Hadoop源码分析三启动及脚本剖析

时间:2021-07-31 03:37:28

1、 启动

hadoop的启动是通过其sbin目录下的脚本来启动的。与启动相关的叫脚本有以下几个:

start-all.shstart-dfs.shstart-yarn.shhadoop-daemon.shyarn-daemon.sh

hadoop-daemon.sh是用来启动与hdfs相关的服务的

yarn-daemon.sh是用来启动和yarn相关的服务

start-dfs.sh是用来启动hdfs集群的

start-yarn.sh是用来启动yarn集群

start-all.sh是用来启动yarn和hdfs集群的

这几个start开头的脚本都是通过调用那两个daemon脚本来启动的。

2、 脚本分析

这里先从start-all.sh开始分析,然后逐步分析其脚本的调用。

start-all.sh脚本内容如下:

?
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
#!/usr/bin/env bash
 
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
 
 
# Start all hadoop daemons.  Run this on master node.
 
echo "This script is Deprecated. Instead use start-dfs.sh and start-yarn.sh"
 
bin=`dirname "${BASH_SOURCE-$0}"`
bin=`cd "$bin"; pwd`
 
DEFAULT_LIBEXEC_DIR="$bin"/../libexec
HADOOP_LIBEXEC_DIR=${HADOOP_LIBEXEC_DIR:-$DEFAULT_LIBEXEC_DIR}
. $HADOOP_LIBEXEC_DIR/hadoop-config.sh
 
# start hdfs daemons if hdfs is present
if [ -f "${HADOOP_HDFS_HOME}"/sbin/start-dfs.sh ]; then
  "${HADOOP_HDFS_HOME}"/sbin/start-dfs.sh --config $HADOOP_CONF_DIR
fi
 
# start yarn daemons if yarn is present
if [ -f "${HADOOP_YARN_HOME}"/sbin/start-yarn.sh ]; then
  "${HADOOP_YARN_HOME}"/sbin/start-yarn.sh --config $HADOOP_CONF_DIR
fi

这个脚本的重点在第31行到末尾,这里是两个if语句,第一个if语句里(第34行)调用的是start-dfs.sh脚本,第二个if语句里(第37行)调用的是start-yarn.sh。

然后我们以start-dfs.sh为例,继续向下分析。

start-dfs.sh的内容如下:

?
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
#!/usr/bin/env bash
 
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
 
 
# Start hadoop dfs daemons.
# Optinally upgrade or rollback dfs state.
# Run this on master node.
 
usage="Usage: start-dfs.sh [-upgrade|-rollback] [other options such as -clusterId]"
 
bin=`dirname "${BASH_SOURCE-$0}"`
bin=`cd "$bin"; pwd`
 
DEFAULT_LIBEXEC_DIR="$bin"/../libexec
HADOOP_LIBEXEC_DIR=${HADOOP_LIBEXEC_DIR:-$DEFAULT_LIBEXEC_DIR}
. $HADOOP_LIBEXEC_DIR/hdfs-config.sh
 
# get arguments
if [[ $# -ge 1 ]]; then
  startOpt="$1"
  shift
  case "$startOpt" in
    -upgrade)
      nameStartOpt="$startOpt"
    ;;
    -rollback)
      dataStartOpt="$startOpt"
    ;;
    *)
      echo $usage
      exit 1
    ;;
  esac
fi
 
#Add other possible options
nameStartOpt="$nameStartOpt $@"
 
#---------------------------------------------------------
# namenodes
 
NAMENODES=$($HADOOP_PREFIX/bin/hdfs getconf -namenodes)
 
echo "Starting namenodes on [$NAMENODES]"
 
"$HADOOP_PREFIX/sbin/hadoop-daemons.sh" \
  --config "$HADOOP_CONF_DIR" \
  --hostnames "$NAMENODES" \
  --script "$bin/hdfs" start namenode $nameStartOpt
 
#---------------------------------------------------------
# datanodes (using default slaves file)
 
if [ -n "$HADOOP_SECURE_DN_USER" ]; then
  echo \
    "Attempting to start secure cluster, skipping datanodes. " \
    "Run start-secure-dns.sh as root to complete startup."
else
  "$HADOOP_PREFIX/sbin/hadoop-daemons.sh" \
    --config "$HADOOP_CONF_DIR" \
    --script "$bin/hdfs" start datanode $dataStartOpt
fi
 
#---------------------------------------------------------
# secondary namenodes (if any)
 
SECONDARY_NAMENODES=$($HADOOP_PREFIX/bin/hdfs getconf -secondarynamenodes 2>/dev/null)
 
if [ -n "$SECONDARY_NAMENODES" ]; then
  echo "Starting secondary namenodes [$SECONDARY_NAMENODES]"
 
  "$HADOOP_PREFIX/sbin/hadoop-daemons.sh" \
      --config "$HADOOP_CONF_DIR" \
      --hostnames "$SECONDARY_NAMENODES" \
      --script "$bin/hdfs" start secondarynamenode
fi
 
#---------------------------------------------------------
# quorumjournal nodes (if any)
 
SHARED_EDITS_DIR=$($HADOOP_PREFIX/bin/hdfs getconf -confKey dfs.namenode.shared.edits.dir 2>&-)
 
case "$SHARED_EDITS_DIR" in
qjournal://*)
  JOURNAL_NODES=$(echo "$SHARED_EDITS_DIR" | sed 's,qjournal://\([^/]*\)/.*,\1,g; s/;/ /g; s/:[0-9]*//g')
  echo "Starting journal nodes [$JOURNAL_NODES]"
  "$HADOOP_PREFIX/sbin/hadoop-daemons.sh" \
      --config "$HADOOP_CONF_DIR" \
      --hostnames "$JOURNAL_NODES" \
      --script "$bin/hdfs" start journalnode ;;
esac
 
#---------------------------------------------------------
# ZK Failover controllers, if auto-HA is enabled
AUTOHA_ENABLED=$($HADOOP_PREFIX/bin/hdfs getconf -confKey dfs.ha.automatic-failover.enabled)
if [ "$(echo "$AUTOHA_ENABLED" | tr A-Z a-z)" = "true" ]; then
  echo "Starting ZK Failover Controllers on NN hosts [$NAMENODES]"
  "$HADOOP_PREFIX/sbin/hadoop-daemons.sh" \
    --config "$HADOOP_CONF_DIR" \
    --hostnames "$NAMENODES" \
    --script "$bin/hdfs" start zkfc
fi
 
# eof

 首先是第23行到30行,这里是在处理hadoop的路径。

然后是第32行到51行是在处理脚本传入的参数。最后就启动hdfs的各个角色了。

首先启动的是namenode,在第60行到63行,这段代码是调用了hadoop-daemons.sh脚本来启动,这个脚本和之前提到的hadoop-daemon.sh脚本的区别在于这个脚本可以在集群的其他机器上启动该启动的角色,而hadoop-daemon.sh只能启动当前机器上的角色。其实hadoop-daemons.sh也是通过调用hadoop-daemon.sh来启动的,这个稍后再分析。这个脚本还有几个参数,其中最重要的是:start namenode。它表示启动namenode。

然后是第65行到第76行,这里是在启动datanode,启动的方式和namenode相同。启动DataNode的代码在第73行。

然后是第78行到第90行,这里在启动secondarynamenode。如果是配置了namenode的高可用,secondarynamenode便不会启动。

然后是第92行到第105行,这里在启动journalnode。如果配置了namenode的高可用,journalnode才会启动。

最后是第107行到末尾,这里启动的是zkfc。同样这个也是要配置高可用才会启动。

如果按照文档(2)中配置的高可用来看,这里启动的角色应该为:namenode、datanode、journalnode、zkfc。

启动上述角色调用的hadoop-daemons.sh脚本内容如下:

?
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
#!/usr/bin/env bash
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# Run a Hadoop command on all slave hosts.
usage="Usage: hadoop-daemons.sh [--config confdir] [--hosts hostlistfile] [start|stop] command args..."
# if no args specified, show usage
if [ $# -le 1 ]; then
  echo $usage
  exit 1
fi
bin=`dirname "${BASH_SOURCE-$0}"`
bin=`cd "$bin"; pwd`
DEFAULT_LIBEXEC_DIR="$bin"/../libexec
HADOOP_LIBEXEC_DIR=${HADOOP_LIBEXEC_DIR:-$DEFAULT_LIBEXEC_DIR}
. $HADOOP_LIBEXEC_DIR/hadoop-config.sh
exec "$bin/slaves.sh" --config $HADOOP_CONF_DIR cd "$HADOOP_PREFIX" \; "$bin/hadoop-daemon.sh" --config $HADOOP_CONF_DIR "$@"

这个脚本的重点在最后一行。这里有两个脚本slaves.sh和hadoop-daemon.sh。slaves.sh脚本会使用ssh登录到指定的服务器中然后执行其中的hadoop-daemon.sh脚本。这里就不分析slaves.sh脚本了。

我们继续看hadoop-daemon.sh脚本。

其内容如下:

?
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
#!/usr/bin/env bash
 
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
 
 
# Runs a Hadoop command as a daemon.
#
# Environment Variables
#
#   HADOOP_CONF_DIR  Alternate conf dir. Default is ${HADOOP_PREFIX}/conf.
#   HADOOP_LOG_DIR   Where log files are stored.  PWD by default.
#   HADOOP_MASTER    host:path where hadoop code should be rsync'd from
#   HADOOP_PID_DIR   The pid files are stored. /tmp by default.
#   HADOOP_IDENT_STRING   A string representing this instance of hadoop. $USER by default
#   HADOOP_NICENESS The scheduling priority for daemons. Defaults to 0.
##
 
usage="Usage: hadoop-daemon.sh [--config <conf-dir>] [--hosts hostlistfile] [--script script] (start|stop) <hadoop-command> <args...>"
 
# if no args specified, show usage
if [ $# -le 1 ]; then
  echo $usage
  exit 1
fi
 
bin=`dirname "${BASH_SOURCE-$0}"`
bin=`cd "$bin"; pwd`
 
DEFAULT_LIBEXEC_DIR="$bin"/../libexec
HADOOP_LIBEXEC_DIR=${HADOOP_LIBEXEC_DIR:-$DEFAULT_LIBEXEC_DIR}
. $HADOOP_LIBEXEC_DIR/hadoop-config.sh
 
# get arguments
 
#default value
hadoopScript="$HADOOP_PREFIX"/bin/hadoop
if [ "--script" = "$1" ]
  then
    shift
    hadoopScript=$1
    shift
fi
startStop=$1
shift
command=$1
shift
 
hadoop_rotate_log ()
{
    log=$1;
    num=5;
    if [ -n "$2" ]; then
    num=$2
    fi
    if [ -f "$log" ]; then # rotate logs
    while [ $num -gt 1 ]; do
        prev=`expr $num - 1`
        [ -f "$log.$prev" ] && mv "$log.$prev" "$log.$num"
        num=$prev
    done
    mv "$log" "$log.$num";
    fi
}
 
if [ -f "${HADOOP_CONF_DIR}/hadoop-env.sh" ]; then
  . "${HADOOP_CONF_DIR}/hadoop-env.sh"
fi
 
# Determine if we're starting a secure datanode, and if so, redefine appropriate variables
if [ "$command" == "datanode" ] && [ "$EUID" -eq 0 ] && [ -n "$HADOOP_SECURE_DN_USER" ]; then
  export HADOOP_PID_DIR=$HADOOP_SECURE_DN_PID_DIR
  export HADOOP_LOG_DIR=$HADOOP_SECURE_DN_LOG_DIR
  export HADOOP_IDENT_STRING=$HADOOP_SECURE_DN_USER
  starting_secure_dn="true"
fi
 
#Determine if we're starting a privileged NFS, if so, redefine the appropriate variables
if [ "$command" == "nfs3" ] && [ "$EUID" -eq 0 ] && [ -n "$HADOOP_PRIVILEGED_NFS_USER" ]; then
    export HADOOP_PID_DIR=$HADOOP_PRIVILEGED_NFS_PID_DIR
    export HADOOP_LOG_DIR=$HADOOP_PRIVILEGED_NFS_LOG_DIR
    export HADOOP_IDENT_STRING=$HADOOP_PRIVILEGED_NFS_USER
    starting_privileged_nfs="true"
fi
 
if [ "$HADOOP_IDENT_STRING" = "" ]; then
  export HADOOP_IDENT_STRING="$USER"
fi
 
 
# get log directory
if [ "$HADOOP_LOG_DIR" = "" ]; then
  export HADOOP_LOG_DIR="$HADOOP_PREFIX/logs"
fi
 
if [ ! -w "$HADOOP_LOG_DIR" ] ; then
  mkdir -p "$HADOOP_LOG_DIR"
  chown $HADOOP_IDENT_STRING $HADOOP_LOG_DIR
fi
 
if [ "$HADOOP_PID_DIR" = "" ]; then
  HADOOP_PID_DIR=/tmp
fi
 
# some variables
export HADOOP_LOGFILE=hadoop-$HADOOP_IDENT_STRING-$command-$HOSTNAME.log
export HADOOP_ROOT_LOGGER=${HADOOP_ROOT_LOGGER:-"INFO,RFA"}
export HADOOP_SECURITY_LOGGER=${HADOOP_SECURITY_LOGGER:-"INFO,RFAS"}
export HDFS_AUDIT_LOGGER=${HDFS_AUDIT_LOGGER:-"INFO,NullAppender"}
log=$HADOOP_LOG_DIR/hadoop-$HADOOP_IDENT_STRING-$command-$HOSTNAME.out
pid=$HADOOP_PID_DIR/hadoop-$HADOOP_IDENT_STRING-$command.pid
HADOOP_STOP_TIMEOUT=${HADOOP_STOP_TIMEOUT:-5}
 
# Set default scheduling priority
if [ "$HADOOP_NICENESS" = "" ]; then
    export HADOOP_NICENESS=0
fi
 
case $startStop in
 
  (start)
 
    [ -w "$HADOOP_PID_DIR" ] ||  mkdir -p "$HADOOP_PID_DIR"
 
    if [ -f $pid ]; then
      if kill -0 `cat $pid` > /dev/null 2>&1; then
        echo $command running as process `cat $pid`.  Stop it first.
        exit 1
      fi
    fi
 
    if [ "$HADOOP_MASTER" != "" ]; then
      echo rsync from $HADOOP_MASTER
      rsync -a -e ssh --delete --exclude=.svn --exclude='logs/*' --exclude='contrib/hod/logs/*' $HADOOP_MASTER/ "$HADOOP_PREFIX"
    fi
 
    hadoop_rotate_log $log
    echo starting $command, logging to $log
    cd "$HADOOP_PREFIX"
    case $command in
      namenode|secondarynamenode|datanode|journalnode|dfs|dfsadmin|fsck|balancer|zkfc)
        if [ -z "$HADOOP_HDFS_HOME" ]; then
          hdfsScript="$HADOOP_PREFIX"/bin/hdfs
        else
          hdfsScript="$HADOOP_HDFS_HOME"/bin/hdfs
        fi
        nohup nice -n $HADOOP_NICENESS $hdfsScript --config $HADOOP_CONF_DIR $command "$@" > "$log" 2>&1 < /dev/null &
      ;;
      (*)
        nohup nice -n $HADOOP_NICENESS $hadoopScript --config $HADOOP_CONF_DIR $command "$@" > "$log" 2>&1 < /dev/null &
      ;;
    esac
    echo $! > $pid
    sleep 1
    head "$log"
    # capture the ulimit output
    if [ "true" = "$starting_secure_dn" ]; then
      echo "ulimit -a for secure datanode user $HADOOP_SECURE_DN_USER" >> $log
      # capture the ulimit info for the appropriate user
      su --shell=/bin/bash $HADOOP_SECURE_DN_USER -c 'ulimit -a' >> $log 2>&1
    elif [ "true" = "$starting_privileged_nfs" ]; then
        echo "ulimit -a for privileged nfs user $HADOOP_PRIVILEGED_NFS_USER" >> $log
        su --shell=/bin/bash $HADOOP_PRIVILEGED_NFS_USER -c 'ulimit -a' >> $log 2>&1
    else
      echo "ulimit -a for user $USER" >> $log
      ulimit -a >> $log 2>&1
    fi
    sleep 3;
    if ! ps -p $! > /dev/null ; then
      exit 1
    fi
    ;;
 
  (stop)
 
    if [ -f $pid ]; then
      TARGET_PID=`cat $pid`
      if kill -0 $TARGET_PID > /dev/null 2>&1; then
        echo stopping $command
        kill $TARGET_PID
        sleep $HADOOP_STOP_TIMEOUT
        if kill -0 $TARGET_PID > /dev/null 2>&1; then
          echo "$command did not stop gracefully after $HADOOP_STOP_TIMEOUT seconds: killing with kill -9"
          kill -9 $TARGET_PID
        fi
      else
        echo no $command to stop
      fi
      rm -f $pid
    else
      echo no $command to stop
    fi
    ;;
 
  (*)
    echo $usage
    exit 1
    ;;
 
esac

这段代码的重点在第131行到结束。这里是真正在启动服务的代码,这个文件在调用的时候,会传入两个重要的参数start/stop xxx。用于启动或停止某些服务。以启动服务为例,其重点在第153行,这里会执行一个hdfsScript脚本。这个参数的定义在第155行,

这里可以看见它实际是hadoop的bin目录下的hdfs文件

文件的内容如下:

?
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
#!/usr/bin/env bash
 
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
 
# Environment Variables
#
#   JSVC_HOME  home directory of jsvc binary.  Required for starting secure
#              datanode.
#
#   JSVC_OUTFILE  path to jsvc output file.  Defaults to
#                 $HADOOP_LOG_DIR/jsvc.out.
#
#   JSVC_ERRFILE  path to jsvc error file.  Defaults to $HADOOP_LOG_DIR/jsvc.err.
 
bin=`which $0`
bin=`dirname ${bin}`
bin=`cd "$bin" > /dev/null; pwd`
 
DEFAULT_LIBEXEC_DIR="$bin"/../libexec
 
HADOOP_LIBEXEC_DIR=${HADOOP_LIBEXEC_DIR:-$DEFAULT_LIBEXEC_DIR}
. $HADOOP_LIBEXEC_DIR/hdfs-config.sh
 
function print_usage(){
  echo "Usage: hdfs [--config confdir] [--loglevel loglevel] COMMAND"
  echo "       where COMMAND is one of:"
  echo "  dfs                  run a filesystem command on the file systems supported in Hadoop."
  echo "  classpath            prints the classpath"
  echo "  namenode -format     format the DFS filesystem"
  echo "  secondarynamenode    run the DFS secondary namenode"
  echo "  namenode             run the DFS namenode"
  echo "  journalnode          run the DFS journalnode"
  echo "  zkfc                 run the ZK Failover Controller daemon"
  echo "  datanode             run a DFS datanode"
  echo "  dfsadmin             run a DFS admin client"
  echo "  haadmin              run a DFS HA admin client"
  echo "  fsck                 run a DFS filesystem checking utility"
  echo "  balancer             run a cluster balancing utility"
  echo "  jmxget               get JMX exported values from NameNode or DataNode."
  echo "  mover                run a utility to move block replicas across"
  echo "                       storage types"
  echo "  oiv                  apply the offline fsimage viewer to an fsimage"
  echo "  oiv_legacy           apply the offline fsimage viewer to an legacy fsimage"
  echo "  oev                  apply the offline edits viewer to an edits file"
  echo "  fetchdt              fetch a delegation token from the NameNode"
  echo "  getconf              get config values from configuration"
  echo "  groups               get the groups which users belong to"
  echo "  snapshotDiff         diff two snapshots of a directory or diff the"
  echo "                       current directory contents with a snapshot"
  echo "  lsSnapshottableDir   list all snapshottable dirs owned by the current user"
  echo "                        Use -help to see options"
  echo "  portmap              run a portmap service"
  echo "  nfs3                 run an NFS version 3 gateway"
  echo "  cacheadmin           configure the HDFS cache"
  echo "  crypto               configure HDFS encryption zones"
  echo "  storagepolicies      list/get/set block storage policies"
  echo "  version              print the version"
  echo ""
  echo "Most commands print help when invoked w/o parameters."
  # There are also debug commands, but they don't show up in this listing.
}
 
if [ $# = 0 ]; then
  print_usage
  exit
fi
 
COMMAND=$1
shift
 
case $COMMAND in
  # usage flags
  --help|-help|-h)
    print_usage
    exit
    ;;
esac
 
# Determine if we're starting a secure datanode, and if so, redefine appropriate variables
if [ "$COMMAND" == "datanode" ] && [ "$EUID" -eq 0 ] && [ -n "$HADOOP_SECURE_DN_USER" ]; then
  if [ -n "$JSVC_HOME" ]; then
    if [ -n "$HADOOP_SECURE_DN_PID_DIR" ]; then
      HADOOP_PID_DIR=$HADOOP_SECURE_DN_PID_DIR
    fi
 
    if [ -n "$HADOOP_SECURE_DN_LOG_DIR" ]; then
      HADOOP_LOG_DIR=$HADOOP_SECURE_DN_LOG_DIR
      HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.log.dir=$HADOOP_LOG_DIR"
    fi
 
    HADOOP_IDENT_STRING=$HADOOP_SECURE_DN_USER
    HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.id.str=$HADOOP_IDENT_STRING"
    starting_secure_dn="true"
  else
    echo "It looks like you're trying to start a secure DN, but \$JSVC_HOME"\
      "isn't set. Falling back to starting insecure DN."
  fi
fi
 
# Determine if we're starting a privileged NFS daemon, and if so, redefine appropriate variables
if [ "$COMMAND" == "nfs3" ] && [ "$EUID" -eq 0 ] && [ -n "$HADOOP_PRIVILEGED_NFS_USER" ]; then
  if [ -n "$JSVC_HOME" ]; then
    if [ -n "$HADOOP_PRIVILEGED_NFS_PID_DIR" ]; then
      HADOOP_PID_DIR=$HADOOP_PRIVILEGED_NFS_PID_DIR
    fi
 
    if [ -n "$HADOOP_PRIVILEGED_NFS_LOG_DIR" ]; then
      HADOOP_LOG_DIR=$HADOOP_PRIVILEGED_NFS_LOG_DIR
      HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.log.dir=$HADOOP_LOG_DIR"
    fi
 
    HADOOP_IDENT_STRING=$HADOOP_PRIVILEGED_NFS_USER
    HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.id.str=$HADOOP_IDENT_STRING"
    starting_privileged_nfs="true"
  else
    echo "It looks like you're trying to start a privileged NFS server, but"\
      "\$JSVC_HOME isn't set. Falling back to starting unprivileged NFS server."
  fi
fi
 
if [ "$COMMAND" = "namenode" ] ; then
  CLASS='org.apache.hadoop.hdfs.server.namenode.NameNode'
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_NAMENODE_OPTS"
elif [ "$COMMAND" = "zkfc" ] ; then
  CLASS='org.apache.hadoop.hdfs.tools.DFSZKFailoverController'
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_ZKFC_OPTS"
elif [ "$COMMAND" = "secondarynamenode" ] ; then
  CLASS='org.apache.hadoop.hdfs.server.namenode.SecondaryNameNode'
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_SECONDARYNAMENODE_OPTS"
elif [ "$COMMAND" = "datanode" ] ; then
  CLASS='org.apache.hadoop.hdfs.server.datanode.DataNode'
  if [ "$starting_secure_dn" = "true" ]; then
    HADOOP_OPTS="$HADOOP_OPTS -jvm server $HADOOP_DATANODE_OPTS"
  else
    HADOOP_OPTS="$HADOOP_OPTS -server $HADOOP_DATANODE_OPTS"
  fi
elif [ "$COMMAND" = "journalnode" ] ; then
  CLASS='org.apache.hadoop.hdfs.qjournal.server.JournalNode'
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_JOURNALNODE_OPTS"
elif [ "$COMMAND" = "dfs" ] ; then
  CLASS=org.apache.hadoop.fs.FsShell
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_CLIENT_OPTS"
elif [ "$COMMAND" = "dfsadmin" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.DFSAdmin
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_CLIENT_OPTS"
elif [ "$COMMAND" = "haadmin" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.DFSHAAdmin
  CLASSPATH=${CLASSPATH}:${TOOL_PATH}
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_CLIENT_OPTS"
elif [ "$COMMAND" = "fsck" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.DFSck
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_CLIENT_OPTS"
elif [ "$COMMAND" = "balancer" ] ; then
  CLASS=org.apache.hadoop.hdfs.server.balancer.Balancer
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_BALANCER_OPTS"
elif [ "$COMMAND" = "mover" ] ; then
  CLASS=org.apache.hadoop.hdfs.server.mover.Mover
  HADOOP_OPTS="${HADOOP_OPTS} ${HADOOP_MOVER_OPTS}"
elif [ "$COMMAND" = "storagepolicies" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.StoragePolicyAdmin
elif [ "$COMMAND" = "jmxget" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.JMXGet
elif [ "$COMMAND" = "oiv" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.offlineImageViewer.OfflineImageViewerPB
elif [ "$COMMAND" = "oiv_legacy" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.offlineImageViewer.OfflineImageViewer
elif [ "$COMMAND" = "oev" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.offlineEditsViewer.OfflineEditsViewer
elif [ "$COMMAND" = "fetchdt" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.DelegationTokenFetcher
elif [ "$COMMAND" = "getconf" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.GetConf
elif [ "$COMMAND" = "groups" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.GetGroups
elif [ "$COMMAND" = "snapshotDiff" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.snapshot.SnapshotDiff
elif [ "$COMMAND" = "lsSnapshottableDir" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.snapshot.LsSnapshottableDir
elif [ "$COMMAND" = "portmap" ] ; then
  CLASS=org.apache.hadoop.portmap.Portmap
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_PORTMAP_OPTS"
elif [ "$COMMAND" = "nfs3" ] ; then
  CLASS=org.apache.hadoop.hdfs.nfs.nfs3.Nfs3
  HADOOP_OPTS="$HADOOP_OPTS $HADOOP_NFS3_OPTS"
elif [ "$COMMAND" = "cacheadmin" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.CacheAdmin
elif [ "$COMMAND" = "crypto" ] ; then
  CLASS=org.apache.hadoop.hdfs.tools.CryptoAdmin
elif [ "$COMMAND" = "version" ] ; then
  CLASS=org.apache.hadoop.util.VersionInfo
elif [ "$COMMAND" = "debug" ]; then
  CLASS=org.apache.hadoop.hdfs.tools.DebugAdmin
elif [ "$COMMAND" = "classpath" ]; then
  if [ "$#" -gt 0 ]; then
    CLASS=org.apache.hadoop.util.Classpath
  else
    # No need to bother starting up a JVM for this simple case.
    if $cygwin; then
      CLASSPATH=$(cygpath -p -w "$CLASSPATH" 2>/dev/null)
    fi
    echo $CLASSPATH
    exit 0
  fi
else
  CLASS="$COMMAND"
fi
 
# cygwin path translation
if $cygwin; then
  CLASSPATH=$(cygpath -p -w "$CLASSPATH" 2>/dev/null)
  HADOOP_LOG_DIR=$(cygpath -w "$HADOOP_LOG_DIR" 2>/dev/null)
  HADOOP_PREFIX=$(cygpath -w "$HADOOP_PREFIX" 2>/dev/null)
  HADOOP_CONF_DIR=$(cygpath -w "$HADOOP_CONF_DIR" 2>/dev/null)
  HADOOP_COMMON_HOME=$(cygpath -w "$HADOOP_COMMON_HOME" 2>/dev/null)
  HADOOP_HDFS_HOME=$(cygpath -w "$HADOOP_HDFS_HOME" 2>/dev/null)
  HADOOP_YARN_HOME=$(cygpath -w "$HADOOP_YARN_HOME" 2>/dev/null)
  HADOOP_MAPRED_HOME=$(cygpath -w "$HADOOP_MAPRED_HOME" 2>/dev/null)
fi
 
export CLASSPATH=$CLASSPATH
 
HADOOP_OPTS="$HADOOP_OPTS -Dhadoop.security.logger=${HADOOP_SECURITY_LOGGER:-INFO,NullAppender}"
 
# Check to see if we should start a secure datanode
if [ "$starting_secure_dn" = "true" ]; then
  if [ "$HADOOP_PID_DIR" = "" ]; then
    HADOOP_SECURE_DN_PID="/tmp/hadoop_secure_dn.pid"
  else
    HADOOP_SECURE_DN_PID="$HADOOP_PID_DIR/hadoop_secure_dn.pid"
  fi
 
  JSVC=$JSVC_HOME/jsvc
  if [ ! -f $JSVC ]; then
    echo "JSVC_HOME is not set correctly so jsvc cannot be found. jsvc is required to run secure datanodes. "
    echo "Please download and install jsvc from http://archive.apache.org/dist/commons/daemon/binaries/ "\
      "and set JSVC_HOME to the directory containing the jsvc binary."
    exit
  fi
 
  if [[ ! $JSVC_OUTFILE ]]; then
    JSVC_OUTFILE="$HADOOP_LOG_DIR/jsvc.out"
  fi
 
  if [[ ! $JSVC_ERRFILE ]]; then
    JSVC_ERRFILE="$HADOOP_LOG_DIR/jsvc.err"
  fi
 
  exec "$JSVC" \
           -Dproc_$COMMAND -outfile "$JSVC_OUTFILE" \
           -errfile "$JSVC_ERRFILE" \
           -pidfile "$HADOOP_SECURE_DN_PID" \
           -nodetach \
           -user "$HADOOP_SECURE_DN_USER" \
            -cp "$CLASSPATH" \
           $JAVA_HEAP_MAX $HADOOP_OPTS \
           org.apache.hadoop.hdfs.server.datanode.SecureDataNodeStarter "$@"
elif [ "$starting_privileged_nfs" = "true" ] ; then
  if [ "$HADOOP_PID_DIR" = "" ]; then
    HADOOP_PRIVILEGED_NFS_PID="/tmp/hadoop_privileged_nfs3.pid"
  else
    HADOOP_PRIVILEGED_NFS_PID="$HADOOP_PID_DIR/hadoop_privileged_nfs3.pid"
  fi
 
  JSVC=$JSVC_HOME/jsvc
  if [ ! -f $JSVC ]; then
    echo "JSVC_HOME is not set correctly so jsvc cannot be found. jsvc is required to run privileged NFS gateways. "
    echo "Please download and install jsvc from http://archive.apache.org/dist/commons/daemon/binaries/ "\
      "and set JSVC_HOME to the directory containing the jsvc binary."
    exit
  fi
 
  if [[ ! $JSVC_OUTFILE ]]; then
    JSVC_OUTFILE="$HADOOP_LOG_DIR/nfs3_jsvc.out"
  fi
 
  if [[ ! $JSVC_ERRFILE ]]; then
    JSVC_ERRFILE="$HADOOP_LOG_DIR/nfs3_jsvc.err"
  fi
 
  exec "$JSVC" \
           -Dproc_$COMMAND -outfile "$JSVC_OUTFILE" \
           -errfile "$JSVC_ERRFILE" \
           -pidfile "$HADOOP_PRIVILEGED_NFS_PID" \
           -nodetach \
           -user "$HADOOP_PRIVILEGED_NFS_USER" \
           -cp "$CLASSPATH" \
           $JAVA_HEAP_MAX $HADOOP_OPTS \
           org.apache.hadoop.hdfs.nfs.nfs3.PrivilegedNfsGatewayStarter "$@"
else
  # run it
  exec "$JAVA" -Dproc_$COMMAND $JAVA_HEAP_MAX $HADOOP_OPTS $CLASS "$@"
fi

这段代码的重点在第134行到219行和第304行。首先看第134行到219行,这段代码虽然很长但全是一堆if else语句。其逻辑也很简单,就是根据传入的COMMAND的值来为CLASS和HADOOP_OPTS赋值。然后是第304行,执行CLASS类。以第134行的namenode类为例,这里CLASS的值为org.apache.hadoop.hdfs.server.namenode.NameNode,这是一个java类随后便会执行这个类启动namenode。

如果想要debug hdfs的源码,最好在这里设置远程调试。因为这里有单个的服务类与启动参数,可以准确的定位需要的服务。

以上就是Hadoop源码分析之启动及脚本剖析的详细内容,本系列下一篇文章传送门Hadoop源码分析四远程debug调试更多关于Hadoop源码分析分析的资料请继续关注服务器之家其它相关文章!

原文链接:https://blog.csdn.net/qq_39210987/article/details/113922124