* Unify the handling of HTTP between SQL and Native
The SqlResource and QueryResource have been
using independent logic for things like error
handling and response context stuff. This
became abundantly clear and painful during a
change I was making for Window Functions, so
I unified them into using the same code for
walking the response and serializing it.
Things are still not perfectly unified (it would
be the absolute best if the SqlResource just
took SQL, planned it and then delegated the
query run entirely to the QueryResource), but
this refactor doesn't take that fully on.
The new code leverages async query processing
from our jetty container, the different
interaction model with the Resource means that
a lot of tests had to be adjusted to align with
the async query model. The semantics of the
tests remain the same with one exception: the
SqlResource used to not log requests that failed
authorization checks, now it does.
This commit updates the task memory allocation logic.
- min task count is 2 and max task count is number of cpus on the machine
- task count increases wrt total task memory
- task memory increases from 512m to 2g
This PR expands `StringDimensionIndexer` to handle conversion of `byte[]` to base64 encoded strings, rather than the current behavior of calling java `toString`.
This issue was uncovered by a regression of sorts introduced by #13519, which updated the protobuf extension to directly convert stuff to java types, resulting in `bytes` typed values being converted as `byte[]` instead of a base64 string which the previous JSON based conversion created. While outputting `byte[]` is more consistent with other input formats, and preferable when the bytes can be consumed directly (such as complex types serde), when fed to a `StringDimensionIndexer`, it resulted in an ugly java `toString` because `processRowValsToUnsortedEncodedKeyComponent` is fed the output of `row.getRaw(..)`. Converting `byte[]` to a base64 string within `StringDimensionIndexer` is consistent with the behavior of calling `row.getDimension(..)` which does do this coercion (and why many tests on binary types appeared to be doing the expected thing).
I added some protobuf `bytes` tests, but they don't really hit the new `StringDimensionIndexer` behavior because they operate on the `InputRow` directly, and call `getDimension` to validate stuff. The parser based version still uses the old conversion mechanisms, so when not using a flattener incorrectly calls `toString` on the `ByteString`. I have encoded this behavior in the test for now, if we either update the parser to use the new flattener or just .. remove parsers we can remove this test stuff.
* bump nested column format version
changes:
* nested field files are now named by their position in field paths list, rather than directly by the path itself. this fixes issues with valid json properties with commas and newlines breaking the csv file meta.smoosh
* update StructuredDataProcessor to deal in NestedPathPart to be consistent with other abstract path handling rather than building JQ syntax strings directly
* add v3 format segment and test
Main change: clarify that the "default value" for casts only applies if
druid.generic.useDefaultValueForNull = true.
Secondary change: adjust a bunch of wording from future to present tense.
* docs: notebook for API tutorial
* Apply suggestions from code review
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
* address the other comments
* typo
* add commentary to outputs
* address feedback from will
* delete unnecessary comment
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
Follow up to #13520
Bytes processed are currently tracked for intermediate stages in MSQ ingestion.
This patch adds the capability to track the bytes processed by an MSQ controller
task while reading from an external input source or a segment source.
Changes:
- Track `processedBytes` for every `InputSource` read in `ExternalInputSliceReader`
- Update `ChannelCounters` with the above obtained `processedBytes` when incrementing
the input file count.
- Update task report structure in docs
The total input processed bytes can be obtained by summing the `processedBytes` as follows:
totalBytes = 0
for every root stage (i.e. a stage which does not have another stage as an input):
for every worker in that stage:
for every input channel: (i.e. channels with prefix "input", e.g. "input0", "input1", etc.)
totalBytes += processedBytes
This commit fixes a bug with nested column "value set" indexes caused by not properly
validating that the globalId looked up for value is present in the global dictionary prior to
looking it up in the local dictionary, which when "adjusting" the global ids for value type
can cause incorrect selection of value indexes.
To use an example of a variant typed nested column with 3 values `["1", null, -2]`.
The string dictionary is `[null, "1"]`, the long dictionary is `[-2]` and our local dictionary is `[0, 1, 2]`.
The code for variant typed indexes checks if the value is present in all global dictionaries
and returns indexes for all matches. So in this case, we first lookup "1" in the string dictionary,
find it at global id 1, all is good. Now, we check the long dictionary for `1`, which due to
`-(insertionpoint + 1)` gives us `-(1 + 2) = -2`. Since the global id space is actually stacked
dictionaries, global ids for long and double values must be "adjusted" by the size of string
dictionary, and size of string + size of long for doubles.
Prior to this patch we were not checking that the globalId is 0 or larger, we then immediately
looked up the `localDictionary.indexOf(-2 + adjustLong) = localDictionary.indexOf(-2 + 2) = localDictionary.indexOf(0)` ... which is an actual value contained in the dictionary! The fix is
to skip the longs completely since there were no global matches.
On to doubles, `-(insertionPoint + 1)` gives us `-(0 + 1) = -1`. The double adjust value is '3'
since 2 strings and 1 long, so `localDictionary.indexOf(-1 + 3)` = `localDictionary.indexOf(2)`
which is also a real value in our local dictionary that is definitely not '1'.
So in this one case, looking for '1' incorrectly ended up matching every row.
Changes:
* Use 80% of memory specified for running services (versus 50% earlier).
* Tasks get either 512m / 1024m or 2048m now (versus 512m or 2048m earlier).
* Add direct memory for router.
* Add validation checks to worker chat handler apis
* Merge things and polishing the error messages.
* Minor error message change
* Fixing race and adding some tests
* Fixing controller fetching stats from wrong workers.
Fixing race
Changing default mode to Parallel
Adding logging.
Fixing exceptions not propagated properly.
* Changing to kernel worker count
* Added a better logic to figure out assigned worker for a stage.
* Nits
* Moving to existing kernel methods
* Adding more coverage
Co-authored-by: cryptoe <karankumar1100@gmail.com>
* Support Framing for Window Aggregations
This adds support for framing over ROWS
for window aggregations.
Still not implemented as yet:
1. RANGE frames
2. Multiple different frames in the same query
3. Frames on last/first functions
This commit adds a new class `InputStats` to track the total bytes processed by a task.
The field `processedBytes` is published in task reports along with other row stats.
Major changes:
- Add class `InputStats` to track processed bytes
- Add method `InputSourceReader.read(InputStats)` to read input rows while counting bytes.
> Since we need to count the bytes, we could not just have a wrapper around `InputSourceReader` or `InputEntityReader` (the way `CountableInputSourceReader` does) because the `InputSourceReader` only deals with `InputRow`s and the byte information is already lost.
- Classic batch: Use the new `InputSourceReader.read(inputStats)` in `AbstractBatchIndexTask`
- Streaming: Increment `processedBytes` in `StreamChunkParser`. This does not use the new `InputSourceReader.read(inputStats)` method.
- Extend `InputStats` with `RowIngestionMeters` so that bytes can be exposed in task reports
Other changes:
- Update tests to verify the value of `processedBytes`
- Rename `MutableRowIngestionMeters` to `SimpleRowIngestionMeters` and remove duplicate class
- Replace `CacheTestSegmentCacheManager` with `NoopSegmentCacheManager`
- Refactor `KafkaIndexTaskTest` and `KinesisIndexTaskTest`
Refactor DataSource to have a getAnalysis method()
This removes various parts of the code where while loops and instanceof
checks were being used to walk through the structure of DataSource objects
in order to build a DataSourceAnalysis. Instead we just ask the DataSource
for its analysis and allow the stack to rebuild whatever structure existed.
* Zero-copy local deep storage.
This is useful for local deep storage, since it reduces disk usage and
makes Historicals able to load segments instantaneously.
Two changes:
1) Introduce "druid.storage.zip" parameter for local storage, which defaults
to false. This changes default behavior from writing an index.zip to writing
a regular directory. This is safe to do even during a rolling update, because
the older code actually already handled unzipped directories being present
on local deep storage.
2) In LocalDataSegmentPuller and LocalDataSegmentPusher, use hard links
instead of copies when possible. (Generally this is possible when the
source and destination directory are on the same filesystem.)
* Druid automated quickstart
* remove conf/druid/single-server/quickstart/_common/historical/jvm.config
* Minor changes in python script
* Add lower bound memory for some services
* Additional runtime properties for services
* Update supervise script to accept command arguments, corresponding changes in druid-quickstart.py
* File end newline
* Limit the ability to start multiple instances of a service, documentation changes
* simplify script arguments
* restore changes in medium profile
* run-druid refactor
* compute and pass middle manager runtime properties to run-druid
supervise script changes to process java opts array
use argparse, leave free memory, logging
* Remove extra quotes from mm task javaopts array
* Update logic to compute minimum memory
* simplify run-druid
* remove debug options from run-druid
* resolve the config_path provided
* comment out service specific runtime properties which are computed in the code
* simplify run-druid
* clean up docs, naming changes
* Throw ValueError exception on illegal state
* update docs
* rename args, compute_only -> compute, run_zk -> zk
* update help documentation
* update help documentation
* move task memory computation into separate method
* Add validation checks
* remove print
* Add validations
* remove start-druid bash script, rename start-druid-main
* Include tasks in lower bound memory calculation
* Fix test
* 256m instead of 256g
* caffeine cache uses 5% of heap
* ensure min task count is 2, task count is monotonic
* update configs and documentation for runtime props in conf/druid/single-server/quickstart
* Update docs
* Specify memory argument for each profile in single-server.md
* Update middleManager runtime.properties
* Move quickstart configs to conf/druid/base, add bash launch script, support python2
* Update supervise script
* rename base config directory to auto
* rename python script, changes to pass repeated args to supervise
* remove exmaples/conf/druid/base dir
* add docs
* restore changes in conf dir
* update start-druid-auto
* remove hashref for commands in supervise script
* start-druid-main java_opts array is comma separated
* update entry point script name in python script
* Update help docs
* documentation changes
* docs changes
* update docs
* add support for running indexer
* update supported services list
* update help
* Update python.md
* remove dir
* update .spelling
* Remove dependency on psutil and pathlib
* update docs
* Update get_physical_memory method
* Update help docs
* update docs
* update method to get physical memory on python
* udpate spelling
* update .spelling
* minor change
* Minor change
* memory comptuation for indexer
* update start-druid
* Update python.md
* Update single-server.md
* Update python.md
* run python3 --version to check if python is installed
* Update supervise script
* start-druid: echo message if python not found
* update anchor text
* minor change
* Update condition in supervise script
* JVM not jvm in docs
The planner sets sqlInsertSegmentGranularity in its context when using
PARTITIONED BY, which sets it on every native query in the stack (as all
native queries for a SQL query typically have the same context).
QueryKit would interpret that as a request to configure bucketing for
all native queries. This isn't useful, as bucketing is only used for
the penultimate stage in INSERT / REPLACE.
So, this patch modifies QueryKit to only look at sqlInsertSegmentGranularity
on the outermost query.
As an additional change, this patch switches the static ObjectMapper to
use the processwide ObjectMapper for deserializing Granularities. Saves
an ObjectMapper instance, and ensures that if there are any special
serdes registered for Granularity, we'll pick them up.
1) Edited the TooManyBuckets error message to mention PARTITIONED BY
instead of segmentGranularity.
2) Added error-code-specific anchors in the docs.
3) Add information to various error codes in the docs about common
causes and solutions.
* Remove stray reference to fix OOM while merging sketches
* Update future to add result from executor service
* Update tests and address review comments
* Address review comments
* Moved mock
* Close threadpool on teardown
* Remove worker task cancel
* add padding and keywords
* add arrayOfDoubles
* Update docs/development/extensions-core/datasketches-tuple.md
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
* Update docs/development/extensions-core/datasketches-tuple.md
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
* Update docs/development/extensions-core/datasketches-tuple.md
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
* Update docs/development/extensions-core/datasketches-tuple.md
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
* Update docs/development/extensions-core/datasketches-tuple.md
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
* partiton int
* fix docs
Co-authored-by: Charles Smith <techdocsmith@gmail.com>
Changes:
- Limit max batch size in `SegmentAllocationQueue` to 500
- Rename `batchAllocationMaxWaitTime` to `batchAllocationWaitTime` since the actual
wait time may exceed this configured value.
- Replace usage of `SegmentInsertAction` in `TaskToolbox` with `SegmentTransactionalInsertAction`
* Processors for Window Processing
This is an initial take on how to use Processors
for Window Processing. A Processor is an interface
that transforms RowsAndColumns objects.
RowsAndColumns objects are essentially combinations
of rows and columns.
The intention is that these Processors are the start
of a set of operators that more closely resemble what
DB engineers would be accustomed to seeing.
* Wire up windowed processors with a query type that
can run them end-to-end. This code can be used to
actually run a query, so yay!
* Wire up windowed processors with a query type that
can run them end-to-end. This code can be used to
actually run a query, so yay!
* Some SQL tests for window functions. Added wikipedia
data to the indexes available to the
SQL queries and tests validating the windowing
functionality as it exists now.
Co-authored-by: Gian Merlino <gianmerlino@gmail.com>
* Switching emitter. This will allow for a per feed emitter designation.
This will work by looking at an event's feed and direct it to a specific emitter. If no specific feed is specified for a feed.
The emitter can direct the event to a default emitter.
* fix checkstyle issues and make docs for switching emitter use basic event feeds
* fix broken docs, add test, and guard against misconfigurations
* add module test
add switching emitter module test
* fix broken SwitchingEmitterModuleTest
* add apache license to top of test
* fix checkstyle issues
* address comments by adding javadocs, removing a todo, and making druid docs more clear
In a cluster with a large number of streaming tasks (~1000), SegmentAllocateActions
on the overlord can often take very long intervals of time to finish thus causing spikes
in the `task/action/run/time`. This may result in lag building up while a task waits for a
segment to get allocated.
The root causes are:
- large number of metadata calls made to the segments and pending segments tables
- `giant` lock held in `TaskLockbox.tryLock()` to acquire task locks and allocate segments
Since the contention typically arises when several tasks of the same datasource try
to allocate segments for the same interval/granularity, the allocation run times can be
improved by batching the requests together.
Changes
- Add flags
- `druid.indexer.tasklock.batchSegmentAllocation` (default `false`)
- `druid.indexer.tasklock.batchAllocationMaxWaitTime` (in millis) (default `1000`)
- Add methods `canPerformAsync` and `performAsync` to `TaskAction`
- Submit each allocate action to a `SegmentAllocationQueue`, and add to correct batch
- Process batch after `batchAllocationMaxWaitTime`
- Acquire `giant` lock just once per batch in `TaskLockbox`
- Reduce metadata calls by batching statements together and updating query filters
- Except for batching, retain the whole behaviour (order of steps, retries, etc.)
- Respond to leadership changes and fail items in queue when not leader
- Emit batch and request level metrics
Kashif Faraz
imply-cheddar
Vadim Ogievetsky
Gian Merlino
Victoria Lim
317brian
Rishabh Singh
Clint Wylie
Adarsh Sanjeev
Rohan Garg
somu-imply
Karan Kumar
Paul Rogers
Jill Osborne
Abhishek Agarwal
Tejaswini Bandlamudi
AmatyaAvadhanula
TSFenwick