hdfs-default.xml 33 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974975976977978979980981982983984985986987988989990991992993994995996997998999100010011002100310041005100610071008100910101011101210131014101510161017
  1. <?xml version="1.0"?>
  2. <?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
  3. <!--
  4. Licensed to the Apache Software Foundation (ASF) under one or more
  5. contributor license agreements. See the NOTICE file distributed with
  6. this work for additional information regarding copyright ownership.
  7. The ASF licenses this file to You under the Apache License, Version 2.0
  8. (the "License"); you may not use this file except in compliance with
  9. the License. You may obtain a copy of the License at
  10. http://www.apache.org/licenses/LICENSE-2.0
  11. Unless required by applicable law or agreed to in writing, software
  12. distributed under the License is distributed on an "AS IS" BASIS,
  13. WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  14. See the License for the specific language governing permissions and
  15. limitations under the License.
  16. -->
  17. <!-- Do not modify this file directly. Instead, copy entries that you -->
  18. <!-- wish to modify from this file into hdfs-site.xml and change them -->
  19. <!-- there. If hdfs-site.xml does not already exist, create it. -->
  20. <configuration>
  21. <property>
  22. <name>hadoop.hdfs.configuration.version</name>
  23. <value>1</value>
  24. <description>version of this configuration file</description>
  25. </property>
  26. <property>
  27. <name>dfs.namenode.logging.level</name>
  28. <value>info</value>
  29. <description>
  30. The logging level for dfs namenode. Other values are "dir" (trace
  31. namespace mutations), "block" (trace block under/over replications
  32. and block creations/deletions), or "all".
  33. </description>
  34. </property>
  35. <property>
  36. <name>dfs.namenode.secondary.http-address</name>
  37. <value>0.0.0.0:50090</value>
  38. <description>
  39. The secondary namenode http server address and port.
  40. If the port is 0 then the server will start on a free port.
  41. </description>
  42. </property>
  43. <property>
  44. <name>dfs.datanode.address</name>
  45. <value>0.0.0.0:50010</value>
  46. <description>
  47. The address where the datanode server will listen to.
  48. If the port is 0 then the server will start on a free port.
  49. </description>
  50. </property>
  51. <property>
  52. <name>dfs.datanode.http.address</name>
  53. <value>0.0.0.0:50075</value>
  54. <description>
  55. The datanode http server address and port.
  56. If the port is 0 then the server will start on a free port.
  57. </description>
  58. </property>
  59. <property>
  60. <name>dfs.datanode.ipc.address</name>
  61. <value>0.0.0.0:50020</value>
  62. <description>
  63. The datanode ipc server address and port.
  64. If the port is 0 then the server will start on a free port.
  65. </description>
  66. </property>
  67. <property>
  68. <name>dfs.datanode.handler.count</name>
  69. <value>3</value>
  70. <description>The number of server threads for the datanode.</description>
  71. </property>
  72. <property>
  73. <name>dfs.namenode.http-address</name>
  74. <value>0.0.0.0:50070</value>
  75. <description>
  76. The address and the base port where the dfs namenode web ui will listen on.
  77. If the port is 0 then the server will start on a free port.
  78. </description>
  79. </property>
  80. <property>
  81. <name>dfs.https.enable</name>
  82. <value>false</value>
  83. <description>Decide if HTTPS(SSL) is supported on HDFS
  84. </description>
  85. </property>
  86. <property>
  87. <name>dfs.client.https.need-auth</name>
  88. <value>false</value>
  89. <description>Whether SSL client certificate authentication is required
  90. </description>
  91. </property>
  92. <property>
  93. <name>dfs.https.server.keystore.resource</name>
  94. <value>ssl-server.xml</value>
  95. <description>Resource file from which ssl server keystore
  96. information will be extracted
  97. </description>
  98. </property>
  99. <property>
  100. <name>dfs.client.https.keystore.resource</name>
  101. <value>ssl-client.xml</value>
  102. <description>Resource file from which ssl client keystore
  103. information will be extracted
  104. </description>
  105. </property>
  106. <property>
  107. <name>dfs.datanode.https.address</name>
  108. <value>0.0.0.0:50475</value>
  109. <description>The datanode secure http server address and port.</description>
  110. </property>
  111. <property>
  112. <name>dfs.namenode.https-address</name>
  113. <value>0.0.0.0:50470</value>
  114. <description>The namenode secure http server address and port.</description>
  115. </property>
  116. <property>
  117. <name>dfs.datanode.dns.interface</name>
  118. <value>default</value>
  119. <description>The name of the Network Interface from which a data node should
  120. report its IP address.
  121. </description>
  122. </property>
  123. <property>
  124. <name>dfs.datanode.dns.nameserver</name>
  125. <value>default</value>
  126. <description>The host name or IP address of the name server (DNS)
  127. which a DataNode should use to determine the host name used by the
  128. NameNode for communication and display purposes.
  129. </description>
  130. </property>
  131. <property>
  132. <name>dfs.namenode.backup.address</name>
  133. <value>0.0.0.0:50100</value>
  134. <description>
  135. The backup node server address and port.
  136. If the port is 0 then the server will start on a free port.
  137. </description>
  138. </property>
  139. <property>
  140. <name>dfs.namenode.backup.http-address</name>
  141. <value>0.0.0.0:50105</value>
  142. <description>
  143. The backup node http server address and port.
  144. If the port is 0 then the server will start on a free port.
  145. </description>
  146. </property>
  147. <property>
  148. <name>dfs.namenode.replication.considerLoad</name>
  149. <value>true</value>
  150. <description>Decide if chooseTarget considers the target's load or not
  151. </description>
  152. </property>
  153. <property>
  154. <name>dfs.default.chunk.view.size</name>
  155. <value>32768</value>
  156. <description>The number of bytes to view for a file on the browser.
  157. </description>
  158. </property>
  159. <property>
  160. <name>dfs.datanode.du.reserved</name>
  161. <value>0</value>
  162. <description>Reserved space in bytes per volume. Always leave this much space free for non dfs use.
  163. </description>
  164. </property>
  165. <property>
  166. <name>dfs.namenode.name.dir</name>
  167. <value>file://${hadoop.tmp.dir}/dfs/name</value>
  168. <description>Determines where on the local filesystem the DFS name node
  169. should store the name table(fsimage). If this is a comma-delimited list
  170. of directories then the name table is replicated in all of the
  171. directories, for redundancy. </description>
  172. </property>
  173. <property>
  174. <name>dfs.namenode.name.dir.restore</name>
  175. <value>false</value>
  176. <description>Set to true to enable NameNode to attempt recovering a
  177. previously failed dfs.namenode.name.dir. When enabled, a recovery of any
  178. failed directory is attempted during checkpoint.</description>
  179. </property>
  180. <property>
  181. <name>dfs.namenode.fs-limits.max-component-length</name>
  182. <value>0</value>
  183. <description>Defines the maximum number of characters in each component
  184. of a path. A value of 0 will disable the check.</description>
  185. </property>
  186. <property>
  187. <name>dfs.namenode.fs-limits.max-directory-items</name>
  188. <value>0</value>
  189. <description>Defines the maximum number of items that a directory may
  190. contain. A value of 0 will disable the check.</description>
  191. </property>
  192. <property>
  193. <name>dfs.namenode.edits.dir</name>
  194. <value>${dfs.namenode.name.dir}</value>
  195. <description>Determines where on the local filesystem the DFS name node
  196. should store the transaction (edits) file. If this is a comma-delimited list
  197. of directories then the transaction file is replicated in all of the
  198. directories, for redundancy. Default value is same as dfs.namenode.name.dir
  199. </description>
  200. </property>
  201. <property>
  202. <name>dfs.namenode.shared.edits.dir</name>
  203. <value></value>
  204. <description>A directory on shared storage between the multiple namenodes
  205. in an HA cluster. This directory will be written by the active and read
  206. by the standby in order to keep the namespaces synchronized. This directory
  207. does not need to be listed in dfs.namenode.edits.dir above. It should be
  208. left empty in a non-HA cluster.
  209. </description>
  210. </property>
  211. <property>
  212. <name>dfs.permissions.enabled</name>
  213. <value>true</value>
  214. <description>
  215. If "true", enable permission checking in HDFS.
  216. If "false", permission checking is turned off,
  217. but all other behavior is unchanged.
  218. Switching from one parameter value to the other does not change the mode,
  219. owner or group of files or directories.
  220. </description>
  221. </property>
  222. <property>
  223. <name>dfs.permissions.superusergroup</name>
  224. <value>supergroup</value>
  225. <description>The name of the group of super-users.</description>
  226. </property>
  227. <!--
  228. <property>
  229. <name>dfs.cluster.administrators</name>
  230. <value>ACL for the admins</value>
  231. <description>This configuration is used to control who can access the
  232. default servlets in the namenode, etc.
  233. </description>
  234. </property>
  235. -->
  236. <property>
  237. <name>dfs.block.access.token.enable</name>
  238. <value>false</value>
  239. <description>
  240. If "true", access tokens are used as capabilities for accessing datanodes.
  241. If "false", no access tokens are checked on accessing datanodes.
  242. </description>
  243. </property>
  244. <property>
  245. <name>dfs.block.access.key.update.interval</name>
  246. <value>600</value>
  247. <description>
  248. Interval in minutes at which namenode updates its access keys.
  249. </description>
  250. </property>
  251. <property>
  252. <name>dfs.block.access.token.lifetime</name>
  253. <value>600</value>
  254. <description>The lifetime of access tokens in minutes.</description>
  255. </property>
  256. <property>
  257. <name>dfs.datanode.data.dir</name>
  258. <value>file://${hadoop.tmp.dir}/dfs/data</value>
  259. <description>Determines where on the local filesystem an DFS data node
  260. should store its blocks. If this is a comma-delimited
  261. list of directories, then data will be stored in all named
  262. directories, typically on different devices.
  263. Directories that do not exist are ignored.
  264. </description>
  265. </property>
  266. <property>
  267. <name>dfs.datanode.data.dir.perm</name>
  268. <value>700</value>
  269. <description>Permissions for the directories on on the local filesystem where
  270. the DFS data node store its blocks. The permissions can either be octal or
  271. symbolic.</description>
  272. </property>
  273. <property>
  274. <name>dfs.replication</name>
  275. <value>3</value>
  276. <description>Default block replication.
  277. The actual number of replications can be specified when the file is created.
  278. The default is used if replication is not specified in create time.
  279. </description>
  280. </property>
  281. <property>
  282. <name>dfs.replication.max</name>
  283. <value>512</value>
  284. <description>Maximal block replication.
  285. </description>
  286. </property>
  287. <property>
  288. <name>dfs.namenode.replication.min</name>
  289. <value>1</value>
  290. <description>Minimal block replication.
  291. </description>
  292. </property>
  293. <property>
  294. <name>dfs.blocksize</name>
  295. <value>67108864</value>
  296. <description>
  297. The default block size for new files, in bytes.
  298. You can use the following suffix (case insensitive):
  299. k(kilo), m(mega), g(giga), t(tera), p(peta), e(exa) to specify the size (such as 128k, 512m, 1g, etc.),
  300. Or provide complete size in bytes (such as 134217728 for 128 MB).
  301. </description>
  302. </property>
  303. <property>
  304. <name>dfs.client.block.write.retries</name>
  305. <value>3</value>
  306. <description>The number of retries for writing blocks to the data nodes,
  307. before we signal failure to the application.
  308. </description>
  309. </property>
  310. <property>
  311. <name>dfs.client.block.write.replace-datanode-on-failure.enable</name>
  312. <value>true</value>
  313. <description>
  314. If there is a datanode/network failure in the write pipeline,
  315. DFSClient will try to remove the failed datanode from the pipeline
  316. and then continue writing with the remaining datanodes. As a result,
  317. the number of datanodes in the pipeline is decreased. The feature is
  318. to add new datanodes to the pipeline.
  319. This is a site-wide property to enable/disable the feature.
  320. When the cluster size is extremely small, e.g. 3 nodes or less, cluster
  321. administrators may want to set the policy to NEVER in the default
  322. configuration file or disable this feature. Otherwise, users may
  323. experience an unusually high rate of pipeline failures since it is
  324. impossible to find new datanodes for replacement.
  325. See also dfs.client.block.write.replace-datanode-on-failure.policy
  326. </description>
  327. </property>
  328. <property>
  329. <name>dfs.client.block.write.replace-datanode-on-failure.policy</name>
  330. <value>DEFAULT</value>
  331. <description>
  332. This property is used only if the value of
  333. dfs.client.block.write.replace-datanode-on-failure.enable is true.
  334. ALWAYS: always add a new datanode when an existing datanode is removed.
  335. NEVER: never add a new datanode.
  336. DEFAULT:
  337. Let r be the replication number.
  338. Let n be the number of existing datanodes.
  339. Add a new datanode only if r is greater than or equal to 3 and either
  340. (1) floor(r/2) is greater than or equal to n; or
  341. (2) r is greater than n and the block is hflushed/appended.
  342. </description>
  343. </property>
  344. <property>
  345. <name>dfs.blockreport.intervalMsec</name>
  346. <value>21600000</value>
  347. <description>Determines block reporting interval in milliseconds.</description>
  348. </property>
  349. <property>
  350. <name>dfs.blockreport.initialDelay</name> <value>0</value>
  351. <description>Delay for first block report in seconds.</description>
  352. </property>
  353. <property>
  354. <name>dfs.datanode.directoryscan.interval</name>
  355. <value>21600</value>
  356. <description>Interval in seconds for Datanode to scan data directories and
  357. reconcile the difference between blocks in memory and on the disk.
  358. </description>
  359. </property>
  360. <property>
  361. <name>dfs.datanode.directoryscan.threads</name>
  362. <value>1</value>
  363. <description>How many threads should the threadpool used to compile reports
  364. for volumes in parallel have.
  365. </description>
  366. </property>
  367. <property>
  368. <name>dfs.heartbeat.interval</name>
  369. <value>3</value>
  370. <description>Determines datanode heartbeat interval in seconds.</description>
  371. </property>
  372. <property>
  373. <name>dfs.namenode.handler.count</name>
  374. <value>10</value>
  375. <description>The number of server threads for the namenode.</description>
  376. </property>
  377. <property>
  378. <name>dfs.namenode.safemode.threshold-pct</name>
  379. <value>0.999f</value>
  380. <description>
  381. Specifies the percentage of blocks that should satisfy
  382. the minimal replication requirement defined by dfs.namenode.replication.min.
  383. Values less than or equal to 0 mean not to wait for any particular
  384. percentage of blocks before exiting safemode.
  385. Values greater than 1 will make safe mode permanent.
  386. </description>
  387. </property>
  388. <property>
  389. <name>dfs.namenode.safemode.min.datanodes</name>
  390. <value>0</value>
  391. <description>
  392. Specifies the number of datanodes that must be considered alive
  393. before the name node exits safemode.
  394. Values less than or equal to 0 mean not to take the number of live
  395. datanodes into account when deciding whether to remain in safe mode
  396. during startup.
  397. Values greater than the number of datanodes in the cluster
  398. will make safe mode permanent.
  399. </description>
  400. </property>
  401. <property>
  402. <name>dfs.namenode.safemode.extension</name>
  403. <value>30000</value>
  404. <description>
  405. Determines extension of safe mode in milliseconds
  406. after the threshold level is reached.
  407. </description>
  408. </property>
  409. <property>
  410. <name>dfs.datanode.balance.bandwidthPerSec</name>
  411. <value>1048576</value>
  412. <description>
  413. Specifies the maximum amount of bandwidth that each datanode
  414. can utilize for the balancing purpose in term of
  415. the number of bytes per second.
  416. </description>
  417. </property>
  418. <property>
  419. <name>dfs.hosts</name>
  420. <value></value>
  421. <description>Names a file that contains a list of hosts that are
  422. permitted to connect to the namenode. The full pathname of the file
  423. must be specified. If the value is empty, all hosts are
  424. permitted.</description>
  425. </property>
  426. <property>
  427. <name>dfs.hosts.exclude</name>
  428. <value></value>
  429. <description>Names a file that contains a list of hosts that are
  430. not permitted to connect to the namenode. The full pathname of the
  431. file must be specified. If the value is empty, no hosts are
  432. excluded.</description>
  433. </property>
  434. <property>
  435. <name>dfs.namenode.max.objects</name>
  436. <value>0</value>
  437. <description>The maximum number of files, directories and blocks
  438. dfs supports. A value of zero indicates no limit to the number
  439. of objects that dfs supports.
  440. </description>
  441. </property>
  442. <property>
  443. <name>dfs.namenode.decommission.interval</name>
  444. <value>30</value>
  445. <description>Namenode periodicity in seconds to check if decommission is
  446. complete.</description>
  447. </property>
  448. <property>
  449. <name>dfs.namenode.decommission.nodes.per.interval</name>
  450. <value>5</value>
  451. <description>The number of nodes namenode checks if decommission is complete
  452. in each dfs.namenode.decommission.interval.</description>
  453. </property>
  454. <property>
  455. <name>dfs.namenode.replication.interval</name>
  456. <value>3</value>
  457. <description>The periodicity in seconds with which the namenode computes
  458. repliaction work for datanodes. </description>
  459. </property>
  460. <property>
  461. <name>dfs.namenode.accesstime.precision</name>
  462. <value>3600000</value>
  463. <description>The access time for HDFS file is precise upto this value.
  464. The default value is 1 hour. Setting a value of 0 disables
  465. access times for HDFS.
  466. </description>
  467. </property>
  468. <property>
  469. <name>dfs.datanode.plugins</name>
  470. <value></value>
  471. <description>Comma-separated list of datanode plug-ins to be activated.
  472. </description>
  473. </property>
  474. <property>
  475. <name>dfs.namenode.plugins</name>
  476. <value></value>
  477. <description>Comma-separated list of namenode plug-ins to be activated.
  478. </description>
  479. </property>
  480. <property>
  481. <name>dfs.stream-buffer-size</name>
  482. <value>4096</value>
  483. <description>The size of buffer to stream files.
  484. The size of this buffer should probably be a multiple of hardware
  485. page size (4096 on Intel x86), and it determines how much data is
  486. buffered during read and write operations.</description>
  487. </property>
  488. <property>
  489. <name>dfs.bytes-per-checksum</name>
  490. <value>512</value>
  491. <description>The number of bytes per checksum. Must not be larger than
  492. dfs.stream-buffer-size</description>
  493. </property>
  494. <property>
  495. <name>dfs.client-write-packet-size</name>
  496. <value>65536</value>
  497. <description>Packet size for clients to write</description>
  498. </property>
  499. <property>
  500. <name>dfs.namenode.checkpoint.dir</name>
  501. <value>file://${hadoop.tmp.dir}/dfs/namesecondary</value>
  502. <description>Determines where on the local filesystem the DFS secondary
  503. name node should store the temporary images to merge.
  504. If this is a comma-delimited list of directories then the image is
  505. replicated in all of the directories for redundancy.
  506. </description>
  507. </property>
  508. <property>
  509. <name>dfs.namenode.checkpoint.edits.dir</name>
  510. <value>${dfs.namenode.checkpoint.dir}</value>
  511. <description>Determines where on the local filesystem the DFS secondary
  512. name node should store the temporary edits to merge.
  513. If this is a comma-delimited list of directoires then teh edits is
  514. replicated in all of the directoires for redundancy.
  515. Default value is same as dfs.namenode.checkpoint.dir
  516. </description>
  517. </property>
  518. <property>
  519. <name>dfs.namenode.checkpoint.period</name>
  520. <value>3600</value>
  521. <description>The number of seconds between two periodic checkpoints.
  522. </description>
  523. </property>
  524. <property>
  525. <name>dfs.namenode.checkpoint.txns</name>
  526. <value>40000</value>
  527. <description>The Secondary NameNode or CheckpointNode will create a checkpoint
  528. of the namespace every 'dfs.namenode.checkpoint.txns' transactions, regardless
  529. of whether 'dfs.namenode.checkpoint.period' has expired.
  530. </description>
  531. </property>
  532. <property>
  533. <name>dfs.namenode.checkpoint.check.period</name>
  534. <value>60</value>
  535. <description>The SecondaryNameNode and CheckpointNode will poll the NameNode
  536. every 'dfs.namenode.checkpoint.check.period' seconds to query the number
  537. of uncheckpointed transactions.
  538. </description>
  539. </property>
  540. <property>
  541. <name>dfs.namenode.num.checkpoints.retained</name>
  542. <value>2</value>
  543. <description>The number of image checkpoint files that will be retained by
  544. the NameNode and Secondary NameNode in their storage directories. All edit
  545. logs necessary to recover an up-to-date namespace from the oldest retained
  546. checkpoint will also be retained.
  547. </description>
  548. </property>
  549. <property>
  550. <name>dfs.namenode.num.extra.edits.retained</name>
  551. <value>1000000</value>
  552. <description>The number of extra transactions which should be retained
  553. beyond what is minimally necessary for a NN restart. This can be useful for
  554. audit purposes or for an HA setup where a remote Standby Node may have
  555. been offline for some time and need to have a longer backlog of retained
  556. edits in order to start again.
  557. Typically each edit is on the order of a few hundred bytes, so the default
  558. of 1 million edits should be on the order of hundreds of MBs or low GBs.
  559. </description>
  560. </property>
  561. <property>
  562. <name>dfs.namenode.delegation.key.update-interval</name>
  563. <value>86400000</value>
  564. <description>The update interval for master key for delegation tokens
  565. in the namenode in milliseconds.
  566. </description>
  567. </property>
  568. <property>
  569. <name>dfs.namenode.delegation.token.max-lifetime</name>
  570. <value>604800000</value>
  571. <description>The maximum lifetime in milliseconds for which a delegation
  572. token is valid.
  573. </description>
  574. </property>
  575. <property>
  576. <name>dfs.namenode.delegation.token.renew-interval</name>
  577. <value>86400000</value>
  578. <description>The renewal interval for delegation token in milliseconds.
  579. </description>
  580. </property>
  581. <property>
  582. <name>dfs.datanode.failed.volumes.tolerated</name>
  583. <value>0</value>
  584. <description>The number of volumes that are allowed to
  585. fail before a datanode stops offering service. By default
  586. any volume failure will cause a datanode to shutdown.
  587. </description>
  588. </property>
  589. <property>
  590. <name>dfs.image.compress</name>
  591. <value>false</value>
  592. <description>Should the dfs image be compressed?
  593. </description>
  594. </property>
  595. <property>
  596. <name>dfs.image.compression.codec</name>
  597. <value>org.apache.hadoop.io.compress.DefaultCodec</value>
  598. <description>If the dfs image is compressed, how should they be compressed?
  599. This has to be a codec defined in io.compression.codecs.
  600. </description>
  601. </property>
  602. <property>
  603. <name>dfs.image.transfer.bandwidthPerSec</name>
  604. <value>0</value>
  605. <description>
  606. Specifies the maximum amount of bandwidth that can be utilized for image
  607. transfer in term of the number of bytes per second.
  608. A default value of 0 indicates that throttling is disabled.
  609. </description>
  610. </property>
  611. <property>
  612. <name>dfs.namenode.support.allow.format</name>
  613. <value>true</value>
  614. <description>Does HDFS namenode allow itself to be formatted?
  615. You may consider setting this to false for any production
  616. cluster, to avoid any possibility of formatting a running DFS.
  617. </description>
  618. </property>
  619. <property>
  620. <name>dfs.datanode.max.transfer.threads</name>
  621. <value>4096</value>
  622. <description>
  623. Specifies the maximum number of threads to use for transferring data
  624. in and out of the DN.
  625. </description>
  626. </property>
  627. <property>
  628. <name>dfs.datanode.readahead.bytes</name>
  629. <value>4193404</value>
  630. <description>
  631. While reading block files, if the Hadoop native libraries are available,
  632. the datanode can use the posix_fadvise system call to explicitly
  633. page data into the operating system buffer cache ahead of the current
  634. reader's position. This can improve performance especially when
  635. disks are highly contended.
  636. This configuration specifies the number of bytes ahead of the current
  637. read position which the datanode will attempt to read ahead. This
  638. feature may be disabled by configuring this property to 0.
  639. If the native libraries are not available, this configuration has no
  640. effect.
  641. </description>
  642. </property>
  643. <property>
  644. <name>dfs.datanode.drop.cache.behind.reads</name>
  645. <value>false</value>
  646. <description>
  647. In some workloads, the data read from HDFS is known to be significantly
  648. large enough that it is unlikely to be useful to cache it in the
  649. operating system buffer cache. In this case, the DataNode may be
  650. configured to automatically purge all data from the buffer cache
  651. after it is delivered to the client. This behavior is automatically
  652. disabled for workloads which read only short sections of a block
  653. (e.g HBase random-IO workloads).
  654. This may improve performance for some workloads by freeing buffer
  655. cache spage usage for more cacheable data.
  656. If the Hadoop native libraries are not available, this configuration
  657. has no effect.
  658. </description>
  659. </property>
  660. <property>
  661. <name>dfs.datanode.drop.cache.behind.writes</name>
  662. <value>false</value>
  663. <description>
  664. In some workloads, the data written to HDFS is known to be significantly
  665. large enough that it is unlikely to be useful to cache it in the
  666. operating system buffer cache. In this case, the DataNode may be
  667. configured to automatically purge all data from the buffer cache
  668. after it is written to disk.
  669. This may improve performance for some workloads by freeing buffer
  670. cache spage usage for more cacheable data.
  671. If the Hadoop native libraries are not available, this configuration
  672. has no effect.
  673. </description>
  674. </property>
  675. <property>
  676. <name>dfs.datanode.sync.behind.writes</name>
  677. <value>false</value>
  678. <description>
  679. If this configuration is enabled, the datanode will instruct the
  680. operating system to enqueue all written data to the disk immediately
  681. after it is written. This differs from the usual OS policy which
  682. may wait for up to 30 seconds before triggering writeback.
  683. This may improve performance for some workloads by smoothing the
  684. IO profile for data written to disk.
  685. If the Hadoop native libraries are not available, this configuration
  686. has no effect.
  687. </description>
  688. </property>
  689. <property>
  690. <name>dfs.client.failover.max.attempts</name>
  691. <value>15</value>
  692. <description>
  693. Expert only. The number of client failover attempts that should be
  694. made before the failover is considered failed.
  695. </description>
  696. </property>
  697. <property>
  698. <name>dfs.client.failover.sleep.base.millis</name>
  699. <value>500</value>
  700. <description>
  701. Expert only. The time to wait, in milliseconds, between failover
  702. attempts increases exponentially as a function of the number of
  703. attempts made so far, with a random factor of +/- 50%. This option
  704. specifies the base value used in the failover calculation. The
  705. first failover will retry immediately. The 2nd failover attempt
  706. will delay at least dfs.client.failover.sleep.base.millis
  707. milliseconds. And so on.
  708. </description>
  709. </property>
  710. <property>
  711. <name>dfs.client.failover.sleep.max.millis</name>
  712. <value>15000</value>
  713. <description>
  714. Expert only. The time to wait, in milliseconds, between failover
  715. attempts increases exponentially as a function of the number of
  716. attempts made so far, with a random factor of +/- 50%. This option
  717. specifies the maximum value to wait between failovers.
  718. Specifically, the time between two failover attempts will not
  719. exceed +/- 50% of dfs.client.failover.sleep.max.millis
  720. milliseconds.
  721. </description>
  722. </property>
  723. <property>
  724. <name>dfs.client.failover.connection.retries</name>
  725. <value>0</value>
  726. <description>
  727. Expert only. Indicates the number of retries a failover IPC client
  728. will make to establish a server connection.
  729. </description>
  730. </property>
  731. <property>
  732. <name>dfs.client.failover.connection.retries.on.timeouts</name>
  733. <value>0</value>
  734. <description>
  735. Expert only. The number of retry attempts a failover IPC client
  736. will make on socket timeout when establishing a server connection.
  737. </description>
  738. </property>
  739. <property>
  740. <name>dfs.nameservices</name>
  741. <value></value>
  742. <description>
  743. Comma-separated list of nameservices.
  744. </description>
  745. </property>
  746. <property>
  747. <name>dfs.nameservice.id</name>
  748. <value></value>
  749. <description>
  750. The ID of this nameservice. If the nameservice ID is not
  751. configured or more than one nameservice is configured for
  752. dfs.nameservices it is determined automatically by
  753. matching the local node's address with the configured address.
  754. </description>
  755. </property>
  756. <property>
  757. <name>dfs.ha.namenodes.EXAMPLENAMESERVICE</name>
  758. <value></value>
  759. <description>
  760. The prefix for a given nameservice, contains a comma-separated
  761. list of namenodes for a given nameservice (eg EXAMPLENAMESERVICE).
  762. </description>
  763. </property>
  764. <property>
  765. <name>dfs.ha.namenode.id</name>
  766. <value></value>
  767. <description>
  768. The ID of this namenode. If the namenode ID is not configured it
  769. is determined automatically by matching the local node's address
  770. with the configured address.
  771. </description>
  772. </property>
  773. <property>
  774. <name>dfs.ha.log-roll.period</name>
  775. <value>120</value>
  776. <description>
  777. How often, in seconds, the StandbyNode should ask the active to
  778. roll edit logs. Since the StandbyNode only reads from finalized
  779. log segments, the StandbyNode will only be as up-to-date as how
  780. often the logs are rolled. Note that failover triggers a log roll
  781. so the StandbyNode will be up to date before it becomes active.
  782. </description>
  783. </property>
  784. <property>
  785. <name>dfs.ha.tail-edits.period</name>
  786. <value>60</value>
  787. <description>
  788. How often, in seconds, the StandbyNode should check for new
  789. finalized log segments in the shared edits log.
  790. </description>
  791. </property>
  792. <property>
  793. <name>dfs.ha.automatic-failover.enabled</name>
  794. <value>false</value>
  795. <description>
  796. Whether automatic failover is enabled. See the HDFS High
  797. Availability documentation for details on automatic HA
  798. configuration.
  799. </description>
  800. </property>
  801. <property>
  802. <name>dfs.support.append</name>
  803. <value>true</value>
  804. <description>
  805. Does HDFS allow appends to files?
  806. </description>
  807. </property>
  808. <property>
  809. <name>dfs.client.local.interfaces</name>
  810. <value></value>
  811. <description>A comma separated list of network interface names to use
  812. for data transfer between the client and datanodes. When creating
  813. a connection to read from or write to a datanode, the client
  814. chooses one of the specified interfaces at random and binds its
  815. socket to the IP of that interface. Individual names may be
  816. specified as either an interface name (eg "eth0"), a subinterface
  817. name (eg "eth0:0"), or an IP address (which may be specified using
  818. CIDR notation to match a range of IPs).
  819. </description>
  820. </property>
  821. <property>
  822. <name>dfs.namenode.kerberos.internal.spnego.principal</name>
  823. <value>${dfs.web.authentication.kerberos.principal}</value>
  824. </property>
  825. <property>
  826. <name>dfs.secondary.namenode.kerberos.internal.spnego.principal</name>
  827. <value>${dfs.web.authentication.kerberos.principal}</value>
  828. </property>
  829. <property>
  830. <name>dfs.namenode.invalidate.work.pct.per.iteration</name>
  831. <value>0.32f</value>
  832. <description>
  833. *Note*: Advanced property. Change with caution.
  834. This determines the percentage amount of block
  835. invalidations (deletes) to do over a single DN heartbeat
  836. deletion command. The final deletion count is determined by applying this
  837. percentage to the number of live nodes in the system.
  838. The resultant number is the number of blocks from the deletion list
  839. chosen for proper invalidation over a single heartbeat of a single DN.
  840. Value should be a positive, non-zero percentage in float notation (X.Yf),
  841. with 1.0f meaning 100%.
  842. </description>
  843. </property>
  844. <property>
  845. <name>dfs.namenode.replication.work.multiplier.per.iteration</name>
  846. <value>2</value>
  847. <description>
  848. *Note*: Advanced property. Change with caution.
  849. This determines the total amount of block transfers to begin in
  850. parallel at a DN, for replication, when such a command list is being
  851. sent over a DN heartbeat by the NN. The actual number is obtained by
  852. multiplying this multiplier with the total number of live nodes in the
  853. cluster. The result number is the number of blocks to begin transfers
  854. immediately for, per DN heartbeat. This number can be any positive,
  855. non-zero integer.
  856. </description>
  857. </property>
  858. <property>
  859. <name>dfs.webhdfs.enabled</name>
  860. <value>false</value>
  861. <description>
  862. Enable WebHDFS (REST API) in Namenodes and Datanodes.
  863. </description>
  864. </property>
  865. <property>
  866. <name>hadoop.fuse.connection.timeout</name>
  867. <value>300</value>
  868. <description>
  869. The minimum number of seconds that we'll cache libhdfs connection objects
  870. in fuse_dfs. Lower values will result in lower memory consumption; higher
  871. values may speed up access by avoiding the overhead of creating new
  872. connection objects.
  873. </description>
  874. </property>
  875. <property>
  876. <name>hadoop.fuse.timer.period</name>
  877. <value>5</value>
  878. <description>
  879. The number of seconds between cache expiry checks in fuse_dfs. Lower values
  880. will result in fuse_dfs noticing changes to Kerberos ticket caches more
  881. quickly.
  882. </description>
  883. </property>
  884. <property>
  885. <name>dfs.metrics.percentiles.intervals</name>
  886. <value></value>
  887. <description>
  888. Comma-delimited set of integers denoting the desired rollover intervals
  889. (in seconds) for percentile latency metrics on the Namenode and Datanode.
  890. By default, percentile latency metrics are disabled.
  891. </description>
  892. </property>
  893. </configuration>