadoop/ConnectionRefused
at org.apache.kylin.engine.mr.common.MapReduceExecutable.doWork(MapReduceExecutable.java:173)
at org.apache.kylin.job.execution.AbstractExecutable.execute(AbstractExecutable.java:167)
at org.apache.kylin.job.execution.DefaultChainedExecutable.doWork(DefaultChainedExecutable.java:71)
at org.apache.kylin.job.execution.AbstractExecutable.execute(AbstractExecutable.java:167)
at org.apache.kylin.job.impl.threadpool.DefaultScheduler$JobRunner.run(DefaultScheduler.java:114)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)

10020端口,需要启动MapReduce JobHistory Server。

查看hadoop的mapred-site.xml有无以下配置:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
<configuration>
<property>
<name>mapreduce.framework.name</name>
<value>yarn</value>
</property>
<property>
<name>mapreduce.jobhistory.address</name>
<value>master:10020</value>
<description>MapReduce JobHistory Server IPC host:port</description>
</property>
<property>
<name>mapreduce.jobhistory.webapp.address</name>
<value>master:19888</value>
<description>MapReduceJobHistory Server Web UI host:port</description>
</property>
</configuration>

master节点:

1
$ mr-jobhistory-daemon.sh start historyserver

问题4:spark构建cube点击build后报错

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Caused by: java.lang.NoClassDefFoundError: org/apache/spark/api/java/function/Function
at org.apache.kylin.engine.spark.SparkBatchCubingJobBuilder2.<init>(SparkBatchCubingJobBuilder2.java:53) ~[kylin-engine-spark-2.6.3.jar:2.6.3]
at org.apache.kylin.engine.spark.SparkBatchCubingEngine2.createBatchCubingJob(SparkBatchCubingEngine2.java:44) ~[kylin-engine-spark-2.6.3.jar:2.6.3]
at org.apache.kylin.engine.EngineFactory.createBatchCubingJob(EngineFactory.java:60) ~[kylin-core-job-2.6.3.jar:2.6.3]
at org.apache.kylin.rest.service.JobService.submitJobInternal(JobService.java:234) ~[kylin-server-base-2.6.3.jar:2.6.3]
at org.apache.kylin.rest.service.JobService.submitJob(JobService.java:202) ~[kylin-server-base-2.6.3.jar:2.6.3]
at org.apache.kylin.rest.controller.CubeController.buildInternal(CubeController.java:395) ~[kylin-server-base-2.6.3.jar:2.6.3]
... 77 more
Caused by: java.lang.ClassNotFoundException: org.apache.spark.api.java.function.Function
at org.apache.catalina.loader.WebappClassLoaderBase.loadClass(WebappClassLoaderBase.java:1928) ~[catalina.jar:7.0.91]
at org.apache.catalina.loader.WebappClassLoaderBase.loadClass(WebappClassLoaderBase.java:1771) ~[catalina.jar:7.0.91]
at org.apache.kylin.engine.spark.SparkBatchCubingJobBuilder2.<init>(SparkBatchCubingJobBuilder2.java:53) ~[kylin-engine-spark-2.6.3.jar:2.6.3]
at org.apache.kylin.engine.spark.SparkBatchCubingEngine2.createBatchCubingJob(SparkBatchCubingEngine2.java:44) ~[kylin-engine-spark-2.6.3.jar:2.6.3]
at org.apache.kylin.engine.EngineFactory.createBatchCubingJob(EngineFactory.java:60) ~[kylin-core-job-2.6.3.jar:2.6.3]
at org.apache.kylin.rest.service.JobService.submitJobInternal(JobService.java:234) ~[kylin-server-base-2.6.3.jar:2.6.3]
at org.apache.kylin.rest.service.JobService.submitJob(JobService.java:202) ~[kylin-server-base-2.6.3.jar:2.6.3]
at org.apache.kylin.rest.controller.CubeController.buildInternal(CubeController.java:395) ~[kylin-server-base-2.6.3.jar:2.6.3]

kylin源码的engine-spark模块,pom.xml中有一些provided的依赖,但是在kylin server启动后并没有在CLASSPATH中找到,所以,简单的方法是把找不到的依赖jar包直接拷贝到$KYLIN_HOME/tomcat/lib下面。

kylin-2.5.2/tomcat/lib/spark-core_2.11-2.1.2.jar(Caused by: java.lang.ClassNotFoundException: org.apache.spark.api.java.function.Function)  
kylin-2.5.2/tomcat/lib/scala-library-2.11.8.jar(Caused by: java.lang.ClassNotFoundException: scala.Serializable)  
1
2
$ cp $KYLIN_HOME/spark/jars/spark-core_2.11-2.1.2.jar $KYLIN_HOME/tomcat/lib
$ cp $KYLIN_HOME/spark/jars/scala-library-2.11.8.jar $KYLIN_HOME/tomcat/lib

重启kylin server生效。

问题5:spark构建cube第二步,找不到HiveConf

在$KYLIN_HOME/bin/kylin.sh中配置HBASE_CLASSPATH_PREFIX。

1
2
3
4
org.apache.kylin.job.exception.ExecuteException: org.apache.kylin.job.exception.ExecuteException: java.lang.NoClassDefFoundError: org/apache/hadoop/hive/conf/HiveConf
Caused by: org.apache.kylin.job.exception.ExecuteException: java.lang.NoClassDefFoundError: org/apache/hadoop/hive/conf/HiveConf
Caused by: java.lang.NoClassDefFoundError: org/apache/hadoop/hive/conf/HiveConf
Caused by: java.lang.ClassNotFoundException: org.apache.hadoop.hive.conf.HiveConf

配置HBASE_CLASSPATH_PREFIX,用到了环境变量里的tomcat_root和$hive_dependency。

1
export HBASE_CLASSPATH_PREFIX=${tomcat_root}/bin/bootstrap.jar:${tomcat_root}/bin/tomcat-juli.jar:${tomcat_root}/lib/*:$hive_dependency:$HBASE_CLASSPATH_PREFIX

问题6:spark构建cube第三步,The auxService:spark_shuffle does not exist

日志位置:

/home/admin/hadoop-2.8.5/logs/userlogs/application_1565839225073_0008/container_1565839225073_0008_01_000001
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
19/08/15 16:52:06 ERROR yarn.YarnAllocator: Failed to launch executor 87 on container container_1565839225073_0008_01_000088
org.apache.spark.SparkException: Exception while starting container container_1565839225073_0008_01_000088 on host slave2
at org.apache.spark.deploy.yarn.ExecutorRunnable.startContainer(ExecutorRunnable.scala:126)
at org.apache.spark.deploy.yarn.ExecutorRunnable.run(ExecutorRunnable.scala:66)
at org.apache.spark.deploy.yarn.YarnAllocator$$anonfun$runAllocatedContainers$1$$anon$1.run(YarnAllocator.scala:520)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
Caused by: org.apache.hadoop.yarn.exceptions.InvalidAuxServiceException: The auxService:spark_shuffle does not exist
at sun.reflect.GeneratedConstructorAccessor24.newInstance(Unknown Source)
at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
at java.lang.reflect.Constructor.newInstance(Constructor.java:423)
at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.instantiateException(SerializedExceptionPBImpl.java:168)
at org.apache.hadoop.yarn.api.records.impl.pb.SerializedExceptionPBImpl.deSerialize(SerializedExceptionPBImpl.java:106)
at org.apache.hadoop.yarn.client.api.impl.NMClientImpl.startContainer(NMClientImpl.java:205)
at org.apache.spark.deploy.yarn.ExecutorRunnable.startContainer(ExecutorRunnable.scala:123)
... 5 more
19/08/15 16:52:07 WARN cluster.YarnClusterScheduler: Initial job has not accepted any resources; check your cluster UI to ensure that workers are registered and have sufficient resources
19/08/15 16:52:09 INFO yarn.ApplicationMaster: Final app status: FAILED, exitCode: 11, (reason: Max number of executor failures (80) reached)
19/08/15 16:52:09 ERROR yarn.ApplicationMaster: User class threw exception: java.lang.RuntimeException: error execute org.apache.kylin.engine.spark.SparkFactDistinct. Root cause: null
java.lang.RuntimeException: error execute org.apache.kylin.engine.spark.SparkFactDistinct. Root cause: null
at org.apache.kylin.common.util.AbstractApplication.execute(AbstractApplication.java:42)
at org.apache.kylin.common.util.SparkEntry.main(SparkEntry.java:44)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:498)
at org.apache.spark.deploy.yarn.ApplicationMaster$$anon$2.run(ApplicationMaster.scala:636)
Caused by: java.lang.InterruptedException
at java.util.concurrent.locks.AbstractQueuedSynchronizer.doAcquireSharedInterruptibly(AbstractQueuedSynchronizer.java:998)
at java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireSharedInterruptibly(AbstractQueuedSynchronizer.java:1304)
at scala.concurrent.impl.Promise$DefaultPromise.tryAwait(Promise.scala:202)
at scala.concurrent.impl.Promise$DefaultPromise.ready(Promise.scala:218)
at scala.concurrent.impl.Promise$DefaultPromise.ready(Promise.scala:153)
at org.apache.spark.scheduler.DAGScheduler.runJob(DAGScheduler.scala:619)
at org.apache.spark.SparkContext.runJob(SparkContext.scala:1928)

修改hadoop的yarn-site.xml

1
2
3
4
5
6
7
8
<property>
<name>yarn.nodemanager.aux-services</name>
<value>spark_shuffle,mapreduce_shuffle</value>
</property>
<property>
<name>yarn.nodemanager.aux-services.spark_shuffle.class</name>
<value>org.apache.spark.network.yarn.YarnShuffleService</value>
</property>

需要引入YarnShuffleService所在的jar包,否则Class org.apache.spark.network.yarn.YarnShuffleService not found。

下载spark源码,编译一下。

1
2
3
4
$ wget https://github.com/apache/spark/archive/v2.3.2.tar.gz
$ tar zxvf v2.3.2.tar.gz
$ cd spark-2.3.2/
$ ./build/mvn -Pyarn -Phadoop-2.7 -Dhadoop.version=2.7.3 -DskipTests clean package

需要用到的jar包位置在spark-2.3.2/common/network-yarn/target/scala-2.11目录下,spark-2.3.2-yarn-shuffle.jar。

分发到/home/admin/hadoop-2.8.5/share/hadoop/yarn/lib,重启yarn集群。

以上参考Spark官方文档http://spark.apache.org/docs/2.3.2/running-on-yarn.html

1
2
3
4
5
6
7
8
9
Configuring the External Shuffle Service
To start the Spark Shuffle Service on each NodeManager in your YARN cluster, follow these instructions:

1. Build Spark with the YARN profile. Skip this step if you are using a pre-packaged distribution.
2. Locate the spark-<version>-yarn-shuffle.jar. This should be under $SPARK_HOME/common/network-yarn/target/scala-<version> if you are building Spark yourself, and under yarn if you are using a distribution.
3. Add this jar to the classpath of all NodeManagers in your cluster.
4. In the yarn-site.xml on each node, add spark_shuffle to yarn.nodemanager.aux-services, then set yarn.nodemanager.aux-services.spark_shuffle.class to org.apache.spark.network.yarn.YarnShuffleService.
5. Increase NodeManager's heap size by setting YARN_HEAPSIZE (1000 by default) in etc/hadoop/yarn-env.sh to avoid garbage collection issues during shuffle.
6. Restart all NodeManagers in your cluster.

运行截图【spark2.1.2是kylin2.5.2自带的spark版本,上述遇到的问题kylin2.3.6和kylin2.5.2是一致的】

image

image

image

邵志鹏 wechat
扫一扫上面的二维码关注我的公众号
0%