2 <?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
5 * Licensed to the Apache Software Foundation (ASF) under one
6 * or more contributor license agreements. See the NOTICE file
7 * distributed with this work for additional information
8 * regarding copyright ownership. The ASF licenses this file
9 * to you under the Apache License, Version 2.0 (the
10 * "License"); you may not use this file except in compliance
11 * with the License. You may obtain a copy of the License at
13 * http://www.apache.org/licenses/LICENSE-2.0
15 * Unless required by applicable law or agreed to in writing, software
16 * distributed under the License is distributed on an "AS IS" BASIS,
17 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
18 * See the License for the specific language governing permissions and
19 * limitations under the License.
26 The important configs. are listed near the top. You should change
27 at least the setting for hbase.tmp.dir. Other settings will change
28 dependent on whether you are running hbase in standalone mode or
29 distributed. See the hbase reference guide for requirements and
30 guidance making configuration.
32 This file does not contain all possible configurations. The file would be
33 much larger if it carried everything. The absent configurations will only be
34 found through source code reading. The idea is that such configurations are
35 exotic and only those who would go to the trouble of reading a particular
36 section in the code would be knowledgeable or invested enough in ever wanting
37 to alter such configurations, so we do not list them here. Listing all
38 possible configurations would overwhelm and obscure the important.
42 <!--Configs you will likely change are listed here at the top of the file.
45 <name>hbase.tmp.dir</name>
46 <value>${java.io.tmpdir}/hbase-${user.name}</value>
47 <description>Temporary directory on the local filesystem.
48 Change this setting to point to a location more permanent
49 than '/tmp', the usual resolve for java.io.tmpdir, as the
50 '/tmp' directory is cleared on machine restart.</description>
53 <name>hbase.rootdir</name>
54 <value>${hbase.tmp.dir}/hbase</value>
55 <description>The directory shared by region servers and into
56 which HBase persists. The URL should be 'fully-qualified'
57 to include the filesystem scheme. For example, to specify the
58 HDFS directory '/hbase' where the HDFS instance's namenode is
59 running at namenode.example.org on port 9000, set this value to:
60 hdfs://namenode.example.org:9000/hbase. By default, we write
61 to whatever ${hbase.tmp.dir} is set too -- usually /tmp --
62 so change this configuration or else all data will be lost on
63 machine restart.</description>
66 <name>hbase.cluster.distributed</name>
68 <description>The mode the cluster will be in. Possible values are
69 false for standalone mode and true for distributed mode. If
70 false, startup will run all HBase and ZooKeeper daemons together
71 in the one JVM.</description>
74 <name>hbase.zookeeper.quorum</name>
75 <value>localhost</value>
76 <description>Comma separated list of servers in the ZooKeeper ensemble
77 (This config. should have been named hbase.zookeeper.ensemble).
78 For example, "host1.mydomain.com,host2.mydomain.com,host3.mydomain.com".
79 By default this is set to localhost for local and pseudo-distributed modes
80 of operation. For a fully-distributed setup, this should be set to a full
81 list of ZooKeeper ensemble servers. If HBASE_MANAGES_ZK is set in hbase-env.sh
82 this is the list of servers which hbase will start/stop ZooKeeper on as
83 part of cluster start/stop. Client-side, we will take this list of
84 ensemble members and put it together with the hbase.zookeeper.property.clientPort
85 config. and pass it into zookeeper constructor as the connectString
86 parameter.</description>
88 <!--The above are the important configurations for getting hbase up
92 <name>zookeeper.recovery.retry.maxsleeptime</name>
94 <description>Max sleep time before retry zookeeper operations in milliseconds,
95 a max time is needed here so that sleep time won't grow unboundedly
99 <name>hbase.local.dir</name>
100 <value>${hbase.tmp.dir}/local/</value>
101 <description>Directory on the local filesystem to be used
102 as a local storage.</description>
105 <!--Master configurations-->
107 <name>hbase.master.port</name>
109 <description>The port the HBase Master should bind to.</description>
112 <name>hbase.master.info.port</name>
114 <description>The port for the HBase Master web UI.
115 Set to -1 if you do not want a UI instance run.</description>
118 <name>hbase.master.info.bindAddress</name>
119 <value>0.0.0.0</value>
120 <description>The bind address for the HBase Master web UI
124 <name>hbase.master.logcleaner.plugins</name>
125 <value>org.apache.hadoop.hbase.master.cleaner.TimeToLiveLogCleaner,org.apache.hadoop.hbase.master.cleaner.TimeToLiveProcedureWALCleaner</value>
126 <description>A comma-separated list of BaseLogCleanerDelegate invoked by
127 the LogsCleaner service. These WAL cleaners are called in order,
128 so put the cleaner that prunes the most files in front. To
129 implement your own BaseLogCleanerDelegate, just put it in HBase's classpath
130 and add the fully qualified class name here. Always add the above
131 default log cleaners in the list.</description>
134 <name>hbase.master.logcleaner.ttl</name>
135 <value>600000</value>
136 <description>How long a WAL remain in the archive ({hbase.rootdir}/oldWALs) directory,
137 after which it will be cleaned by a Master thread. The value is in milliseconds.</description>
140 <name>hbase.master.procedurewalcleaner.ttl</name>
141 <value>604800000</value>
142 <description>How long a Procedure WAL will remain in the
143 archive directory, after which it will be cleaned
144 by a Master thread. The value is in milliseconds.</description>
147 <name>hbase.master.hfilecleaner.plugins</name>
148 <value>org.apache.hadoop.hbase.master.cleaner.TimeToLiveHFileCleaner</value>
149 <description>A comma-separated list of BaseHFileCleanerDelegate invoked by
150 the HFileCleaner service. These HFiles cleaners are called in order,
151 so put the cleaner that prunes the most files in front. To
152 implement your own BaseHFileCleanerDelegate, just put it in HBase's classpath
153 and add the fully qualified class name here. Always add the above
154 default log cleaners in the list as they will be overwritten in
155 hbase-site.xml.</description>
158 <name>hbase.master.infoserver.redirect</name>
160 <description>Whether or not the Master listens to the Master web
161 UI port (hbase.master.info.port) and redirects requests to the web
162 UI server shared by the Master and RegionServer. Config. makes
163 sense when Master is serving Regions (not the default).</description>
166 <name>hbase.master.fileSplitTimeout</name>
167 <value>600000</value>
168 <description>Splitting a region, how long to wait on the file-splitting
169 step before aborting the attempt. Default: 600000. This setting used
170 to be known as hbase.regionserver.fileSplitTimeout in hbase-1.x.
171 Split is now run master-side hence the rename (If a
172 'hbase.master.fileSplitTimeout' setting found, will use it to
173 prime the current 'hbase.master.fileSplitTimeout'
174 Configuration.</description>
177 <!--RegionServer configurations-->
179 <name>hbase.regionserver.port</name>
181 <description>The port the HBase RegionServer binds to.</description>
184 <name>hbase.regionserver.info.port</name>
186 <description>The port for the HBase RegionServer web UI
187 Set to -1 if you do not want the RegionServer UI to run.</description>
190 <name>hbase.regionserver.info.bindAddress</name>
191 <value>0.0.0.0</value>
192 <description>The address for the HBase RegionServer web UI</description>
195 <name>hbase.regionserver.info.port.auto</name>
197 <description>Whether or not the Master or RegionServer
198 UI should search for a port to bind to. Enables automatic port
199 search if hbase.regionserver.info.port is already in use.
200 Useful for testing, turned off by default.</description>
203 <name>hbase.regionserver.handler.count</name>
205 <description>Count of RPC Listener instances spun up on RegionServers.
206 Same property is used by the Master for count of master handlers.
207 Too many handlers can be counter-productive. Make it a multiple of
208 CPU count. If mostly read-only, handlers count close to cpu count
209 does well. Start with twice the CPU count and tune from there.</description>
212 <name>hbase.ipc.server.callqueue.handler.factor</name>
214 <description>Factor to determine the number of call queues.
215 A value of 0 means a single queue shared between all the handlers.
216 A value of 1 means that each handler has its own queue.</description>
219 <name>hbase.ipc.server.callqueue.read.ratio</name>
221 <description>Split the call queues into read and write queues.
222 The specified interval (which should be between 0.0 and 1.0)
223 will be multiplied by the number of call queues.
224 A value of 0 indicate to not split the call queues, meaning that both read and write
225 requests will be pushed to the same set of queues.
226 A value lower than 0.5 means that there will be less read queues than write queues.
227 A value of 0.5 means there will be the same number of read and write queues.
228 A value greater than 0.5 means that there will be more read queues than write queues.
229 A value of 1.0 means that all the queues except one are used to dispatch read requests.
231 Example: Given the total number of call queues being 10
232 a read.ratio of 0 means that: the 10 queues will contain both read/write requests.
233 a read.ratio of 0.3 means that: 3 queues will contain only read requests
234 and 7 queues will contain only write requests.
235 a read.ratio of 0.5 means that: 5 queues will contain only read requests
236 and 5 queues will contain only write requests.
237 a read.ratio of 0.8 means that: 8 queues will contain only read requests
238 and 2 queues will contain only write requests.
239 a read.ratio of 1 means that: 9 queues will contain only read requests
240 and 1 queues will contain only write requests.
244 <name>hbase.ipc.server.callqueue.scan.ratio</name>
246 <description>Given the number of read call queues, calculated from the total number
247 of call queues multiplied by the callqueue.read.ratio, the scan.ratio property
248 will split the read call queues into small-read and long-read queues.
249 A value lower than 0.5 means that there will be less long-read queues than short-read queues.
250 A value of 0.5 means that there will be the same number of short-read and long-read queues.
251 A value greater than 0.5 means that there will be more long-read queues than short-read queues
252 A value of 0 or 1 indicate to use the same set of queues for gets and scans.
254 Example: Given the total number of read call queues being 8
255 a scan.ratio of 0 or 1 means that: 8 queues will contain both long and short read requests.
256 a scan.ratio of 0.3 means that: 2 queues will contain only long-read requests
257 and 6 queues will contain only short-read requests.
258 a scan.ratio of 0.5 means that: 4 queues will contain only long-read requests
259 and 4 queues will contain only short-read requests.
260 a scan.ratio of 0.8 means that: 6 queues will contain only long-read requests
261 and 2 queues will contain only short-read requests.
265 <name>hbase.regionserver.msginterval</name>
267 <description>Interval between messages from the RegionServer to Master
268 in milliseconds.</description>
271 <name>hbase.regionserver.logroll.period</name>
272 <value>3600000</value>
273 <description>Period at which we will roll the commit log regardless
274 of how many edits it has.</description>
277 <name>hbase.regionserver.logroll.errors.tolerated</name>
279 <description>The number of consecutive WAL close errors we will allow
280 before triggering a server abort. A setting of 0 will cause the
281 region server to abort if closing the current WAL writer fails during
282 log rolling. Even a small value (2 or 3) will allow a region server
283 to ride over transient HDFS errors.</description>
286 <name>hbase.regionserver.hlog.reader.impl</name>
287 <value>org.apache.hadoop.hbase.regionserver.wal.ProtobufLogReader</value>
288 <description>The WAL file reader implementation.</description>
291 <name>hbase.regionserver.hlog.writer.impl</name>
292 <value>org.apache.hadoop.hbase.regionserver.wal.ProtobufLogWriter</value>
293 <description>The WAL file writer implementation.</description>
296 <name>hbase.regionserver.global.memstore.size</name>
298 <description>Maximum size of all memstores in a region server before new
299 updates are blocked and flushes are forced. Defaults to 40% of heap (0.4).
300 Updates are blocked and flushes are forced until size of all memstores
301 in a region server hits hbase.regionserver.global.memstore.size.lower.limit.
302 The default value in this configuration has been intentionally left empty in order to
303 honor the old hbase.regionserver.global.memstore.upperLimit property if present.
307 <name>hbase.regionserver.global.memstore.size.lower.limit</name>
309 <description>Maximum size of all memstores in a region server before flushes
310 are forced. Defaults to 95% of hbase.regionserver.global.memstore.size
311 (0.95). A 100% value for this value causes the minimum possible flushing
312 to occur when updates are blocked due to memstore limiting. The default
313 value in this configuration has been intentionally left empty in order to
314 honor the old hbase.regionserver.global.memstore.lowerLimit property if
319 <name>hbase.systemtables.compacting.memstore.type</name>
321 <description>Determines the type of memstore to be used for system tables like
322 META, namespace tables etc. By default NONE is the type and hence we use the
323 default memstore for all the system tables. If we need to use compacting
324 memstore for system tables then set this property to BASIC/EAGER
328 <name>hbase.regionserver.optionalcacheflushinterval</name>
329 <value>3600000</value>
331 Maximum amount of time an edit lives in memory before being automatically flushed.
332 Default 1 hour. Set it to 0 to disable automatic flushing.
336 <name>hbase.regionserver.dns.interface</name>
337 <value>default</value>
338 <description>The name of the Network Interface from which a region server
339 should report its IP address.</description>
342 <name>hbase.regionserver.dns.nameserver</name>
343 <value>default</value>
344 <description>The host name or IP address of the name server (DNS)
345 which a region server should use to determine the host name used by the
346 master for communication and display purposes.</description>
349 <name>hbase.regionserver.region.split.policy</name>
350 <value>org.apache.hadoop.hbase.regionserver.SteppingSplitPolicy</value>
352 A split policy determines when a region should be split. The various
353 other split policies that are available currently are BusyRegionSplitPolicy,
354 ConstantSizeRegionSplitPolicy, DisabledRegionSplitPolicy,
355 DelimitedKeyPrefixRegionSplitPolicy, KeyPrefixRegionSplitPolicy, and
356 SteppingSplitPolicy. DisabledRegionSplitPolicy blocks manual region splitting.
360 <name>hbase.regionserver.regionSplitLimit</name>
363 Limit for the number of regions after which no more region splitting
364 should take place. This is not hard limit for the number of regions
365 but acts as a guideline for the regionserver to stop splitting after
366 a certain limit. Default is set to 1000.
370 <!--ZooKeeper configuration-->
372 <name>zookeeper.session.timeout</name>
374 <description>ZooKeeper session timeout in milliseconds. It is used in two different ways.
375 First, this value is used in the ZK client that HBase uses to connect to the ensemble.
376 It is also used by HBase when it starts a ZK server and it is passed as the 'maxSessionTimeout'.
377 See http://hadoop.apache.org/zookeeper/docs/current/zookeeperProgrammers.html#ch_zkSessions.
378 For example, if an HBase region server connects to a ZK ensemble that's also managed
379 by HBase, then the session timeout will be the one specified by this configuration.
380 But, a region server that connects to an ensemble managed with a different configuration
381 will be subjected that ensemble's maxSessionTimeout. So, even though HBase might propose
382 using 90 seconds, the ensemble can have a max timeout lower than this and it will take
383 precedence. The current default that ZK ships with is 40 seconds, which is lower than
388 <name>zookeeper.znode.parent</name>
389 <value>/hbase</value>
390 <description>Root ZNode for HBase in ZooKeeper. All of HBase's ZooKeeper
391 files that are configured with a relative path will go under this node.
392 By default, all of HBase's ZooKeeper file paths are configured with a
393 relative path, so they will all go under this directory unless changed.
397 <name>zookeeper.znode.acl.parent</name>
399 <description>Root ZNode for access control lists.</description>
402 <name>hbase.zookeeper.dns.interface</name>
403 <value>default</value>
404 <description>The name of the Network Interface from which a ZooKeeper server
405 should report its IP address.</description>
408 <name>hbase.zookeeper.dns.nameserver</name>
409 <value>default</value>
410 <description>The host name or IP address of the name server (DNS)
411 which a ZooKeeper server should use to determine the host name used by the
412 master for communication and display purposes.</description>
415 The following three properties are used together to create the list of
416 host:peer_port:leader_port quorum servers for ZooKeeper.
419 <name>hbase.zookeeper.peerport</name>
421 <description>Port used by ZooKeeper peers to talk to each other.
422 See http://hadoop.apache.org/zookeeper/docs/r3.1.1/zookeeperStarted.html#sc_RunningReplicatedZooKeeper
423 for more information.</description>
426 <name>hbase.zookeeper.leaderport</name>
428 <description>Port used by ZooKeeper for leader election.
429 See http://hadoop.apache.org/zookeeper/docs/r3.1.1/zookeeperStarted.html#sc_RunningReplicatedZooKeeper
430 for more information.</description>
432 <!-- End of properties used to generate ZooKeeper host:port quorum list. -->
435 Beginning of properties that are directly mapped from ZooKeeper's zoo.cfg.
436 All properties with an "hbase.zookeeper.property." prefix are converted for
437 ZooKeeper's configuration. Hence, if you want to add an option from zoo.cfg,
438 e.g. "initLimit=10" you would append the following to your configuration:
440 <name>hbase.zookeeper.property.initLimit</name>
445 <name>hbase.zookeeper.property.initLimit</name>
447 <description>Property from ZooKeeper's config zoo.cfg.
448 The number of ticks that the initial synchronization phase can take.</description>
451 <name>hbase.zookeeper.property.syncLimit</name>
453 <description>Property from ZooKeeper's config zoo.cfg.
454 The number of ticks that can pass between sending a request and getting an
455 acknowledgment.</description>
458 <name>hbase.zookeeper.property.dataDir</name>
459 <value>${hbase.tmp.dir}/zookeeper</value>
460 <description>Property from ZooKeeper's config zoo.cfg.
461 The directory where the snapshot is stored.</description>
464 <name>hbase.zookeeper.property.clientPort</name>
466 <description>Property from ZooKeeper's config zoo.cfg.
467 The port at which the clients will connect.</description>
470 <name>hbase.zookeeper.property.maxClientCnxns</name>
472 <description>Property from ZooKeeper's config zoo.cfg.
473 Limit on number of concurrent connections (at the socket level) that a
474 single client, identified by IP address, may make to a single member of
475 the ZooKeeper ensemble. Set high to avoid zk connection issues running
476 standalone and pseudo-distributed.</description>
478 <!-- End of properties that are directly mapped from ZooKeeper's zoo.cfg -->
480 <!--Client configurations-->
482 <name>hbase.client.write.buffer</name>
483 <value>2097152</value>
484 <description>Default size of the BufferedMutator write buffer in bytes.
485 A bigger buffer takes more memory -- on both the client and server
486 side since server instantiates the passed write buffer to process
487 it -- but a larger buffer size reduces the number of RPCs made.
488 For an estimate of server-side memory-used, evaluate
489 hbase.client.write.buffer * hbase.regionserver.handler.count</description>
492 <name>hbase.client.pause</name>
494 <description>General client pause value. Used mostly as value to wait
495 before running a retry of a failed get, region lookup, etc.
496 See hbase.client.retries.number for description of how we backoff from
497 this initial pause amount and how this pause works w/ retries.</description>
500 <name>hbase.client.pause.cqtbe</name>
502 <description>Whether or not to use a special client pause for
503 CallQueueTooBigException (cqtbe). Set this property to a higher value
504 than hbase.client.pause if you observe frequent CQTBE from the same
505 RegionServer and the call queue there keeps full</description>
508 <name>hbase.client.retries.number</name>
510 <description>Maximum retries. Used as maximum for all retryable
511 operations such as the getting of a cell's value, starting a row update,
512 etc. Retry interval is a rough function based on hbase.client.pause. At
513 first we retry at this interval but then with backoff, we pretty quickly reach
514 retrying every ten seconds. See HConstants#RETRY_BACKOFF for how the backup
515 ramps up. Change this setting and hbase.client.pause to suit your workload.</description>
518 <name>hbase.client.max.total.tasks</name>
520 <description>The maximum number of concurrent mutation tasks a single HTable instance will
521 send to the cluster.</description>
524 <name>hbase.client.max.perserver.tasks</name>
526 <description>The maximum number of concurrent mutation tasks a single HTable instance will
527 send to a single region server.</description>
530 <name>hbase.client.max.perregion.tasks</name>
532 <description>The maximum number of concurrent mutation tasks the client will
533 maintain to a single Region. That is, if there is already
534 hbase.client.max.perregion.tasks writes in progress for this region, new puts
535 won't be sent to this region until some writes finishes.</description>
538 <name>hbase.client.perserver.requests.threshold</name>
539 <value>2147483647</value>
540 <description>The max number of concurrent pending requests for one server in all client threads
541 (process level). Exceeding requests will be thrown ServerTooBusyException immediately to prevent
542 user's threads being occupied and blocked by only one slow region server. If you use a fix
543 number of threads to access HBase in a synchronous way, set this to a suitable value which is
544 related to the number of threads will help you. See
545 https://issues.apache.org/jira/browse/HBASE-16388 for details.</description>
548 <name>hbase.client.scanner.caching</name>
549 <value>2147483647</value>
550 <description>Number of rows that we try to fetch when calling next
551 on a scanner if it is not served from (local, client) memory. This configuration
552 works together with hbase.client.scanner.max.result.size to try and use the
553 network efficiently. The default value is Integer.MAX_VALUE by default so that
554 the network will fill the chunk size defined by hbase.client.scanner.max.result.size
555 rather than be limited by a particular number of rows since the size of rows varies
556 table to table. If you know ahead of time that you will not require more than a certain
557 number of rows from a scan, this configuration should be set to that row limit via
558 Scan#setCaching. Higher caching values will enable faster scanners but will eat up more
559 memory and some calls of next may take longer and longer times when the cache is empty.
560 Do not set this value such that the time between invocations is greater than the scanner
561 timeout; i.e. hbase.client.scanner.timeout.period</description>
564 <name>hbase.client.keyvalue.maxsize</name>
565 <value>10485760</value>
566 <description>Specifies the combined maximum allowed size of a KeyValue
567 instance. This is to set an upper boundary for a single entry saved in a
568 storage file. Since they cannot be split it helps avoiding that a region
569 cannot be split any further because the data is too large. It seems wise
570 to set this to a fraction of the maximum region size. Setting it to zero
571 or less disables the check.</description>
574 <name>hbase.server.keyvalue.maxsize</name>
575 <value>10485760</value>
576 <description>Maximum allowed size of an individual cell, inclusive of value and all key
577 components. A value of 0 or less disables the check.
578 The default value is 10MB.
579 This is a safety setting to protect the server from OOM situations.
583 <name>hbase.client.scanner.timeout.period</name>
585 <description>Client scanner lease period in milliseconds.</description>
588 <name>hbase.client.localityCheck.threadPoolSize</name>
592 <!--Miscellaneous configuration-->
594 <name>hbase.bulkload.retries.number</name>
596 <description>Maximum retries. This is maximum number of iterations
597 to atomic bulk loads are attempted in the face of splitting operations
598 0 means never give up.</description>
601 <name>hbase.master.balancer.maxRitPercent</name>
603 <description>The max percent of regions in transition when balancing.
604 The default value is 1.0. So there are no balancer throttling. If set this config to 0.01,
605 It means that there are at most 1% regions in transition when balancing.
606 Then the cluster's availability is at least 99% when balancing.</description>
609 <name>hbase.balancer.period
611 <value>300000</value>
612 <description>Period at which the region balancer runs in the Master.</description>
615 <name>hbase.normalizer.period</name>
616 <value>300000</value>
617 <description>Period at which the region normalizer runs in the Master.</description>
620 <name>hbase.normalizer.min.region.count</name>
622 <description>configure the minimum number of regions</description>
625 <name>hbase.regions.slop</name>
627 <description>Rebalance if any regionserver has average + (average * slop) regions.
628 The default value of this parameter is 0.001 in StochasticLoadBalancer (the default load balancer),
629 while the default is 0.2 in other load balancers (i.e., SimpleLoadBalancer).</description>
632 <name>hbase.server.thread.wakefrequency</name>
634 <description>Time to sleep in between searches for work (in milliseconds).
635 Used as sleep interval by service threads such as log roller.</description>
638 <name>hbase.server.versionfile.writeattempts</name>
641 How many times to retry attempting to write a version file
642 before just aborting. Each attempt is separated by the
643 hbase.server.thread.wakefrequency milliseconds.</description>
646 <name>hbase.hregion.memstore.flush.size</name>
647 <value>134217728</value>
649 Memstore will be flushed to disk if size of the memstore
650 exceeds this number of bytes. Value is checked by a thread that runs
651 every hbase.server.thread.wakefrequency.</description>
654 <name>hbase.hregion.percolumnfamilyflush.size.lower.bound.min</name>
655 <value>16777216</value>
657 If FlushLargeStoresPolicy is used and there are multiple column families,
658 then every time that we hit the total memstore limit, we find out all the
659 column families whose memstores exceed a "lower bound" and only flush them
660 while retaining the others in memory. The "lower bound" will be
661 "hbase.hregion.memstore.flush.size / column_family_number" by default
662 unless value of this property is larger than that. If none of the families
663 have their memstore size more than lower bound, all the memstores will be
664 flushed (just as usual).
668 <name>hbase.hregion.preclose.flush.size</name>
669 <value>5242880</value>
671 If the memstores in a region are this size or larger when we go
672 to close, run a "pre-flush" to clear out memstores before we put up
673 the region closed flag and take the region offline. On close,
674 a flush is run under the close flag to empty memory. During
675 this time the region is offline and we are not taking on any writes.
676 If the memstore content is large, this flush could take a long time to
677 complete. The preflush is meant to clean out the bulk of the memstore
678 before putting up the close flag and taking the region offline so the
679 flush that runs under the close flag has little to do.</description>
682 <name>hbase.hregion.memstore.block.multiplier</name>
685 Block updates if memstore has hbase.hregion.memstore.block.multiplier
686 times hbase.hregion.memstore.flush.size bytes. Useful preventing
687 runaway memstore during spikes in update traffic. Without an
688 upper-bound, memstore fills such that when it flushes the
689 resultant flush files take a long time to compact or split, or
690 worse, we OOME.</description>
693 <name>hbase.hregion.memstore.mslab.enabled</name>
696 Enables the MemStore-Local Allocation Buffer,
697 a feature which works to prevent heap fragmentation under
698 heavy write loads. This can reduce the frequency of stop-the-world
699 GC pauses on large heaps.</description>
702 <name>hbase.hregion.max.filesize</name>
703 <value>10737418240</value>
705 Maximum HFile size. If the sum of the sizes of a region's HFiles has grown to exceed this
706 value, the region is split in two.</description>
709 <name>hbase.hregion.majorcompaction</name>
710 <value>604800000</value>
711 <description>Time between major compactions, expressed in milliseconds. Set to 0 to disable
712 time-based automatic major compactions. User-requested and size-based major compactions will
713 still run. This value is multiplied by hbase.hregion.majorcompaction.jitter to cause
714 compaction to start at a somewhat-random time during a given window of time. The default value
715 is 7 days, expressed in milliseconds. If major compactions are causing disruption in your
716 environment, you can configure them to run at off-peak times for your deployment, or disable
717 time-based major compactions by setting this parameter to 0, and run major compactions in a
718 cron job or by another external mechanism.</description>
721 <name>hbase.hregion.majorcompaction.jitter</name>
723 <description>A multiplier applied to hbase.hregion.majorcompaction to cause compaction to occur
724 a given amount of time either side of hbase.hregion.majorcompaction. The smaller the number,
725 the closer the compactions will happen to the hbase.hregion.majorcompaction
726 interval.</description>
729 <name>hbase.hstore.compactionThreshold</name>
731 <description> If more than this number of StoreFiles exist in any one Store
732 (one StoreFile is written per flush of MemStore), a compaction is run to rewrite all
733 StoreFiles into a single StoreFile. Larger values delay compaction, but when compaction does
734 occur, it takes longer to complete.</description>
737 <name>hbase.regionserver.compaction.enabled</name>
739 <description>Enable/disable compactions on by setting true/false.
740 We can further switch compactions dynamically with the
741 compaction_switch shell command.</description>
744 <name>hbase.hstore.flusher.count</name>
746 <description> The number of flush threads. With fewer threads, the MemStore flushes will be
747 queued. With more threads, the flushes will be executed in parallel, increasing the load on
748 HDFS, and potentially causing more compactions. </description>
751 <name>hbase.hstore.blockingStoreFiles</name>
753 <description> If more than this number of StoreFiles exist in any one Store (one StoreFile
754 is written per flush of MemStore), updates are blocked for this region until a compaction is
755 completed, or until hbase.hstore.blockingWaitTime has been exceeded.</description>
758 <name>hbase.hstore.blockingWaitTime</name>
760 <description> The time for which a region will block updates after reaching the StoreFile limit
761 defined by hbase.hstore.blockingStoreFiles. After this time has elapsed, the region will stop
762 blocking updates even if a compaction has not been completed.</description>
765 <name>hbase.hstore.compaction.min</name>
767 <description>The minimum number of StoreFiles which must be eligible for compaction before
768 compaction can run. The goal of tuning hbase.hstore.compaction.min is to avoid ending up with
769 too many tiny StoreFiles to compact. Setting this value to 2 would cause a minor compaction
770 each time you have two StoreFiles in a Store, and this is probably not appropriate. If you
771 set this value too high, all the other values will need to be adjusted accordingly. For most
772 cases, the default value is appropriate. In previous versions of HBase, the parameter
773 hbase.hstore.compaction.min was named hbase.hstore.compactionThreshold.</description>
776 <name>hbase.hstore.compaction.max</name>
778 <description>The maximum number of StoreFiles which will be selected for a single minor
779 compaction, regardless of the number of eligible StoreFiles. Effectively, the value of
780 hbase.hstore.compaction.max controls the length of time it takes a single compaction to
781 complete. Setting it larger means that more StoreFiles are included in a compaction. For most
782 cases, the default value is appropriate.</description>
785 <name>hbase.hstore.compaction.min.size</name>
786 <value>134217728</value>
787 <description>A StoreFile (or a selection of StoreFiles, when using ExploringCompactionPolicy)
788 smaller than this size will always be eligible for minor compaction.
789 HFiles this size or larger are evaluated by hbase.hstore.compaction.ratio to determine if
790 they are eligible. Because this limit represents the "automatic include" limit for all
791 StoreFiles smaller than this value, this value may need to be reduced in write-heavy
792 environments where many StoreFiles in the 1-2 MB range are being flushed, because every
793 StoreFile will be targeted for compaction and the resulting StoreFiles may still be under the
794 minimum size and require further compaction. If this parameter is lowered, the ratio check is
795 triggered more quickly. This addressed some issues seen in earlier versions of HBase but
796 changing this parameter is no longer necessary in most situations. Default: 128 MB expressed
797 in bytes.</description>
800 <name>hbase.hstore.compaction.max.size</name>
801 <value>9223372036854775807</value>
802 <description>A StoreFile (or a selection of StoreFiles, when using ExploringCompactionPolicy)
803 larger than this size will be excluded from compaction. The effect of
804 raising hbase.hstore.compaction.max.size is fewer, larger StoreFiles that do not get
805 compacted often. If you feel that compaction is happening too often without much benefit, you
806 can try raising this value. Default: the value of LONG.MAX_VALUE, expressed in bytes.</description>
809 <name>hbase.hstore.compaction.ratio</name>
811 <description>For minor compaction, this ratio is used to determine whether a given StoreFile
812 which is larger than hbase.hstore.compaction.min.size is eligible for compaction. Its
813 effect is to limit compaction of large StoreFiles. The value of hbase.hstore.compaction.ratio
814 is expressed as a floating-point decimal. A large ratio, such as 10, will produce a single
815 giant StoreFile. Conversely, a low value, such as .25, will produce behavior similar to the
816 BigTable compaction algorithm, producing four StoreFiles. A moderate value of between 1.0 and
817 1.4 is recommended. When tuning this value, you are balancing write costs with read costs.
818 Raising the value (to something like 1.4) will have more write costs, because you will
819 compact larger StoreFiles. However, during reads, HBase will need to seek through fewer
820 StoreFiles to accomplish the read. Consider this approach if you cannot take advantage of
821 Bloom filters. Otherwise, you can lower this value to something like 1.0 to reduce the
822 background cost of writes, and use Bloom filters to control the number of StoreFiles touched
823 during reads. For most cases, the default value is appropriate.</description>
826 <name>hbase.hstore.compaction.ratio.offpeak</name>
828 <description>Allows you to set a different (by default, more aggressive) ratio for determining
829 whether larger StoreFiles are included in compactions during off-peak hours. Works in the
830 same way as hbase.hstore.compaction.ratio. Only applies if hbase.offpeak.start.hour and
831 hbase.offpeak.end.hour are also enabled.</description>
834 <name>hbase.hstore.time.to.purge.deletes</name>
836 <description>The amount of time to delay purging of delete markers with future timestamps. If
837 unset, or set to 0, all delete markers, including those with future timestamps, are purged
838 during the next major compaction. Otherwise, a delete marker is kept until the major compaction
839 which occurs after the marker's timestamp plus the value of this setting, in milliseconds.
843 <name>hbase.offpeak.start.hour</name>
845 <description>The start of off-peak hours, expressed as an integer between 0 and 23, inclusive.
846 Set to -1 to disable off-peak.</description>
849 <name>hbase.offpeak.end.hour</name>
851 <description>The end of off-peak hours, expressed as an integer between 0 and 23, inclusive. Set
852 to -1 to disable off-peak.</description>
855 <name>hbase.regionserver.thread.compaction.throttle</name>
856 <value>2684354560</value>
857 <description>There are two different thread pools for compactions, one for large compactions and
858 the other for small compactions. This helps to keep compaction of lean tables (such as
859 hbase:meta) fast. If a compaction is larger than this threshold, it
860 goes into the large compaction pool. In most cases, the default value is appropriate. Default:
861 2 x hbase.hstore.compaction.max x hbase.hregion.memstore.flush.size (which defaults to 128MB).
862 The value field assumes that the value of hbase.hregion.memstore.flush.size is unchanged from
863 the default.</description>
866 <name>hbase.regionserver.majorcompaction.pagecache.drop</name>
868 <description>Specifies whether to drop pages read/written into the system page cache by
869 major compactions. Setting it to true helps prevent major compactions from
870 polluting the page cache, which is almost always required, especially for clusters
871 with low/moderate memory to storage ratio.</description>
874 <name>hbase.regionserver.minorcompaction.pagecache.drop</name>
876 <description>Specifies whether to drop pages read/written into the system page cache by
877 minor compactions. Setting it to true helps prevent minor compactions from
878 polluting the page cache, which is most beneficial on clusters with low
879 memory to storage ratio or very write heavy clusters. You may want to set it to
880 false under moderate to low write workload when bulk of the reads are
881 on the most recently written data.</description>
884 <name>hbase.hstore.compaction.kv.max</name>
886 <description>The maximum number of KeyValues to read and then write in a batch when flushing or
887 compacting. Set this lower if you have big KeyValues and problems with Out Of Memory
888 Exceptions Set this higher if you have wide, small rows. </description>
891 <name>hbase.storescanner.parallel.seek.enable</name>
894 Enables StoreFileScanner parallel-seeking in StoreScanner,
895 a feature which can reduce response latency under special conditions.</description>
898 <name>hbase.storescanner.parallel.seek.threads</name>
901 The default thread pool size if parallel-seeking feature enabled.</description>
904 <name>hfile.block.cache.size</name>
906 <description>Percentage of maximum heap (-Xmx setting) to allocate to block cache
907 used by a StoreFile. Default of 0.4 means allocate 40%.
908 Set to 0 to disable but it's not recommended; you need at least
909 enough cache to hold the storefile indices.</description>
912 <name>hfile.block.index.cacheonwrite</name>
914 <description>This allows to put non-root multi-level index blocks into the block
915 cache at the time the index is being written.</description>
918 <name>hfile.index.block.max.size</name>
919 <value>131072</value>
920 <description>When the size of a leaf-level, intermediate-level, or root-level
921 index block in a multi-level block index grows to this size, the
922 block is written out and a new block is started.</description>
925 <name>hbase.bucketcache.ioengine</name>
927 <description>Where to store the contents of the bucketcache. One of: offheap,
928 file, files or mmap. If a file or files, set it to file(s):PATH_TO_FILE.
929 mmap means the content will be in an mmaped file. Use mmap:PATH_TO_FILE.
930 See http://hbase.apache.org/book.html#offheap.blockcache for more information.
934 <name>hbase.bucketcache.size</name>
936 <description>A float that EITHER represents a percentage of total heap memory
937 size to give to the cache (if < 1.0) OR, it is the total capacity in
938 megabytes of BucketCache. Default: 0.0</description>
941 <name>hbase.bucketcache.bucket.sizes</name>
943 <description>A comma-separated list of sizes for buckets for the bucketcache.
944 Can be multiple sizes. List block sizes in order from smallest to largest.
945 The sizes you use will depend on your data access patterns.
946 Must be a multiple of 256 else you will run into
947 'java.io.IOException: Invalid HFile block magic' when you go to read from cache.
948 If you specify no values here, then you pick up the default bucketsizes set
949 in code (See BucketAllocator#DEFAULT_BUCKET_SIZES).
953 <name>hfile.format.version</name>
955 <description>The HFile format version to use for new files.
956 Version 3 adds support for tags in hfiles (See http://hbase.apache.org/book.html#hbase.tags).
957 Also see the configuration 'hbase.replication.rpc.codec'.
961 <name>hfile.block.bloom.cacheonwrite</name>
963 <description>Enables cache-on-write for inline blocks of a compound Bloom filter.</description>
966 <name>io.storefile.bloom.block.size</name>
967 <value>131072</value>
968 <description>The size in bytes of a single block ("chunk") of a compound Bloom
969 filter. This size is approximate, because Bloom blocks can only be
970 inserted at data block boundaries, and the number of keys per data
971 block varies.</description>
974 <name>hbase.rs.cacheblocksonwrite</name>
976 <description>Whether an HFile block should be added to the block cache when the
977 block is finished.</description>
980 <name>hbase.rpc.timeout</name>
982 <description>This is for the RPC layer to define how long (millisecond) HBase client applications
983 take for a remote call to time out. It uses pings to check connections
984 but will eventually throw a TimeoutException.</description>
987 <name>hbase.client.operation.timeout</name>
988 <value>1200000</value>
989 <description>Operation timeout is a top-level restriction (millisecond) that makes sure a
990 blocking operation in Table will not be blocked more than this. In each operation, if rpc
991 request fails because of timeout or other reason, it will retry until success or throw
992 RetriesExhaustedException. But if the total time being blocking reach the operation timeout
993 before retries exhausted, it will break early and throw SocketTimeoutException.</description>
996 <name>hbase.cells.scanned.per.heartbeat.check</name>
998 <description>The number of cells scanned in between heartbeat checks. Heartbeat
999 checks occur during the processing of scans to determine whether or not the
1000 server should stop scanning in order to send back a heartbeat message to the
1001 client. Heartbeat messages are used to keep the client-server connection alive
1002 during long running scans. Small values mean that the heartbeat checks will
1003 occur more often and thus will provide a tighter bound on the execution time of
1004 the scan. Larger values mean that the heartbeat checks occur less frequently
1008 <name>hbase.rpc.shortoperation.timeout</name>
1009 <value>10000</value>
1010 <description>This is another version of "hbase.rpc.timeout". For those RPC operation
1011 within cluster, we rely on this configuration to set a short timeout limitation
1012 for short operation. For example, short rpc timeout for region server's trying
1013 to report to active master can benefit quicker master failover process.</description>
1016 <name>hbase.ipc.client.tcpnodelay</name>
1018 <description>Set no delay on rpc socket connections. See
1019 http://docs.oracle.com/javase/1.5.0/docs/api/java/net/Socket.html#getTcpNoDelay()</description>
1022 <name>hbase.regionserver.hostname</name>
1024 <description>This config is for experts: don't set its value unless you really know what you are doing.
1025 When set to a non-empty value, this represents the (external facing) hostname for the underlying server.
1026 See https://issues.apache.org/jira/browse/HBASE-12954 for details.</description>
1029 <name>hbase.regionserver.hostname.disable.master.reversedns</name>
1030 <value>false</value>
1031 <description>This config is for experts: don't set its value unless you really know what you are doing.
1032 When set to true, regionserver will use the current node hostname for the servername and HMaster will
1033 skip reverse DNS lookup and use the hostname sent by regionserver instead. Note that this config and
1034 hbase.regionserver.hostname are mutually exclusive. See https://issues.apache.org/jira/browse/HBASE-18226
1035 for more details.</description>
1037 <!-- The following properties configure authentication information for
1038 HBase processes when using Kerberos security. There are no default
1039 values, included here for documentation purposes -->
1041 <name>hbase.master.keytab.file</name>
1043 <description>Full path to the kerberos keytab file to use for logging in
1044 the configured HMaster server principal.</description>
1047 <name>hbase.master.kerberos.principal</name>
1049 <description>Ex. "hbase/_HOST@EXAMPLE.COM". The kerberos principal name
1050 that should be used to run the HMaster process. The principal name should
1051 be in the form: user/hostname@DOMAIN. If "_HOST" is used as the hostname
1052 portion, it will be replaced with the actual hostname of the running
1053 instance.</description>
1056 <name>hbase.regionserver.keytab.file</name>
1058 <description>Full path to the kerberos keytab file to use for logging in
1059 the configured HRegionServer server principal.</description>
1062 <name>hbase.regionserver.kerberos.principal</name>
1064 <description>Ex. "hbase/_HOST@EXAMPLE.COM". The kerberos principal name
1065 that should be used to run the HRegionServer process. The principal name
1066 should be in the form: user/hostname@DOMAIN. If "_HOST" is used as the
1067 hostname portion, it will be replaced with the actual hostname of the
1068 running instance. An entry for this principal must exist in the file
1069 specified in hbase.regionserver.keytab.file</description>
1071 <!-- Additional configuration specific to HBase security -->
1073 <name>hadoop.policy.file</name>
1074 <value>hbase-policy.xml</value>
1075 <description>The policy configuration file used by RPC servers to make
1076 authorization decisions on client requests. Only used when HBase
1077 security is enabled.</description>
1080 <name>hbase.superuser</name>
1082 <description>List of users or groups (comma-separated), who are allowed
1083 full privileges, regardless of stored ACLs, across the cluster.
1084 Only used when HBase security is enabled.</description>
1087 <name>hbase.auth.key.update.interval</name>
1088 <value>86400000</value>
1089 <description>The update interval for master key for authentication tokens
1090 in servers in milliseconds. Only used when HBase security is enabled.</description>
1093 <name>hbase.auth.token.max.lifetime</name>
1094 <value>604800000</value>
1095 <description>The maximum lifetime in milliseconds after which an
1096 authentication token expires. Only used when HBase security is enabled.</description>
1099 <name>hbase.ipc.client.fallback-to-simple-auth-allowed</name>
1100 <value>false</value>
1101 <description>When a client is configured to attempt a secure connection, but attempts to
1102 connect to an insecure server, that server may instruct the client to
1103 switch to SASL SIMPLE (unsecure) authentication. This setting controls
1104 whether or not the client will accept this instruction from the server.
1105 When false (the default), the client will not allow the fallback to SIMPLE
1106 authentication, and will abort the connection.</description>
1109 <name>hbase.ipc.server.fallback-to-simple-auth-allowed</name>
1110 <value>false</value>
1111 <description>When a server is configured to require secure connections, it will
1112 reject connection attempts from clients using SASL SIMPLE (unsecure) authentication.
1113 This setting allows secure servers to accept SASL SIMPLE connections from clients
1114 when the client requests. When false (the default), the server will not allow the fallback
1115 to SIMPLE authentication, and will reject the connection. WARNING: This setting should ONLY
1116 be used as a temporary measure while converting clients over to secure authentication. It
1117 MUST BE DISABLED for secure operation.</description>
1120 <name>hbase.display.keys</name>
1122 <description>When this is set to true the webUI and such will display all start/end keys
1123 as part of the table details, region names, etc. When this is set to false,
1124 the keys are hidden.</description>
1127 <name>hbase.coprocessor.enabled</name>
1129 <description>Enables or disables coprocessor loading. If 'false'
1130 (disabled), any other coprocessor related configuration will be ignored.
1134 <name>hbase.coprocessor.user.enabled</name>
1136 <description>Enables or disables user (aka. table) coprocessor loading.
1137 If 'false' (disabled), any table coprocessor attributes in table
1138 descriptors will be ignored. If "hbase.coprocessor.enabled" is 'false'
1139 this setting has no effect.
1143 <name>hbase.coprocessor.region.classes</name>
1145 <description>A comma-separated list of region observer or endpoint coprocessors
1146 that are loaded by default on all tables. For any override coprocessor method,
1147 these classes will be called in order. After implementing your own Coprocessor,
1148 add it to HBase's classpath and add the fully qualified class name here.
1149 A coprocessor can also be loaded on demand by setting HTableDescriptor or the
1150 HBase shell.</description>
1153 <name>hbase.coprocessor.master.classes</name>
1155 <description>A comma-separated list of
1156 org.apache.hadoop.hbase.coprocessor.MasterObserver coprocessors that are
1157 loaded by default on the active HMaster process. For any implemented
1158 coprocessor methods, the listed classes will be called in order. After
1159 implementing your own MasterObserver, just put it in HBase's classpath
1160 and add the fully qualified class name here.</description>
1163 <name>hbase.coprocessor.abortonerror</name>
1165 <description>Set to true to cause the hosting server (master or regionserver)
1166 to abort if a coprocessor fails to load, fails to initialize, or throws an
1167 unexpected Throwable object. Setting this to false will allow the server to
1168 continue execution but the system wide state of the coprocessor in question
1169 will become inconsistent as it will be properly executing in only a subset
1170 of servers, so this is most useful for debugging only.</description>
1173 <name>hbase.rest.port</name>
1175 <description>The port for the HBase REST server.</description>
1178 <name>hbase.rest.readonly</name>
1179 <value>false</value>
1180 <description>Defines the mode the REST server will be started in. Possible values are:
1181 false: All HTTP methods are permitted - GET/PUT/POST/DELETE.
1182 true: Only the GET method is permitted.</description>
1185 <name>hbase.rest.threads.max</name>
1187 <description>The maximum number of threads of the REST server thread pool.
1188 Threads in the pool are reused to process REST requests. This
1189 controls the maximum number of requests processed concurrently.
1190 It may help to control the memory used by the REST server to
1191 avoid OOM issues. If the thread pool is full, incoming requests
1192 will be queued up and wait for some free threads.</description>
1195 <name>hbase.rest.threads.min</name>
1197 <description>The minimum number of threads of the REST server thread pool.
1198 The thread pool always has at least these number of threads so
1199 the REST server is ready to serve incoming requests.</description>
1202 <name>hbase.rest.support.proxyuser</name>
1203 <value>false</value>
1204 <description>Enables running the REST server to support proxy-user mode.</description>
1206 <property skipInDoc="true">
1207 <name>hbase.defaults.for.version</name>
1208 <value>@@@VERSION@@@</value>
1209 <description>This defaults file was compiled for version ${project.version}. This variable is used
1210 to make sure that a user doesn't have an old version of hbase-default.xml on the
1211 classpath.</description>
1214 <name>hbase.defaults.for.version.skip</name>
1215 <value>false</value>
1216 <description>Set to true to skip the 'hbase.defaults.for.version' check.
1217 Setting this to true can be useful in contexts other than
1218 the other side of a maven generation; i.e. running in an
1219 IDE. You'll want to set this boolean to true to avoid
1220 seeing the RuntimeException complaint: "hbase-default.xml file
1221 seems to be for and old version of HBase (\${hbase.version}), this
1222 version is X.X.X-SNAPSHOT"</description>
1225 <name>hbase.table.lock.enable</name>
1227 <description>Set to true to enable locking the table in zookeeper for schema change operations.
1228 Table locking from master prevents concurrent schema modifications to corrupt table
1229 state.</description>
1232 <name>hbase.table.max.rowsize</name>
1233 <value>1073741824</value>
1235 Maximum size of single row in bytes (default is 1 Gb) for Get'ting
1236 or Scan'ning without in-row scan flag set. If row size exceeds this limit
1237 RowTooBigException is thrown to client.
1241 <name>hbase.thrift.minWorkerThreads</name>
1243 <description>The "core size" of the thread pool. New threads are created on every
1244 connection until this many threads are created.</description>
1247 <name>hbase.thrift.maxWorkerThreads</name>
1249 <description>The maximum size of the thread pool. When the pending request queue
1250 overflows, new threads are created until their number reaches this number.
1251 After that, the server starts dropping connections.</description>
1254 <name>hbase.thrift.maxQueuedRequests</name>
1256 <description>The maximum number of pending Thrift connections waiting in the queue. If
1257 there are no idle threads in the pool, the server queues requests. Only
1258 when the queue overflows, new threads are added, up to
1259 hbase.thrift.maxQueuedRequests threads.</description>
1262 <name>hbase.regionserver.thrift.framed</name>
1263 <value>false</value>
1264 <description>Use Thrift TFramedTransport on the server side.
1265 This is the recommended transport for thrift servers and requires a similar setting
1266 on the client side. Changing this to false will select the default transport,
1267 vulnerable to DoS when malformed requests are issued due to THRIFT-601.
1271 <name>hbase.regionserver.thrift.framed.max_frame_size_in_mb</name>
1273 <description>Default frame size when using framed transport, in MB</description>
1276 <name>hbase.regionserver.thrift.compact</name>
1277 <value>false</value>
1278 <description>Use Thrift TCompactProtocol binary serialization protocol.</description>
1281 <name>hbase.rootdir.perms</name>
1283 <description>FS Permissions for the root data subdirectory in a secure (kerberos) setup.
1284 When master starts, it creates the rootdir with this permissions or sets the permissions
1285 if it does not match.</description>
1288 <name>hbase.wal.dir.perms</name>
1290 <description>FS Permissions for the root WAL directory in a secure(kerberos) setup.
1291 When master starts, it creates the WAL dir with this permissions or sets the permissions
1292 if it does not match.</description>
1295 <name>hbase.data.umask.enable</name>
1296 <value>false</value>
1297 <description>Enable, if true, that file permissions should be assigned
1298 to the files written by the regionserver</description>
1301 <name>hbase.data.umask</name>
1303 <description>File permissions that should be used to write data
1304 files when hbase.data.umask.enable is true</description>
1307 <name>hbase.snapshot.enabled</name>
1309 <description>Set to true to allow snapshots to be taken / restored / cloned.</description>
1312 <name>hbase.snapshot.restore.take.failsafe.snapshot</name>
1314 <description>Set to true to take a snapshot before the restore operation.
1315 The snapshot taken will be used in case of failure, to restore the previous state.
1316 At the end of the restore operation this snapshot will be deleted</description>
1319 <name>hbase.snapshot.restore.failsafe.name</name>
1320 <value>hbase-failsafe-{snapshot.name}-{restore.timestamp}</value>
1321 <description>Name of the failsafe snapshot taken by the restore operation.
1322 You can use the {snapshot.name}, {table.name} and {restore.timestamp} variables
1323 to create a name based on what you are restoring.</description>
1326 <name>hbase.snapshot.working.dir</name>
1328 <description>Location where the snapshotting process will occur. The location of the
1329 completed snapshots will not change, but the temporary directory where the snapshot
1330 process occurs will be set to this location. This can be a separate filesystem than
1331 the root directory, for performance increase purposes. See HBASE-21098 for more
1332 information</description>
1335 <name>hbase.server.compactchecker.interval.multiplier</name>
1337 <description>The number that determines how often we scan to see if compaction is necessary.
1338 Normally, compactions are done after some events (such as memstore flush), but if
1339 region didn't receive a lot of writes for some time, or due to different compaction
1340 policies, it may be necessary to check it periodically. The interval between checks is
1341 hbase.server.compactchecker.interval.multiplier multiplied by
1342 hbase.server.thread.wakefrequency.</description>
1345 <name>hbase.lease.recovery.timeout</name>
1346 <value>900000</value>
1347 <description>How long we wait on dfs lease recovery in total before giving up.</description>
1350 <name>hbase.lease.recovery.dfs.timeout</name>
1351 <value>64000</value>
1352 <description>How long between dfs recover lease invocations. Should be larger than the sum of
1353 the time it takes for the namenode to issue a block recovery command as part of
1354 datanode; dfs.heartbeat.interval and the time it takes for the primary
1355 datanode, performing block recovery to timeout on a dead datanode; usually
1356 dfs.client.socket-timeout. See the end of HBASE-8389 for more.</description>
1359 <name>hbase.column.max.version</name>
1361 <description>New column family descriptors will use this value as the default number of versions
1362 to keep.</description>
1365 <name>dfs.client.read.shortcircuit</name>
1366 <value>false</value>
1368 If set to true, this configuration parameter enables short-circuit local
1373 <name>dfs.domain.socket.path</name>
1376 This is a path to a UNIX domain socket that will be used for
1377 communication between the DataNode and local HDFS clients, if
1378 dfs.client.read.shortcircuit is set to true. If the string "_PORT" is
1379 present in this path, it will be replaced by the TCP port of the DataNode.
1380 Be careful about permissions for the directory that hosts the shared
1381 domain socket; dfsclient will complain if open to other users than the HBase user.
1385 <name>hbase.dfs.client.read.shortcircuit.buffer.size</name>
1386 <value>131072</value>
1387 <description>If the DFSClient configuration
1388 dfs.client.read.shortcircuit.buffer.size is unset, we will
1389 use what is configured here as the short circuit read default
1390 direct byte buffer size. DFSClient native default is 1MB; HBase
1391 keeps its HDFS files open so number of file blocks * 1MB soon
1392 starts to add up and threaten OOME because of a shortage of
1393 direct memory. So, we set it down from the default. Make
1394 it > the default hbase block size set in the HColumnDescriptor
1395 which is usually 64k.
1399 <name>hbase.regionserver.checksum.verify</name>
1402 If set to true (the default), HBase verifies the checksums for hfile
1403 blocks. HBase writes checksums inline with the data when it writes out
1404 hfiles. HDFS (as of this writing) writes checksums to a separate file
1405 than the data file necessitating extra seeks. Setting this flag saves
1406 some on i/o. Checksum verification by HDFS will be internally disabled
1407 on hfile streams when this flag is set. If the hbase-checksum verification
1408 fails, we will switch back to using HDFS checksums (so do not disable HDFS
1409 checksums! And besides this feature applies to hfiles only, not to WALs).
1410 If this parameter is set to false, then hbase will not verify any checksums,
1411 instead it will depend on checksum verification being done in the HDFS client.
1415 <name>hbase.hstore.bytes.per.checksum</name>
1416 <value>16384</value>
1418 Number of bytes in a newly created checksum chunk for HBase-level
1419 checksums in hfile blocks.
1423 <name>hbase.hstore.checksum.algorithm</name>
1424 <value>CRC32C</value>
1426 Name of an algorithm that is used to compute checksums. Possible values
1427 are NULL, CRC32, CRC32C.
1431 <name>hbase.client.scanner.max.result.size</name>
1432 <value>2097152</value>
1433 <description>Maximum number of bytes returned when calling a scanner's next method.
1434 Note that when a single row is larger than this limit the row is still returned completely.
1435 The default value is 2MB, which is good for 1ge networks.
1436 With faster and/or high latency networks this value should be increased.
1440 <name>hbase.server.scanner.max.result.size</name>
1441 <value>104857600</value>
1442 <description>Maximum number of bytes returned when calling a scanner's next method.
1443 Note that when a single row is larger than this limit the row is still returned completely.
1444 The default value is 100MB.
1445 This is a safety setting to protect the server from OOM situations.
1449 <name>hbase.status.published</name>
1450 <value>false</value>
1452 This setting activates the publication by the master of the status of the region server.
1453 When a region server dies and its recovery starts, the master will push this information
1454 to the client application, to let them cut the connection immediately instead of waiting
1459 <name>hbase.status.publisher.class</name>
1460 <value>org.apache.hadoop.hbase.master.ClusterStatusPublisher$MulticastPublisher</value>
1462 Implementation of the status publication with a multicast message.
1466 <name>hbase.status.listener.class</name>
1467 <value>org.apache.hadoop.hbase.client.ClusterStatusListener$MulticastListener</value>
1469 Implementation of the status listener with a multicast message.
1473 <name>hbase.status.multicast.address.ip</name>
1474 <value>226.1.1.3</value>
1476 Multicast address to use for the status publication by multicast.
1480 <name>hbase.status.multicast.address.port</name>
1481 <value>16100</value>
1483 Multicast port to use for the status publication by multicast.
1487 <name>hbase.dynamic.jars.dir</name>
1488 <value>${hbase.rootdir}/lib</value>
1490 The directory from which the custom filter JARs can be loaded
1491 dynamically by the region server without the need to restart. However,
1492 an already loaded filter/co-processor class would not be un-loaded. See
1493 HBASE-1936 for more details.
1495 Does not apply to coprocessors.
1499 <name>hbase.security.authentication</name>
1500 <value>simple</value>
1502 Controls whether or not secure authentication is enabled for HBase.
1503 Possible values are 'simple' (no authentication), and 'kerberos'.
1507 <name>hbase.rest.filter.classes</name>
1508 <value>org.apache.hadoop.hbase.rest.filter.GzipFilter</value>
1510 Servlet filters for REST service.
1514 <name>hbase.master.loadbalancer.class</name>
1515 <value>org.apache.hadoop.hbase.master.balancer.StochasticLoadBalancer</value>
1517 Class used to execute the regions balancing when the period occurs.
1518 See the class comment for more on how it works
1519 http://hbase.apache.org/devapidocs/org/apache/hadoop/hbase/master/balancer/StochasticLoadBalancer.html
1520 It replaces the DefaultLoadBalancer as the default (since renamed
1521 as the SimpleLoadBalancer).
1525 <name>hbase.master.loadbalance.bytable</name>
1526 <value>false</value>
1527 <description>Factor Table name when the balancer runs.
1532 <name>hbase.master.normalizer.class</name>
1533 <value>org.apache.hadoop.hbase.master.normalizer.SimpleRegionNormalizer</value>
1535 Class used to execute the region normalization when the period occurs.
1536 See the class comment for more on how it works
1537 http://hbase.apache.org/devapidocs/org/apache/hadoop/hbase/master/normalizer/SimpleRegionNormalizer.html
1541 <name>hbase.rest.csrf.enabled</name>
1542 <value>false</value>
1544 Set to true to enable protection against cross-site request forgery (CSRF)
1548 <name>hbase.rest-csrf.browser-useragents-regex</name>
1549 <value>^Mozilla.*,^Opera.*</value>
1551 A comma-separated list of regular expressions used to match against an HTTP
1552 request's User-Agent header when protection against cross-site request
1553 forgery (CSRF) is enabled for REST server by setting
1554 hbase.rest.csrf.enabled to true. If the incoming User-Agent matches
1555 any of these regular expressions, then the request is considered to be sent
1556 by a browser, and therefore CSRF prevention is enforced. If the request's
1557 User-Agent does not match any of these regular expressions, then the request
1558 is considered to be sent by something other than a browser, such as scripted
1559 automation. In this case, CSRF is not a potential attack vector, so
1560 the prevention is not enforced. This helps achieve backwards-compatibility
1561 with existing automation that has not been updated to send the CSRF
1566 <name>hbase.security.exec.permission.checks</name>
1567 <value>false</value>
1569 If this setting is enabled and ACL based access control is active (the
1570 AccessController coprocessor is installed either as a system coprocessor
1571 or on a table as a table coprocessor) then you must grant all relevant
1572 users EXEC privilege if they require the ability to execute coprocessor
1573 endpoint calls. EXEC privilege, like any other permission, can be
1574 granted globally to a user, or to a user on a per table or per namespace
1575 basis. For more information on coprocessor endpoints, see the coprocessor
1576 section of the HBase online manual. For more information on granting or
1577 revoking permissions using the AccessController, see the security
1578 section of the HBase online manual.
1582 <name>hbase.procedure.regionserver.classes</name>
1584 <description>A comma-separated list of
1585 org.apache.hadoop.hbase.procedure.RegionServerProcedureManager procedure managers that are
1586 loaded by default on the active HRegionServer process. The lifecycle methods (init/start/stop)
1587 will be called by the active HRegionServer process to perform the specific globally barriered
1588 procedure. After implementing your own RegionServerProcedureManager, just put it in
1589 HBase's classpath and add the fully qualified class name here.
1593 <name>hbase.procedure.master.classes</name>
1595 <description>A comma-separated list of
1596 org.apache.hadoop.hbase.procedure.MasterProcedureManager procedure managers that are
1597 loaded by default on the active HMaster process. A procedure is identified by its signature and
1598 users can use the signature and an instant name to trigger an execution of a globally barriered
1599 procedure. After implementing your own MasterProcedureManager, just put it in HBase's classpath
1600 and add the fully qualified class name here.</description>
1603 <name>hbase.coordinated.state.manager.class</name>
1604 <value>org.apache.hadoop.hbase.coordination.ZkCoordinatedStateManager</value>
1605 <description>Fully qualified name of class implementing coordinated state manager.</description>
1608 <name>hbase.regionserver.storefile.refresh.period</name>
1611 The period (in milliseconds) for refreshing the store files for the secondary regions. 0
1612 means this feature is disabled. Secondary regions sees new files (from flushes and
1613 compactions) from primary once the secondary region refreshes the list of files in the
1614 region (there is no notification mechanism). But too frequent refreshes might cause
1615 extra Namenode pressure. If the files cannot be refreshed for longer than HFile TTL
1616 (hbase.master.hfilecleaner.ttl) the requests are rejected. Configuring HFile TTL to a larger
1617 value is also recommended with this setting.
1621 <name>hbase.region.replica.replication.enabled</name>
1622 <value>false</value>
1624 Whether asynchronous WAL replication to the secondary region replicas is enabled or not.
1625 If this is enabled, a replication peer named "region_replica_replication" will be created
1626 which will tail the logs and replicate the mutations to region replicas for tables that
1627 have region replication > 1. If this is enabled once, disabling this replication also
1628 requires disabling the replication peer using shell or Admin java class.
1629 Replication to secondary region replicas works over standard inter-cluster replication.
1633 <name>hbase.http.filter.initializers</name>
1634 <value>org.apache.hadoop.hbase.http.lib.StaticUserWebFilter</value>
1636 A comma separated list of class names. Each class in the list must extend
1637 org.apache.hadoop.hbase.http.FilterInitializer. The corresponding Filter will
1638 be initialized. Then, the Filter will be applied to all user facing jsp
1639 and servlet web pages.
1640 The ordering of the list defines the ordering of the filters.
1641 The default StaticUserWebFilter add a user principal as defined by the
1642 hbase.http.staticuser.user property.
1646 <name>hbase.security.visibility.mutations.checkauths</name>
1647 <value>false</value>
1649 This property if enabled, will check whether the labels in the visibility
1650 expression are associated with the user issuing the mutation
1654 <name>hbase.http.max.threads</name>
1657 The maximum number of threads that the HTTP Server will create in its
1662 <name>hbase.replication.rpc.codec</name>
1663 <value>org.apache.hadoop.hbase.codec.KeyValueCodecWithTags</value>
1665 The codec that is to be used when replication is enabled so that
1666 the tags are also replicated. This is used along with HFileV3 which
1667 supports tags in them. If tags are not used or if the hfile version used
1668 is HFileV2 then KeyValueCodec can be used as the replication codec. Note that
1669 using KeyValueCodecWithTags for replication when there are no tags causes no harm.
1673 <name>hbase.replication.source.maxthreads</name>
1676 The maximum number of threads any replication source will use for
1677 shipping edits to the sinks in parallel. This also limits the number of
1678 chunks each replication batch is broken into. Larger values can improve
1679 the replication throughput between the master and slave clusters. The
1680 default of 10 will rarely need to be changed.
1683 <!-- Static Web User Filter properties. -->
1685 <name>hbase.http.staticuser.user</name>
1686 <value>dr.stack</value>
1688 The user name to filter as, on static web filters
1689 while rendering content. An example use is the HDFS
1690 web UI (user to be used for browsing files).
1694 <name>hbase.regionserver.handler.abort.on.error.percent</name>
1696 <description>The percent of region server RPC threads failed to abort RS.
1697 -1 Disable aborting; 0 Abort if even a single handler has died;
1698 0.x Abort only when this percent of handlers have died;
1699 1 Abort only all of the handers have died.</description>
1701 <!-- Mob properties. -->
1703 <name>hbase.mob.file.cache.size</name>
1706 Number of opened file handlers to cache.
1707 A larger value will benefit reads by providing more file handlers per mob
1708 file cache and would reduce frequent file opening and closing.
1709 However, if this is set too high, this could lead to a "too many opened file handlers"
1710 The default value is 1000.
1714 <name>hbase.mob.cache.evict.period</name>
1717 The amount of time in seconds before the mob cache evicts cached mob files.
1718 The default value is 3600 seconds.
1722 <name>hbase.mob.cache.evict.remain.ratio</name>
1725 The ratio (between 0.0 and 1.0) of files that remains cached after an eviction
1726 is triggered when the number of cached mob files exceeds the hbase.mob.file.cache.size.
1727 The default value is 0.5f.
1731 <name>hbase.master.mob.ttl.cleaner.period</name>
1732 <value>86400</value>
1734 The period that ExpiredMobFileCleanerChore runs. The unit is second.
1735 The default value is one day. The MOB file name uses only the date part of
1736 the file creation time in it. We use this time for deciding TTL expiry of
1737 the files. So the removal of TTL expired files might be delayed. The max
1738 delay might be 24 hrs.
1742 <name>hbase.mob.compaction.mergeable.threshold</name>
1743 <value>1342177280</value>
1745 If the size of a mob file is less than this value, it's regarded as a small
1746 file and needs to be merged in mob compaction. The default value is 1280MB.
1750 <name>hbase.mob.delfile.max.count</name>
1753 The max number of del files that is allowed in the mob compaction.
1754 In the mob compaction, when the number of existing del files is larger than
1755 this value, they are merged until number of del files is not larger this value.
1756 The default value is 3.
1760 <name>hbase.mob.compaction.batch.size</name>
1763 The max number of the mob files that is allowed in a batch of the mob compaction.
1764 The mob compaction merges the small mob files to bigger ones. If the number of the
1765 small files is very large, it could lead to a "too many opened file handlers" in the merge.
1766 And the merge has to be split into batches. This value limits the number of mob files
1767 that are selected in a batch of the mob compaction. The default value is 100.
1771 <name>hbase.mob.compaction.chore.period</name>
1772 <value>604800</value>
1774 The period that MobCompactionChore runs. The unit is second.
1775 The default value is one week.
1779 <name>hbase.mob.compactor.class</name>
1780 <value>org.apache.hadoop.hbase.mob.compactions.PartitionedMobCompactor</value>
1782 Implementation of mob compactor, the default one is PartitionedMobCompactor.
1786 <name>hbase.mob.compaction.threads.max</name>
1789 The max number of threads used in MobCompactor.
1793 <name>hbase.snapshot.master.timeout.millis</name>
1794 <value>300000</value>
1796 Timeout for master for the snapshot procedure execution.
1800 <name>hbase.snapshot.region.timeout</name>
1801 <value>300000</value>
1803 Timeout for regionservers to keep threads in snapshot request pool waiting.
1807 <name>hbase.rpc.rows.warning.threshold</name>
1810 Number of rows in a batch operation above which a warning will be logged.
1814 <name>hbase.master.wait.on.service.seconds</name>
1816 <description>Default is 5 minutes. Make it 30 seconds for tests. See
1817 HBASE-19794 for some context.</description>