Changes:
- Add `Range` header to the request before opening the connection
- Use header `Content-Range` instead of `Accept-Ranges` as `Content-Range` is guaranteed to be populated if the server is returning a partial response
* changes:
* remove SystemSchema duplicate ServerInventoryView in broker
* suppress duplicate segment added/removed warnings in HttpServerInventoryView when doing a full sync
* fixes
Fixes#12022
### Description
The current implementations of memory estimation in `OnHeapIncrementalIndex` and `StringDimensionIndexer` tend to over-estimate which leads to more persistence cycles than necessary.
This PR replaces the max estimation mechanism with getting the incremental memory used by the aggregator or indexer at each invocation of `aggregate` or `encode` respectively.
### Changes
- Add new flag `useMaxMemoryEstimates` in the task context. This overrides the same flag in DefaultTaskConfig i.e. `druid.indexer.task.default.context` map
- Add method `AggregatorFactory.factorizeWithSize()` that returns an `AggregatorAndSize` which contains
the aggregator instance and the estimated initial size of the aggregator
- Add method `Aggregator.aggregateWithSize()` which returns the incremental memory used by this aggregation step
- Update the method `DimensionIndexer.processRowValsToKeyComponent()` to return the encoded key component as well as its effective size in bytes
- Update `OnHeapIncrementalIndex` to use the new estimations only if `useMaxMemoryEstimates = false`
Follow up to #12205 to allow druid-mysql-extensions to work with mysql connector/j 8.x again, which does not contain MySQLTransientException, and while would have had the same problem as mariadb if a transient exception was checked, the new check eagerly loads the class when starting up, causing immediate failure.
Fixes an issue where a load-drop-load sequence for a segment and historical doesn't work correctly for http based load queue peon. The first cycle of load-drop works fine - the problem comes when there is an attempt to reload the segment. The historical caches load success for some recent segments and makes the reload as a no-op. But it doesn't consider that fact that the segment was also dropped in between the load requests.
This change invalidates the cache after a client tries to fetch a success result.
* VirtualizedColumnSelectorFactory: Allow virtual columns to reference each other.
This matches the behavior of QueryableIndex and IncrementalIndex based cursors.
* Fixes to getColumnCapabilities.
Fixes a bug because of which some SQL queries cannot be parsed using druid convention. Specifically, these queries translate to an inline datasource and have some null values. Calcite internally uses NULL as SQL type for these literals and that is not supported by the druid.
I am now allowing null column types to be returned while building RowSignature in org.apache.druid.sql.calcite.table.RowSignatures#fromRelDataType. RowSignature already allows null column type for any column. Doing so should also fix bindable queries such as select (1,2). When such queries are run with headers set to true, we get an exception in org.apache.druid.sql.http.ArrayWriter#writeHeader. This is again a similar exception to the one addressed in this PR. Because SQL type for the result column is RECORD and that doesn't have a corresponding columnType.
* init multiValue column group by
* Changing sorting to Lexicographic as default
* Adding initial tests
* 1.Fixing test cases adding
2.Optimized inmem structs
* Linking SQL layer to native layer
* Adding multiDimension support to group by column strategy
* 1. Removing array coercion in Calcite layer
2. Removing ResultRowDeserializer
* 1. Supporting all primitive array types
2. Removing dimension spec as part of columnSelector
* 1. Supporting all primitive array types
2. Removing dimension spec as part of columnSelector
* 1. Checkstyle things
2. Removing flag
* Minor naming things
* CheckStyle Things
* Fixing test case
* Fixing hashing
* 1. Adding the MV function
2. Added few test cases
* 1. Adding MV function test cases
* Adding Selector strategy function test cases
* Fixing ClientQuerySegmentWalkerTest
* Adding GroupByQueryRunnerTest test cases
* Fixing test cases
* Adding few more test cases
* Fixing Exception asset statement and intellij inspection
* Adding null compatibility tests
* Review comments
* Fixing few failing tests
* Fixing few failing tests
* Do no convert to topN Q incase of group by on array
* Fixing checkstyle
* Fixing differences between jdk's class cast exception message
* 1. Fixing ordering if the grouping key is an array
* Fixing DefaultLimitSpec
* Fixing CalciteArraysQueryTest
* Dummy commit for LGTM
* changes:
* only coerce multi-value string null values when `ExpressionPlan.Trait.NEEDS_APPLIED` is set
* correct return type inference for ARRAY_APPEND,ARRAY_PREPEND,ARRAY_SLICE,ARRAY_CONCAT
* fix bug with ExprEval.ofType when actual type of object from binding doesn't match its claimed type
* Review comments
* Fixing test cases
* Fixing spot bugs
* Fixing strict compile
Co-authored-by: Clint Wylie <cwylie@apache.org>
This PR changes the value of the property `druid.sql.planner.useGroupingSetForExactDistinct` from `false` to `true` in the runtime.properties files, so that newer installations have this property as `true`, while the default still remains as `false`.
The flag determines how queries which contain an aggregation over `DISTINCT` like `SELECT COUNT(DISTINCT foo.dim1) FILTER(WHERE foo.cnt = 1), SUM(foo.cnt) FROM druid.foo` get planned by Calcite. With the flag being set to false, it plans it via joins, whereas with it being set to true, the query is set using grouping sets.
There is a known issue with Calcite (https://github.com/apache/druid/issues/7953), where an NPE is thrown while planning the above query with joins. There is no such issue while planning the query using grouping sets.
Makes kinesis ingestion resilient to `LimitExceededException` caused by resharding.
Replace `describeStream` with `listShards` (recommended) to get shard related info.
`describeStream` has a limit (100) to the number of shards returned per call and a low default TPS limit of 10.
`listShards` returns the info for at most 1000 shards and has a higher TPS limit of 100 as well.
Key changed/added classes in this PR
* `KinesisRecordSupplier`
* `KinesisAdminClient`
Fixed an issue where the provisionerService which can be used to spawn resources as needed is left running on a non-leader coordinator/overlord, after it is removed from leadership. Provisioning should only be done by the leader. To fix the issue, a call to stop the provisionerService was added to the stop() method of HttpRemoteTaskRunner class. The provisionerService was properly closed on other TaskRunner types.
This fixes a bug that causes TaskClient in overlord to continuously retry to pause tasks. This can happen when a task is not responding to the pause command. Ideally, in such a case when the task is unresponsive, the overlord would have given up after a few retries and would have killed the task. However, due to this bug, retries go on forever.
* add new doc
* Apply suggestions from code review
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
* reorder query laning properties
* rename doc
* new name in doc header
* organize material into "service tiering" section
* text edits and update sidebars.json
* update query laning
* how queries get assigned to lanes
* add more details to intro; use more consistent terminology
* more content
* Apply suggestions from code review
Co-authored-by: Jihoon Son <jihoonson@apache.org>
* Update docs/operations/mixed-workloads.md
* Apply suggestions from code review
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
* typo
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
Co-authored-by: Jihoon Son <jihoonson@apache.org>
Based on profiling data, about 25% of the time de-serializing DataSchema
is spent on formatting strings in validateId.
This can add up quickly, especially when de-serializing task information
in the overlord, where in can consume almost 2% of CPU if there are many
tasks.
Since the formatting is unnecessary unless the checks fail, we can
leverage the built-in formatting of Preconditions.checkArgument instead
to avoid the cost.
tejaswini-imply
Clint Wylie
Maytas Monsereenusorn
Vadim Ogievetsky
Kashif Faraz
AshishKapoor
zachjsh
Rohan Garg
Jihoon Son
Gian Merlino
Victoria Lim
Abhishek Agarwal
TSFenwick
Karan Kumar
Suneet Saldanha
somu-imply
Laksh Singla
JoyKing
AmatyaAvadhanula
Uwe Schindler
Benedict Jin
Ivan Vankovich
aggarwalakshay
Jonathan Wei
Xavier Léauté