[120] | 1 | /** |
---|
| 2 | * Licensed to the Apache Software Foundation (ASF) under one |
---|
| 3 | * or more contributor license agreements. See the NOTICE file |
---|
| 4 | * distributed with this work for additional information |
---|
| 5 | * regarding copyright ownership. The ASF licenses this file |
---|
| 6 | * to you under the Apache License, Version 2.0 (the |
---|
| 7 | * "License"); you may not use this file except in compliance |
---|
| 8 | * with the License. You may obtain a copy of the License at |
---|
| 9 | * |
---|
| 10 | * http://www.apache.org/licenses/LICENSE-2.0 |
---|
| 11 | * |
---|
| 12 | * Unless required by applicable law or agreed to in writing, software |
---|
| 13 | * distributed under the License is distributed on an "AS IS" BASIS, |
---|
| 14 | * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
---|
| 15 | * See the License for the specific language governing permissions and |
---|
| 16 | * limitations under the License. |
---|
| 17 | */ |
---|
| 18 | package org.apache.hadoop.hdfs; |
---|
| 19 | |
---|
| 20 | import java.io.IOException; |
---|
| 21 | import java.net.UnknownHostException; |
---|
| 22 | import java.security.NoSuchAlgorithmException; |
---|
| 23 | import java.security.SecureRandom; |
---|
| 24 | import java.util.Random; |
---|
| 25 | |
---|
| 26 | import org.apache.hadoop.conf.Configuration; |
---|
| 27 | import org.apache.hadoop.fs.FileSystem; |
---|
| 28 | import org.apache.hadoop.hdfs.protocol.Block; |
---|
| 29 | import org.apache.hadoop.hdfs.server.common.HdfsConstants.StartupOption; |
---|
| 30 | import org.apache.hadoop.hdfs.server.datanode.SimulatedFSDataset; |
---|
| 31 | import org.apache.hadoop.hdfs.server.namenode.CreateEditsLog; |
---|
| 32 | import org.apache.hadoop.net.DNS; |
---|
| 33 | |
---|
| 34 | |
---|
| 35 | /** |
---|
| 36 | * |
---|
| 37 | |
---|
| 38 | * |
---|
| 39 | * This program starts a mini cluster of data nodes |
---|
| 40 | * (ie a mini cluster without the name node), all within one address space. |
---|
| 41 | * It is assumed that the name node has been started separately prior |
---|
| 42 | * to running this program. |
---|
| 43 | * |
---|
| 44 | * A use case of this is to run a real name node with a large number of |
---|
| 45 | * simulated data nodes for say a NN benchmark. |
---|
| 46 | * |
---|
| 47 | * Synopisis: |
---|
| 48 | * DataNodeCluster -n numDatNodes [-racks numRacks] -simulated |
---|
| 49 | * [-inject startingBlockId numBlocksPerDN] |
---|
| 50 | * [ -r replicationForInjectedBlocks ] |
---|
| 51 | * [-d editsLogDirectory] |
---|
| 52 | * |
---|
| 53 | * if -simulated is specified then simulated data nodes are started. |
---|
| 54 | * if -inject is specified then blocks are injected in each datanode; |
---|
| 55 | * -inject option is valid only for simulated data nodes. |
---|
| 56 | * |
---|
| 57 | * See Also @link #CreateEditsLog for creating a edits log file to |
---|
| 58 | * inject a matching set of blocks into into a name node. |
---|
| 59 | * Typical use of -inject is to inject blocks into a set of datanodes |
---|
| 60 | * using this DataNodeCLuster command |
---|
| 61 | * and then to inject the same blocks into a name node using the |
---|
| 62 | * CreateEditsLog command. |
---|
| 63 | * |
---|
| 64 | */ |
---|
| 65 | |
---|
| 66 | public class DataNodeCluster { |
---|
| 67 | static final String DATANODE_DIRS = "/tmp/DataNodeCluster"; |
---|
| 68 | static String dataNodeDirs = DATANODE_DIRS; |
---|
| 69 | static final String USAGE = |
---|
| 70 | "Usage: datanodecluster " + |
---|
| 71 | " -n <numDataNodes> " + |
---|
| 72 | " [-racks <numRacks>] " + |
---|
| 73 | " [-simulated] " + |
---|
| 74 | " [-inject startingBlockId numBlocksPerDN]" + |
---|
| 75 | " [-r replicationFactorForInjectedBlocks]" + |
---|
| 76 | " [-d dataNodeDirs]\n" + |
---|
| 77 | " Default datanode direcory is " + DATANODE_DIRS + "\n" + |
---|
| 78 | " Default replication factor for injected blocks is 1\n" + |
---|
| 79 | " Defaul rack is used if -racks is not specified\n" + |
---|
| 80 | " Data nodes are simulated if -simulated OR conf file specifies simulated\n"; |
---|
| 81 | |
---|
| 82 | |
---|
| 83 | static void printUsageExit() { |
---|
| 84 | System.out.println(USAGE); |
---|
| 85 | System.exit(-1); |
---|
| 86 | } |
---|
| 87 | static void printUsageExit(String err) { |
---|
| 88 | System.out.println(err); |
---|
| 89 | printUsageExit(); |
---|
| 90 | } |
---|
| 91 | |
---|
| 92 | public static void main(String[] args) { |
---|
| 93 | int numDataNodes = 0; |
---|
| 94 | int numRacks = 0; |
---|
| 95 | boolean inject = false; |
---|
| 96 | long startingBlockId = 1; |
---|
| 97 | int numBlocksPerDNtoInject = 0; |
---|
| 98 | int replication = 1; |
---|
| 99 | |
---|
| 100 | Configuration conf = new Configuration(); |
---|
| 101 | |
---|
| 102 | for (int i = 0; i < args.length; i++) { // parse command line |
---|
| 103 | if (args[i].equals("-n")) { |
---|
| 104 | if (++i >= args.length || args[i].startsWith("-")) { |
---|
| 105 | printUsageExit("missing number of nodes"); |
---|
| 106 | } |
---|
| 107 | numDataNodes = Integer.parseInt(args[i]); |
---|
| 108 | } else if (args[i].equals("-racks")) { |
---|
| 109 | if (++i >= args.length || args[i].startsWith("-")) { |
---|
| 110 | printUsageExit("Missing number of racks"); |
---|
| 111 | } |
---|
| 112 | numRacks = Integer.parseInt(args[i]); |
---|
| 113 | } else if (args[i].equals("-r")) { |
---|
| 114 | if (++i >= args.length || args[i].startsWith("-")) { |
---|
| 115 | printUsageExit("Missing replicaiton factor"); |
---|
| 116 | } |
---|
| 117 | replication = Integer.parseInt(args[i]); |
---|
| 118 | } else if (args[i].equals("-d")) { |
---|
| 119 | if (++i >= args.length || args[i].startsWith("-")) { |
---|
| 120 | printUsageExit("Missing datanode dirs parameter"); |
---|
| 121 | } |
---|
| 122 | dataNodeDirs = args[i]; |
---|
| 123 | } else if (args[i].equals("-simulated")) { |
---|
| 124 | conf.setBoolean(SimulatedFSDataset.CONFIG_PROPERTY_SIMULATED, true); |
---|
| 125 | } else if (args[i].equals("-inject")) { |
---|
| 126 | if (!conf.getBoolean(SimulatedFSDataset.CONFIG_PROPERTY_SIMULATED, |
---|
| 127 | false) ) { |
---|
| 128 | System.out.print("-inject is valid only for simulated"); |
---|
| 129 | printUsageExit(); |
---|
| 130 | } |
---|
| 131 | inject = true; |
---|
| 132 | if (++i >= args.length || args[i].startsWith("-")) { |
---|
| 133 | printUsageExit( |
---|
| 134 | "Missing starting block and number of blocks per DN to inject"); |
---|
| 135 | } |
---|
| 136 | startingBlockId = Integer.parseInt(args[i]); |
---|
| 137 | if (++i >= args.length || args[i].startsWith("-")) { |
---|
| 138 | printUsageExit("Missing number of blocks to inject"); |
---|
| 139 | } |
---|
| 140 | numBlocksPerDNtoInject = Integer.parseInt(args[i]); |
---|
| 141 | } else { |
---|
| 142 | printUsageExit(); |
---|
| 143 | } |
---|
| 144 | } |
---|
| 145 | if (numDataNodes <= 0 || replication <= 0 ) { |
---|
| 146 | printUsageExit("numDataNodes and replication have to be greater than zero"); |
---|
| 147 | } |
---|
| 148 | if (replication > numDataNodes) { |
---|
| 149 | printUsageExit("Replication must be less than or equal to numDataNodes"); |
---|
| 150 | |
---|
| 151 | } |
---|
| 152 | String nameNodeAdr = FileSystem.getDefaultUri(conf).getAuthority(); |
---|
| 153 | if (nameNodeAdr == null) { |
---|
| 154 | System.out.println("No name node address and port in config"); |
---|
| 155 | System.exit(-1); |
---|
| 156 | } |
---|
| 157 | boolean simulated = |
---|
| 158 | conf.getBoolean(SimulatedFSDataset.CONFIG_PROPERTY_SIMULATED, false); |
---|
| 159 | System.out.println("Starting " + numDataNodes + |
---|
| 160 | (simulated ? " Simulated " : " ") + |
---|
| 161 | " Data Nodes that will connect to Name Node at " + nameNodeAdr); |
---|
| 162 | |
---|
| 163 | System.setProperty("test.build.data", dataNodeDirs); |
---|
| 164 | |
---|
| 165 | MiniDFSCluster mc = new MiniDFSCluster(); |
---|
| 166 | try { |
---|
| 167 | mc.formatDataNodeDirs(); |
---|
| 168 | } catch (IOException e) { |
---|
| 169 | System.out.println("Error formating data node dirs:" + e); |
---|
| 170 | } |
---|
| 171 | |
---|
| 172 | String[] rack4DataNode = null; |
---|
| 173 | if (numRacks > 0) { |
---|
| 174 | System.out.println("Using " + numRacks + " racks: "); |
---|
| 175 | String rackPrefix = getUniqueRackPrefix(); |
---|
| 176 | |
---|
| 177 | rack4DataNode = new String[numDataNodes]; |
---|
| 178 | for (int i = 0; i < numDataNodes; ++i ) { |
---|
| 179 | //rack4DataNode[i] = racks[i%numRacks]; |
---|
| 180 | rack4DataNode[i] = rackPrefix + "-" + i%numRacks; |
---|
| 181 | System.out.println("Data Node " + i + " using " + rack4DataNode[i]); |
---|
| 182 | |
---|
| 183 | |
---|
| 184 | } |
---|
| 185 | } |
---|
| 186 | try { |
---|
| 187 | mc.startDataNodes(conf, numDataNodes, true, StartupOption.REGULAR, |
---|
| 188 | rack4DataNode); |
---|
| 189 | if (inject) { |
---|
| 190 | long blockSize = 10; |
---|
| 191 | System.out.println("Injecting " + numBlocksPerDNtoInject + |
---|
| 192 | " blocks in each DN starting at blockId " + startingBlockId + |
---|
| 193 | " with blocksize of " + blockSize); |
---|
| 194 | Block[] blocks = new Block[numBlocksPerDNtoInject]; |
---|
| 195 | long blkid = startingBlockId; |
---|
| 196 | for (int i_dn = 0; i_dn < numDataNodes; ++i_dn) { |
---|
| 197 | for (int i = 0; i < blocks.length; ++i) { |
---|
| 198 | blocks[i] = new Block(blkid++, blockSize, |
---|
| 199 | CreateEditsLog.BLOCK_GENERATION_STAMP); |
---|
| 200 | } |
---|
| 201 | for (int i = 1; i <= replication; ++i) { |
---|
| 202 | // inject blocks for dn_i into dn_i and replica in dn_i's neighbors |
---|
| 203 | mc.injectBlocks((i_dn + i- 1)% numDataNodes, blocks); |
---|
| 204 | System.out.println("Injecting blocks of dn " + i_dn + " into dn" + |
---|
| 205 | ((i_dn + i- 1)% numDataNodes)); |
---|
| 206 | } |
---|
| 207 | } |
---|
| 208 | System.out.println("Created blocks from Bids " |
---|
| 209 | + startingBlockId + " to " + (blkid -1)); |
---|
| 210 | } |
---|
| 211 | |
---|
| 212 | } catch (IOException e) { |
---|
| 213 | System.out.println("Error creating data node:" + e); |
---|
| 214 | } |
---|
| 215 | } |
---|
| 216 | |
---|
| 217 | /* |
---|
| 218 | * There is high probability that the rack id generated here will |
---|
| 219 | * not conflict with those of other data node cluster. |
---|
| 220 | * Not perfect but mostly unique rack ids are good enough |
---|
| 221 | */ |
---|
| 222 | static private String getUniqueRackPrefix() { |
---|
| 223 | |
---|
| 224 | String ip = "unknownIP"; |
---|
| 225 | try { |
---|
| 226 | ip = DNS.getDefaultIP("default"); |
---|
| 227 | } catch (UnknownHostException ignored) { |
---|
| 228 | System.out.println("Could not find ip address of \"default\" inteface."); |
---|
| 229 | } |
---|
| 230 | |
---|
| 231 | int rand = 0; |
---|
| 232 | try { |
---|
| 233 | rand = SecureRandom.getInstance("SHA1PRNG").nextInt(Integer.MAX_VALUE); |
---|
| 234 | } catch (NoSuchAlgorithmException e) { |
---|
| 235 | rand = (new Random()).nextInt(Integer.MAX_VALUE); |
---|
| 236 | } |
---|
| 237 | return "/Rack-" + rand + "-"+ ip + "-" + |
---|
| 238 | System.currentTimeMillis(); |
---|
| 239 | } |
---|
| 240 | } |
---|