diff --git a/docs/configuration/index.md b/docs/configuration/index.md index aa1cd930617..43d1a14903e 100644 --- a/docs/configuration/index.md +++ b/docs/configuration/index.md @@ -841,7 +841,7 @@ These Overlord static configurations can be defined in the `overlord/runtime.pro |Property|Description|Default| |--------|-----------|-------| -|`druid.indexer.runner.type`|Choices "local" or "remote". Indicates whether tasks should be run locally or in a distributed environment. Experimental task runner "httpRemote" is also available which is same as "remote" but uses HTTP to interact with Middle Manaters instead of Zookeeper.|local| +|`druid.indexer.runner.type`|Choices "local" or "remote". Indicates whether tasks should be run locally or in a distributed environment. Experimental task runner "httpRemote" is also available which is same as "remote" but uses HTTP to interact with Middle Managers instead of Zookeeper.|local| |`druid.indexer.storage.type`|Choices are "local" or "metadata". Indicates whether incoming tasks should be stored locally (in heap) or in metadata storage. Storing incoming tasks in metadata storage allows for tasks to be resumed if the Overlord should fail.|local| |`druid.indexer.storage.recentlyFinishedThreshold`|A duration of time to store task results.|PT24H| |`druid.indexer.tasklock.forceTimeChunkLock`|_**Setting this to false is still experimental**_
If set, all tasks are enforced to use time chunk lock. If not set, each task automatically chooses a lock type to use. This configuration can be overwritten by setting `forceTimeChunkLock` in the [task context](../ingestion/tasks.md#context). See [Task Locking & Priority](../ingestion/tasks.md#context) for more details about locking in tasks.|true| @@ -1049,6 +1049,8 @@ Amazon's EC2 is currently the only supported autoscaler. This section contains the configuration options for the processes that reside on Data servers (MiddleManagers/Peons and Historicals) in the suggested [three-server configuration](../design/processes.html#server-types). +Configuration options for the experimental [Indexer process](../design/indexer.md) are also provided here. + ### MiddleManager and Peons These MiddleManager and Peon configurations can be defined in the `middleManager/runtime.properties` file. @@ -1173,6 +1175,89 @@ then the value from the configuration below is used: |--------|-----------|-------| |`druid.peon.defaultSegmentWriteOutMediumFactory.type`|`tmpFile` or `offHeapMemory`, see explanation above|`tmpFile`| +### Indexer + +#### Indexer Process Configuration + +|Property|Description|Default| +|--------|-----------|-------| +|`druid.host`|The host for the current process. This is used to advertise the current processes location as reachable from another process and should generally be specified such that `http://${druid.host}/` could actually talk to this process|InetAddress.getLocalHost().getCanonicalHostName()| +|`druid.bindOnHost`|Indicating whether the process's internal jetty server bind on `druid.host`. Default is false, which means binding to all interfaces.|false| +|`druid.plaintextPort`|This is the port to actually listen on; unless port mapping is used, this will be the same port as is on `druid.host`|8091| +|`druid.tlsPort`|TLS port for HTTPS connector, if [druid.enableTlsPort](../operations/tls-support.md) is set then this config will be used. If `druid.host` contains port then that port will be ignored. This should be a non-negative Integer.|8283| +|`druid.service`|The name of the service. This is used as a dimension when emitting metrics and alerts to differentiate between the various services|druid/indexer| + +#### Indexer General Configuration + +|Property|Description|Default| +|--------|-----------|-------| +|`druid.worker.version`|Version identifier for the Indexer.|0| +|`druid.worker.capacity`|Maximum number of tasks the Indexer can accept.|Number of available processors - 1| +|`druid.worker.globalIngestionHeapLimitBytes`|Total amount of heap available for ingestion processing. This is applied by automatically setting the `maxBytesInMemory` property on tasks.|60% of configured JVM heap| +|`druid.worker.numConcurrentMerges`|Maximum number of segment persist or merge operations that can run concurrently across all tasks.|`druid.worker.capacity` / 2, rounded down| +|`druid.indexer.task.baseDir`|Base temporary working directory.|`System.getProperty("java.io.tmpdir")`| +|`druid.indexer.task.baseTaskDir`|Base temporary working directory for tasks.|`${druid.indexer.task.baseDir}/persistent/tasks`| +|`druid.indexer.task.defaultHadoopCoordinates`|Hadoop version to use with HadoopIndexTasks that do not request a particular version.|org.apache.hadoop:hadoop-client:2.8.3| +|`druid.indexer.task.gracefulShutdownTimeout`|Wait this long on Indexer restart for restorable tasks to gracefully exit.|PT5M| +|`druid.indexer.task.hadoopWorkingPath`|Temporary working directory for Hadoop tasks.|`/tmp/druid-indexing`| +|`druid.indexer.task.restoreTasksOnRestart`|If true, the Indexer will attempt to stop tasks gracefully on shutdown and restore them on restart.|false| +|`druid.peon.taskActionClient.retry.minWait`|The minimum retry time to communicate with Overlord.|PT5S| +|`druid.peon.taskActionClient.retry.maxWait`|The maximum retry time to communicate with Overlord.|PT1M| +|`druid.peon.taskActionClient.retry.maxRetryCount`|The maximum number of retries to communicate with Overlord.|60| + +#### Indexer Concurrent Requests + +Druid uses Jetty to serve HTTP requests. + +|Property|Description|Default| +|--------|-----------|-------| +|`druid.server.http.numThreads`|Number of threads for HTTP requests. Please see the [Indexer Server HTTP threads](../design/indexer.html#server-http-threads) documentation for more details on how the Indexer uses this configuration.|max(10, (Number of cores * 17) / 16 + 2) + 30| +|`druid.server.http.queueSize`|Size of the worker queue used by Jetty server to temporarily store incoming client connections. If this value is set and a request is rejected by jetty because queue is full then client would observe request failure with TCP connection being closed immediately with a completely empty response from server.|Unbounded| +|`druid.server.http.maxIdleTime`|The Jetty max idle time for a connection.|PT5M| +|`druid.server.http.enableRequestLimit`|If enabled, no requests would be queued in jetty queue and "HTTP 429 Too Many Requests" error response would be sent. |false| +|`druid.server.http.defaultQueryTimeout`|Query timeout in millis, beyond which unfinished queries will be cancelled|300000| +|`druid.server.http.gracefulShutdownTimeout`|The maximum amount of time Jetty waits after receiving shutdown signal. After this timeout the threads will be forcefully shutdown. This allows any queries that are executing to complete.|`PT0S` (do not wait)| +|`druid.server.http.unannouncePropagationDelay`|How long to wait for zookeeper unannouncements to propagate before shutting down Jetty. This is a minimum and `druid.server.http.gracefulShutdownTimeout` does not start counting down until after this period elapses.|`PT0S` (do not wait)| +|`druid.server.http.maxQueryTimeout`|Maximum allowed value (in milliseconds) for `timeout` parameter. See [query-context](../querying/query-context.html) to know more about `timeout`. Query is rejected if the query context `timeout` is greater than this value. |Long.MAX_VALUE| +|`druid.server.http.maxRequestHeaderSize`|Maximum size of a request header in bytes. Larger headers consume more memory and can make a server more vulnerable to denial of service attacks.|8 * 1024| + +#### Indexer Processing Resources + +|Property|Description|Default| +|--------|-----------|-------| +|`druid.processing.buffer.sizeBytes`|This specifies a buffer size for the storage of intermediate results. The computation engine in the Indexer processes will use a scratch buffer of this size to do all of their intermediate computations off-heap. Larger values allow for more aggregations in a single pass over the data while smaller values can require more passes depending on the query that is being executed.|auto (max 1GB)| +|`druid.processing.buffer.poolCacheMaxCount`|processing buffer pool caches the buffers for later use, this is the maximum count cache will grow to. note that pool can create more buffers than it can cache if necessary.|Integer.MAX_VALUE| +|`druid.processing.formatString`|Indexer processes use this format string to name their processing threads.|processing-%s| +|`druid.processing.numMergeBuffers`|The number of direct memory buffers available for merging query results. The buffers are sized by `druid.processing.buffer.sizeBytes`. This property is effectively a concurrency limit for queries that require merging buffers. If you are using any queries that require merge buffers (currently, just groupBy v2) then you should have at least two of these.|`max(2, druid.processing.numThreads / 4)`| +|`druid.processing.numThreads`|The number of processing threads to have available for parallel processing of segments. Our rule of thumb is `num_cores - 1`, which means that even under heavy load there will still be one core available to do background tasks like talking with ZooKeeper and pulling down segments. If only one core is available, this property defaults to the value `1`.|Number of cores - 1 (or 1)| +|`druid.processing.columnCache.sizeBytes`|Maximum size in bytes for the dimension value lookup cache. Any value greater than `0` enables the cache. It is currently disabled by default. Enabling the lookup cache can significantly improve the performance of aggregators operating on dimension values, such as the JavaScript aggregator, or cardinality aggregator, but can slow things down if the cache hit rate is low (i.e. dimensions with few repeating values). Enabling it may also require additional garbage collection tuning to avoid long GC pauses.|`0` (disabled)| +|`druid.processing.fifo`|If the processing queue should treat tasks of equal priority in a FIFO manner|`false`| +|`druid.processing.tmpDir`|Path where temporary files created while processing a query should be stored. If specified, this configuration takes priority over the default `java.io.tmpdir` path.|path represented by `java.io.tmpdir`| + +The amount of direct memory needed by Druid is at least +`druid.processing.buffer.sizeBytes * (druid.processing.numMergeBuffers + druid.processing.numThreads + 1)`. You can +ensure at least this amount of direct memory is available by providing `-XX:MaxDirectMemorySize=` at the command +line. + +#### Query Configurations + +See [general query configuration](#general-query-configuration). + +#### Indexer Caching + +You can optionally configure caching to be enabled on the Indexer by setting caching configs here. + +|Property|Possible Values|Description|Default| +|--------|---------------|-----------|-------| +|`druid.realtime.cache.useCache`|true, false|Enable the cache on the realtime.|false| +|`druid.realtime.cache.populateCache`|true, false|Populate the cache on the realtime.|false| +|`druid.realtime.cache.unCacheable`|All druid query types|All query types to not cache.|`["groupBy", "select"]`| +|`druid.realtime.cache.maxEntrySize`|Maximum cache entry size in bytes.|1_000_000| + +See [cache configuration](#cache-configuration) for how to configure cache settings. + +Note that only local caches such as the `local`-type cache and `caffeine` cache are supported. If a remote cache such as `memcached` is used, it will be ignored. + ### Historical For general Historical Process information, see [here](../design/historical.md). @@ -1188,7 +1273,6 @@ These Historical configurations can be defined in the `historical/runtime.proper |`druid.tlsPort`|TLS port for HTTPS connector, if [druid.enableTlsPort](../operations/tls-support.md) is set then this config will be used. If `druid.host` contains port then that port will be ignored. This should be a non-negative Integer.|8283| |`druid.service`|The name of the service. This is used as a dimension when emitting metrics and alerts to differentiate between the various services|druid/historical| - #### Historical General Configuration |Property|Description|Default| @@ -1269,6 +1353,8 @@ See [cache configuration](#cache-configuration) for how to configure cache setti This section contains the configuration options for the processes that reside on Query servers (Brokers) in the suggested [three-server configuration](../design/processes.html#server-types). +Configuration options for the experimental [Router process](../design/router.md) are also provided here. + ### Broker For general Broker process information, see [here](../design/broker.md). @@ -1613,3 +1699,28 @@ Supported query contexts: |`maxIntermediateRows`|Can be used to lower the value of `druid.query.groupBy.maxIntermediateRows` for this query.|None| |`maxResults`|Can be used to lower the value of `druid.query.groupBy.maxResults` for this query.|None| |`useOffheap`|Set to true to store aggregations off-heap when merging results.|false| + +### Router + +#### Router Process Configs + +|Property|Description|Default| +|--------|-----------|-------| +|`druid.host`|The host for the current process. This is used to advertise the current processes location as reachable from another process and should generally be specified such that `http://${druid.host}/` could actually talk to this process|InetAddress.getLocalHost().getCanonicalHostName()| +|`druid.bindOnHost`|Indicating whether the process's internal jetty server bind on `druid.host`. Default is false, which means binding to all interfaces.|false| +|`druid.plaintextPort`|This is the port to actually listen on; unless port mapping is used, this will be the same port as is on `druid.host`|8888| +|`druid.tlsPort`|TLS port for HTTPS connector, if [druid.enableTlsPort](../operations/tls-support.md) is set then this config will be used. If `druid.host` contains port then that port will be ignored. This should be a non-negative Integer.|9088| +|`druid.service`|The name of the service. This is used as a dimension when emitting metrics and alerts to differentiate between the various services|druid/router| + +#### Runtime Configuration + +|Property|Possible Values|Description|Default| +|--------|---------------|-----------|-------| +|`druid.router.defaultBrokerServiceName`|Any string.|The default Broker to connect to in case service discovery fails.|druid/broker| +|`druid.router.tierToBrokerMap`|An ordered JSON map of tiers to Broker names. The priority of Brokers is based on the ordering.|Queries for a certain tier of data are routed to their appropriate Broker.|{"_default_tier": ""}| +|`druid.router.defaultRule`|Any string.|The default rule for all datasources.|"_default"| +|`druid.router.pollPeriod`|Any ISO8601 duration.|How often to poll for new rules.|PT1M| +|`druid.router.strategies`|An ordered JSON array of objects.|Please see [Router Strategies](../design/router.html#router-strategies) for details.|[{"type":"timeBoundary"},{"type":"priority"}]| +|`druid.router.avatica.balancer.type`|String representing an AvaticaConnectionBalancer name. Please see [Avatica Query Balancing](../design/router.html#avatica-query-balancing)|Class to use for balancing Avatica queries across Brokers|rendezvousHash| +|`druid.router.http.maxRequestBufferSize`|Maximum size of the buffer used to write requests when forwarding them to the Broker. This should be set to atleast the maxHeaderSize allowed on the Broker|8 * 1024| +|`druid.router.managementProxy.enabled`|Enables the Router's [management proxy](../design/router.html#router-as-management-proxy) functionality.|false| diff --git a/docs/design/indexer.md b/docs/design/indexer.md index 4221b4e9ff6..93871d46aab 100644 --- a/docs/design/indexer.md +++ b/docs/design/indexer.md @@ -29,23 +29,31 @@ The Apache Druid (incubating) Indexer process is an alternative to the MiddleMan The Indexer is designed to be easier to configure and deploy compared to the MiddleManager + Peon system and to better enable resource sharing across tasks. -## Running +### Configuration + +For Apache Druid (incubating) Indexer Process Configuration, see [Indexer Configuration](../configuration/index.html#indexer). + +### HTTP endpoints + +The Indexer process shares the same HTTP endpoints as the [MiddleManager](../operations/api-reference.html#middlemanager). + +### Running ``` org.apache.druid.cli.Main server indexer ``` -## Task Resource Sharing +### Task resource sharing The following resources are shared across all tasks running inside an Indexer process. -### Query resources +#### Query resources The query processing threads and buffers are shared across all tasks. The Indexer will serve queries from a single endpoint shared by all tasks. -If [query caching](#indexer-caching) is enabled, the query cache is also shared across all tasks. +If [query caching](../configuration/index.html#indexer-caching) is enabled, the query cache is also shared across all tasks. -### Server HTTP threads +#### Server HTTP threads The Indexer maintains two equally sized pools of HTTP threads. @@ -55,7 +63,7 @@ The size of the pools are configured by the `druid.server.http.numThreads` confi In addition to these two pools, 2 separate threads are allocated for lookup handling. If lookups are not used, these threads will not be used. -### Memory Sharing +#### Memory sharing The Indexer uses the `druid.worker.globalIngestionHeapLimitBytes` configuration to impose a global heap limit across all of the tasks it is running. @@ -71,86 +79,13 @@ This means that the peak in-heap usage for row data can be up to approximately ` The remaining portion of the heap is reserved for query processing and segment persist/merge operations, and miscellaneous heap usage. -#### Concurrent Segment Persist/Merge Limits +#### Concurrent segment persist/merge limits To help reduce peak memory usage, the Indexer imposes a limit on the number of concurrent segment persist/merge operations across all running tasks. By default, the number of concurrent persist/merge operations is limited to (`druid.worker.capacity` / 2), rounded down. This limit can be configured with the `druid.worker.numConcurrentMerges` property. -## Runtime Configuration - -In addition to the [common configurations](../configuration/index.html#common-configurations), the Indexer accepts the following configurations: - -|Property|Description|Default| -|--------|-----------|-------| -|`druid.worker.version`|Version identifier for the Indexer.|0| -|`druid.worker.capacity`|Maximum number of tasks the Indexer can accept.|Number of available processors - 1| -|`druid.worker.globalIngestionHeapLimitBytes`|Total amount of heap available for ingestion processing. This is applied by automatically setting the `maxBytesInMemory` property on tasks.|60% of configured JVM heap| -|`druid.worker.numConcurrentMerges`|Maximum number of segment persist or merge operations that can run concurrently across all tasks.|`druid.worker.capacity` / 2, rounded down| -|`druid.indexer.task.baseDir`|Base temporary working directory.|`System.getProperty("java.io.tmpdir")`| -|`druid.indexer.task.baseTaskDir`|Base temporary working directory for tasks.|`${druid.indexer.task.baseDir}/persistent/tasks`| -|`druid.indexer.task.defaultHadoopCoordinates`|Hadoop version to use with HadoopIndexTasks that do not request a particular version.|org.apache.hadoop:hadoop-client:2.8.3| -|`druid.indexer.task.gracefulShutdownTimeout`|Wait this long on Indexer restart for restorable tasks to gracefully exit.|PT5M| -|`druid.indexer.task.hadoopWorkingPath`|Temporary working directory for Hadoop tasks.|`/tmp/druid-indexing`| -|`druid.indexer.task.restoreTasksOnRestart`|If true, the Indexer will attempt to stop tasks gracefully on shutdown and restore them on restart.|false| -|`druid.peon.taskActionClient.retry.minWait`|The minimum retry time to communicate with Overlord.|PT5S| -|`druid.peon.taskActionClient.retry.maxWait`|The maximum retry time to communicate with Overlord.|PT1M| -|`druid.peon.taskActionClient.retry.maxRetryCount`|The maximum number of retries to communicate with Overlord.|60| - -### Concurrent Requests - -Druid uses Jetty to serve HTTP requests. - -|Property|Description|Default| -|--------|-----------|-------| -|`druid.server.http.numThreads`|Number of threads for HTTP requests. Please see the [Server HTTP threads](#server-http-threads) section for more details on how the Indexer uses this configuration.|max(10, (Number of cores * 17) / 16 + 2) + 30| -|`druid.server.http.queueSize`|Size of the worker queue used by Jetty server to temporarily store incoming client connections. If this value is set and a request is rejected by jetty because queue is full then client would observe request failure with TCP connection being closed immediately with a completely empty response from server.|Unbounded| -|`druid.server.http.maxIdleTime`|The Jetty max idle time for a connection.|PT5M| -|`druid.server.http.enableRequestLimit`|If enabled, no requests would be queued in jetty queue and "HTTP 429 Too Many Requests" error response would be sent. |false| -|`druid.server.http.defaultQueryTimeout`|Query timeout in millis, beyond which unfinished queries will be cancelled|300000| -|`druid.server.http.gracefulShutdownTimeout`|The maximum amount of time Jetty waits after receiving shutdown signal. After this timeout the threads will be forcefully shutdown. This allows any queries that are executing to complete.|`PT0S` (do not wait)| -|`druid.server.http.unannouncePropagationDelay`|How long to wait for zookeeper unannouncements to propagate before shutting down Jetty. This is a minimum and `druid.server.http.gracefulShutdownTimeout` does not start counting down until after this period elapses.|`PT0S` (do not wait)| -|`druid.server.http.maxQueryTimeout`|Maximum allowed value (in milliseconds) for `timeout` parameter. See [query-context](../querying/query-context.html) to know more about `timeout`. Query is rejected if the query context `timeout` is greater than this value. |Long.MAX_VALUE| -|`druid.server.http.maxRequestHeaderSize`|Maximum size of a request header in bytes. Larger headers consume more memory and can make a server more vulnerable to denial of service attacks.|8 * 1024| - -### Processing - -|Property|Description|Default| -|--------|-----------|-------| -|`druid.processing.buffer.sizeBytes`|This specifies a buffer size for the storage of intermediate results. The computation engine in the Indexer processes will use a scratch buffer of this size to do all of their intermediate computations off-heap. Larger values allow for more aggregations in a single pass over the data while smaller values can require more passes depending on the query that is being executed.|auto (max 1GB)| -|`druid.processing.buffer.poolCacheMaxCount`|processing buffer pool caches the buffers for later use, this is the maximum count cache will grow to. note that pool can create more buffers than it can cache if necessary.|Integer.MAX_VALUE| -|`druid.processing.formatString`|Indexer processes use this format string to name their processing threads.|processing-%s| -|`druid.processing.numMergeBuffers`|The number of direct memory buffers available for merging query results. The buffers are sized by `druid.processing.buffer.sizeBytes`. This property is effectively a concurrency limit for queries that require merging buffers. If you are using any queries that require merge buffers (currently, just groupBy v2) then you should have at least two of these.|`max(2, druid.processing.numThreads / 4)`| -|`druid.processing.numThreads`|The number of processing threads to have available for parallel processing of segments. Our rule of thumb is `num_cores - 1`, which means that even under heavy load there will still be one core available to do background tasks like talking with ZooKeeper and pulling down segments. If only one core is available, this property defaults to the value `1`.|Number of cores - 1 (or 1)| -|`druid.processing.columnCache.sizeBytes`|Maximum size in bytes for the dimension value lookup cache. Any value greater than `0` enables the cache. It is currently disabled by default. Enabling the lookup cache can significantly improve the performance of aggregators operating on dimension values, such as the JavaScript aggregator, or cardinality aggregator, but can slow things down if the cache hit rate is low (i.e. dimensions with few repeating values). Enabling it may also require additional garbage collection tuning to avoid long GC pauses.|`0` (disabled)| -|`druid.processing.fifo`|If the processing queue should treat tasks of equal priority in a FIFO manner|`false`| -|`druid.processing.tmpDir`|Path where temporary files created while processing a query should be stored. If specified, this configuration takes priority over the default `java.io.tmpdir` path.|path represented by `java.io.tmpdir`| - -The amount of direct memory needed by Druid is at least -`druid.processing.buffer.sizeBytes * (druid.processing.numMergeBuffers + druid.processing.numThreads + 1)`. You can -ensure at least this amount of direct memory is available by providing `-XX:MaxDirectMemorySize=` at the command -line. - -### Query Configurations - -See [general query configuration](../configuration/index.html#general-query-configuration). - -### Indexer Caching - -You can optionally configure caching to be enabled on the Indexer by setting caching configs here. - -|Property|Possible Values|Description|Default| -|--------|---------------|-----------|-------| -|`druid.realtime.cache.useCache`|true, false|Enable the cache on the realtime.|false| -|`druid.realtime.cache.populateCache`|true, false|Populate the cache on the realtime.|false| -|`druid.realtime.cache.unCacheable`|All druid query types|All query types to not cache.|`["groupBy", "select"]`| -|`druid.realtime.cache.maxEntrySize`|Maximum cache entry size in bytes.|1_000_000| - -See [cache configuration](../configuration/index.html#cache-configuration) for how to configure cache settings. - -Note that only local caches such as the `local`-type cache and `caffeine` cache are supported. If a remote cache such as `memcached` is used, it will be ignored. - -## Current Limitations +### Current limitations Separate task logs are not currently supported when using the Indexer; all task log messages will instead be logged in the Indexer process log. diff --git a/docs/design/router.md b/docs/design/router.md index 68b348a7393..b00cf874c04 100644 --- a/docs/design/router.md +++ b/docs/design/router.md @@ -32,16 +32,133 @@ For query routing purposes, you should only ever need the Router process if you In addition to query routing, the Router also runs the [Druid Console](../operations/management-uis.html#druid-console), a management UI for datasources, segments, tasks, data processes (Historicals and MiddleManagers), and coordinator dynamic configuration. The user can also run SQL and native Druid queries within the console. +### Configuration -Running -------- +For Apache Druid (incubating) Router Process Configuration, see [Router Configuration](../configuration/index.html#router). + +### HTTP endpoints + +For a list of API endpoints supported by the Router, see [Router API](../operations/api-reference.html#router). + +### Running ``` org.apache.druid.cli.Main server router ``` -Example Production Configuration --------------------------------- +### Router as management proxy + +The Router can be configured to forward requests to the active Coordinator or Overlord process. This may be useful for +setting up a highly available cluster in situations where the HTTP redirect mechanism of the inactive -> active +Coordinator/Overlord does not function correctly (servers are behind a load balancer, the hostname used in the redirect +is only resolvable internally, etc.). + +#### Enabling the management proxy + +To enable this functionality, set the following in the Router's runtime.properties: + +``` +druid.router.managementProxy.enabled=true +``` + +#### Management proxy routing + +The management proxy supports implicit and explicit routes. Implicit routes are those where the destination can be +determined from the original request path based on Druid API path conventions. For the Coordinator the convention is +`/druid/coordinator/*` and for the Overlord the convention is `/druid/indexer/*`. These are convenient because they mean +that using the management proxy does not require modifying the API request other than issuing the request to the Router +instead of the Coordinator or Overlord. Most Druid API requests can be routed implicitly. + +Explicit routes are those where the request to the Router contains a path prefix indicating which process the request +should be routed to. For the Coordinator this prefix is `/proxy/coordinator` and for the Overlord it is `/proxy/overlord`. +This is required for API calls with an ambiguous destination. For example, the `/status` API is present on all Druid +processes, so explicit routing needs to be used to indicate the proxy destination. + +This is summarized in the table below: + +|Request Route|Destination|Rewritten Route|Example| +|-------------|-----------|---------------|-------| +|`/druid/coordinator/*`|Coordinator|`/druid/coordinator/*`|`router:8888/druid/coordinator/v1/datasources` -> `coordinator:8081/druid/coordinator/v1/datasources`| +|`/druid/indexer/*`|Overlord|`/druid/indexer/*`|`router:8888/druid/indexer/v1/task` -> `overlord:8090/druid/indexer/v1/task`| +|`/proxy/coordinator/*`|Coordinator|`/*`|`router:8888/proxy/coordinator/status` -> `coordinator:8081/status`| +|`/proxy/overlord/*`|Overlord|`/*`|`router:8888/proxy/overlord/druid/indexer/v1/isLeader` -> `overlord:8090/druid/indexer/v1/isLeader`| + +### Router strategies + +The Router has a configurable list of strategies for how it selects which Brokers to route queries to. The order of the strategies matter because as soon as a strategy condition is matched, a Broker is selected. + +#### timeBoundary + +```json +{ + "type":"timeBoundary" +} +``` + +Including this strategy means all timeBoundary queries are always routed to the highest priority Broker. + +#### priority + +```json +{ + "type":"priority", + "minPriority":0, + "maxPriority":1 +} +``` + +Queries with a priority set to less than minPriority are routed to the lowest priority Broker. Queries with priority set to greater than maxPriority are routed to the highest priority Broker. By default, minPriority is 0 and maxPriority is 1. Using these default values, if a query with priority 0 (the default query priority is 0) is sent, the query skips the priority selection logic. + +#### JavaScript + +Allows defining arbitrary routing rules using a JavaScript function. The function is passed the configuration and the query to be executed, and returns the tier it should be routed to, or null for the default tier. + +*Example*: a function that sends queries containing more than three aggregators to the lowest priority Broker. + +```json +{ + "type" : "javascript", + "function" : "function (config, query) { if (query.getAggregatorSpecs && query.getAggregatorSpecs().size() >= 3) { var size = config.getTierToBrokerMap().values().size(); if (size > 0) { return config.getTierToBrokerMap().values().toArray()[size-1] } else { return config.getDefaultBrokerServiceName() } } else { return null } }" +} +``` + +> JavaScript-based functionality is disabled by default. Please refer to the Druid [JavaScript programming guide](../development/javascript.md) for guidelines about using Druid's JavaScript functionality, including instructions on how to enable it. + + +### Avatica query balancing + +All Avatica JDBC requests with a given connection ID must be routed to the same Broker, since Druid Brokers do not share connection state with each other. + +To accomplish this, Druid provides two built-in balancers that use rendezvous hashing and consistent hashing of a request's connection ID respectively to assign requests to Brokers. + +Note that when multiple Routers are used, all Routers should have identical balancer configuration to ensure that they make the same routing decisions. + +#### Rendezvous hash balancer + +This balancer uses [Rendezvous Hashing](https://en.wikipedia.org/wiki/Rendezvous_hashing) on an Avatica request's connection ID to assign the request to a Broker. + +To use this balancer, specify the following property: + +``` +druid.router.avatica.balancer.type=rendezvousHash +``` + +If no `druid.router.avatica.balancer` property is set, the Router will also default to using the Rendezvous Hash Balancer. + +#### Consistent hash balancer + +This balancer uses [Consistent Hashing](https://en.wikipedia.org/wiki/Consistent_hashing) on an Avatica request's connection ID to assign the request to a Broker. + +To use this balancer, specify the following property: + +``` +druid.router.avatica.balancer.type=consistentHash +``` + +This is a non-default implementation that is provided for experimentation purposes. The consistent hasher has longer setup times on initialization and when the set of Brokers changes, but has a faster Broker assignment time than the rendezous hasher when tested with 5 Brokers. Benchmarks for both implementations have been provided in `ConsistentHasherBenchmark` and `RendezvousHasherBenchmark`. The consistent hasher also requires locking, while the rendezvous hasher does not. + + +### Example production configuration In this example, we have two tiers in our production cluster: `hot` and `_default_tier`. Queries for the `hot` tier are routed through the `broker-hot` set of Brokers, and queries for the `_default_tier` are routed through the `broker-cold` set of Brokers. If any exceptions or network problems occur, queries are routed to the `broker-cold` set of brokers. In our example, we are running with a c3.2xlarge EC2 instance. We assume a `common.runtime.properties` already exists. @@ -87,158 +204,3 @@ druid.router.http.numMaxThreads=100 druid.server.http.numThreads=100 ``` -Runtime Configuration ---------------------- - -The Router module uses several of the default modules in [Configuration](../configuration/index.md) and has the following set of configurations as well: - -|Property|Possible Values|Description|Default| -|--------|---------------|-----------|-------| -|`druid.router.defaultBrokerServiceName`|Any string.|The default Broker to connect to in case service discovery fails.|druid/broker| -|`druid.router.tierToBrokerMap`|An ordered JSON map of tiers to Broker names. The priority of Brokers is based on the ordering.|Queries for a certain tier of data are routed to their appropriate Broker.|{"_default_tier": ""}| -|`druid.router.defaultRule`|Any string.|The default rule for all datasources.|"_default"| -|`druid.router.pollPeriod`|Any ISO8601 duration.|How often to poll for new rules.|PT1M| -|`druid.router.strategies`|An ordered JSON array of objects.|All custom strategies to use for routing.|[{"type":"timeBoundary"},{"type":"priority"}]| -|`druid.router.avatica.balancer.type`|String representing an AvaticaConnectionBalancer name|Class to use for balancing Avatica queries across Brokers|rendezvousHash| -|`druid.router.http.maxRequestBufferSize`|Maximum size of the buffer used to write requests when forwarding them to the Broker. This should be set to atleast the maxHeaderSize allowed on the Broker|8 * 1024| - - -Router Strategies ------------------ -The Router has a configurable list of strategies for how it selects which Brokers to route queries to. The order of the strategies matter because as soon as a strategy condition is matched, a Broker is selected. - -### timeBoundary - -```json -{ - "type":"timeBoundary" -} -``` - -Including this strategy means all timeBoundary queries are always routed to the highest priority Broker. - -### priority - -```json -{ - "type":"priority", - "minPriority":0, - "maxPriority":1 -} -``` - -Queries with a priority set to less than minPriority are routed to the lowest priority Broker. Queries with priority set to greater than maxPriority are routed to the highest priority Broker. By default, minPriority is 0 and maxPriority is 1. Using these default values, if a query with priority 0 (the default query priority is 0) is sent, the query skips the priority selection logic. - -### JavaScript - -Allows defining arbitrary routing rules using a JavaScript function. The function is passed the configuration and the query to be executed, and returns the tier it should be routed to, or null for the default tier. - -*Example*: a function that sends queries containing more than three aggregators to the lowest priority Broker. - -```json -{ - "type" : "javascript", - "function" : "function (config, query) { if (query.getAggregatorSpecs && query.getAggregatorSpecs().size() >= 3) { var size = config.getTierToBrokerMap().values().size(); if (size > 0) { return config.getTierToBrokerMap().values().toArray()[size-1] } else { return config.getDefaultBrokerServiceName() } } else { return null } }" -} -``` - -> JavaScript-based functionality is disabled by default. Please refer to the Druid [JavaScript programming guide](../development/javascript.md) for guidelines about using Druid's JavaScript functionality, including instructions on how to enable it. - - -Avatica Query Balancing --------------- - -All Avatica JDBC requests with a given connection ID must be routed to the same Broker, since Druid Brokers do not share connection state with each other. - -To accomplish this, Druid provides two built-in balancers that use rendezvous hashing and consistent hashing of a request's connection ID respectively to assign requests to Brokers. - -Note that when multiple Routers are used, all Routers should have identical balancer configuration to ensure that they make the same routing decisions. - -### Rendezvous hash balancer - -This balancer uses [Rendezvous Hashing](https://en.wikipedia.org/wiki/Rendezvous_hashing) on an Avatica request's connection ID to assign the request to a Broker. - -To use this balancer, specify the following property: - -``` -druid.router.avatica.balancer.type=rendezvousHash -``` - -If no `druid.router.avatica.balancer` property is set, the Router will also default to using the Rendezvous Hash Balancer. - -### Consistent hash balancer - -This balancer uses [Consistent Hashing](https://en.wikipedia.org/wiki/Consistent_hashing) on an Avatica request's connection ID to assign the request to a Broker. - -To use this balancer, specify the following property: - -``` -druid.router.avatica.balancer.type=consistentHash -``` - -This is a non-default implementation that is provided for experimentation purposes. The consistent hasher has longer setup times on initialization and when the set of Brokers changes, but has a faster Broker assignment time than the rendezous hasher when tested with 5 Brokers. Benchmarks for both implementations have been provided in `ConsistentHasherBenchmark` and `RendezvousHasherBenchmark`. The consistent hasher also requires locking, while the rendezvous hasher does not. - -HTTP endpoints --------------- - -The Router process exposes several HTTP endpoints for interactions. - -### GET - -* `/status` - -Returns the Druid version, loaded extensions, memory used, total memory and other useful information about the process. - -* `/druid/v2/datasources` - -Returns a list of queryable datasources. - -* `/druid/v2/datasources/{dataSourceName}` - -Returns the dimensions and metrics of the datasource. - -* `/druid/v2/datasources/{dataSourceName}/dimensions` - -Returns the dimensions of the datasource. - -* `/druid/v2/datasources/{dataSourceName}/metrics` - -Returns the metrics of the datasource. - -Router as Management Proxy --------------------------- - -The Router can be configured to forward requests to the active Coordinator or Overlord process. This may be useful for -setting up a highly available cluster in situations where the HTTP redirect mechanism of the inactive -> active -Coordinator/Overlord does not function correctly (servers are behind a load balancer, the hostname used in the redirect -is only resolvable internally, etc.). - -### Enabling the Management Proxy - -To enable this functionality, set the following in the Router's runtime.properties: - -``` -druid.router.managementProxy.enabled=true -``` - -### Routing - -The management proxy supports implicit and explicit routes. Implicit routes are those where the destination can be -determined from the original request path based on Druid API path conventions. For the Coordinator the convention is -`/druid/coordinator/*` and for the Overlord the convention is `/druid/indexer/*`. These are convenient because they mean -that using the management proxy does not require modifying the API request other than issuing the request to the Router -instead of the Coordinator or Overlord. Most Druid API requests can be routed implicitly. - -Explicit routes are those where the request to the Router contains a path prefix indicating which process the request -should be routed to. For the Coordinator this prefix is `/proxy/coordinator` and for the Overlord it is `/proxy/overlord`. -This is required for API calls with an ambiguous destination. For example, the `/status` API is present on all Druid -processes, so explicit routing needs to be used to indicate the proxy destination. - -This is summarized in the table below: - -|Request Route|Destination|Rewritten Route|Example| -|-------------|-----------|---------------|-------| -|`/druid/coordinator/*`|Coordinator|`/druid/coordinator/*`|`router:8888/druid/coordinator/v1/datasources` -> `coordinator:8081/druid/coordinator/v1/datasources`| -|`/druid/indexer/*`|Overlord|`/druid/indexer/*`|`router:8888/druid/indexer/v1/task` -> `overlord:8090/druid/indexer/v1/task`| -|`/proxy/coordinator/*`|Coordinator|`/*`|`router:8888/proxy/coordinator/status` -> `coordinator:8081/status`| -|`/proxy/overlord/*`|Overlord|`/*`|`router:8888/proxy/overlord/druid/indexer/v1/isLeader` -> `overlord:8090/druid/indexer/v1/isLeader`| diff --git a/docs/operations/api-reference.md b/docs/operations/api-reference.md index d271a950d72..9b06eeed0d3 100644 --- a/docs/operations/api-reference.md +++ b/docs/operations/api-reference.md @@ -765,3 +765,22 @@ The endpoint for submitting queries. Accepts an option `?pretty` that pretty pri Returns segment information lists including server locations for the given query.. +### Router + +#### GET + +* `/druid/v2/datasources` + +Returns a list of queryable datasources. + +* `/druid/v2/datasources/{dataSourceName}` + +Returns the dimensions and metrics of the datasource. + +* `/druid/v2/datasources/{dataSourceName}/dimensions` + +Returns the dimensions of the datasource. + +* `/druid/v2/datasources/{dataSourceName}/metrics` + +Returns the metrics of the datasource. \ No newline at end of file