* dynamic coord config adding more balancing control
add new dynamic coordinator config, maxSegmentsToConsiderPerMove. This
config caps the number of segments that are iterated over when selecting
a segment to move. The default value combined with current balancing
strategies will still iterate over all provided segments. However,
setting this value to something > 0 will cap the number of segments
visited. This could make sense in cases where a cluster has a very large
number of segments and the admins prefer less iterations vs a thorough
consideration of all segments provided.
* fix checkstyle failure
* Make doc more detailed for admin to understand when/why to use new config
* refactor PR to use a % of segments instead of raw number
* update the docs
* remove bad doc line
* fix typo in name of new dynamic config
* update RservoirSegmentSampler to gracefully deal with values > 100%
* add handler for <= 0 in ReservoirSegmentSampler
* fixup CoordinatorDynamicConfigTest naming and argument ordering
* fix items in docs after spellcheck flags
* Fix lgtm flag on missing space in string literal
* improve documentation for new config
* Add default value to config docs and add advice in cluster tuning doc
* Add percentOfSegmentsToConsiderPerMove to web console coord config dialog
* update jest snapshot after console change
* fix spell checker errors
* Improve debug logging in getRandomSegmentBalancerHolder to cover all bad inputs for % of segments to consider
* add new config back to web console module after merge with master
* fix ReservoirSegmentSamplerTest
* fix line breaks in coordinator console dialog
* Add a test that helps ensure not regressions for percentOfSegmentsToConsiderPerMove
* Make improvements based off of feedback in review
* additional cleanup coming from review
* Add a warning log if limit on segments to consider for move can't be calcluated
* remove unused import
* fix tests for CoordinatorDynamicConfig
* remove precondition test that is redundant in CoordinatorDynamicConfig Builder class
* add query through to server selector
* add nullable extensions, deprecate old methods with defaults
* style changes
* add nullable to ServerSelectorStrategy
* fix test coverage
* missing override in test
* add null check
* CompressionUtils: Read the entire stream when unzipping from a stream.
Should fix#6905 by making sure we avoid closing partially-read streams.
* CHECKSTYLE!
* integration test for coordinator and overlord leadership, added sys.servers is_leader column
* docs
* remove not needed
* fix comments
* fix compile heh
* oof
* revert unintended
* fix tests, split out docker-compose file selection from starting cluster, use docker-compose down to stop cluster
* fixes
* style
* dang
* heh
* scripts are hard
* fix spelling
* fix thing that must not matter since was already wrong ip, log when test fails
* needs more heap
* fix merge
* less aggro
* add a travls job to do integrate test on K8s
* revert build_run_cluster.sh
* revert msic
* run IT test
* ready to test
* modify before/after script
* done
* change mod for script
* done
* add env DRUID_OPERATOR_VERSION=0.0.3
* change version
Co-authored-by: yuezhang <yuezhang@freewheel.tv>
Added the status/selfDiscovered endpoint to indexer. Per the api-reference doc, all services support status/selfDiscovered endpoint. So this change would fix that expected behavior.
Also added example config files for indexer process that can be used to spin up the indexer process.
* fix race condition with DruidSchema tables and dataSourcesNeedingRebuild
* rework to see if it passes analysis
* more better
* maybe this
* re-arrange and comments
* no need for intervals
* don't set redundant fields
* fix tests
* better filter control
* work with not
* wrap callout with form group
* update snapshot
* add split hint
* highlight issues with spec
* fixes
* fix default value
* move intervals back to partition step
* work with all sorts of chars
* fix enabled view
Currently `getConfPath` function returns an empty string for `indexer` service. So the config directory path for this service is not set properly when installed using docker environment.
* Fixes and tests related to the Indexer process.
Three bugs fixed:
1) Indexers would not announce themselves as segment servers if they
did not have storage locations defined. This used to work, but was
broken in #9971. Fixed this by adding an "isSegmentServer" method
to ServerType and updating SegmentLoadDropHandler to always announce
if this method returns true.
2) Certain batch task types were written in a way that assumed "isReady"
would be called before "run", which is not guaranteed. In particular,
they relied on it in order to initialize "taskLockHelper". Fixed this
by updating AbstractBatchIndexTask to ensure "isReady" is called
before "run" for these tasks.
3) UnifiedIndexerAppenderatorsManager did not properly handle complex
datasources. Introduced DataSourceAnalysis in order to fix this.
Test changes:
1) Add a new "docker-compose.cli-indexer.yml" config that spins up an
Indexer instead of a MiddleManager.
2) Introduce a "USE_INDEXER" environment variable that determines if
docker-compose will start up an Indexer or a MiddleManager.
3) Duplicate all the jdk8 tests and run them in both MiddleManager and
Indexer mode.
4) Various adjustments to encourage fail-fast errors in the Docker
build scripts.
5) Various adjustments to speed up integration tests and reduce memory
usage.
6) Add another Mac-specific approach to determining a machine's own IP.
This was useful on my development machine.
7) Update segment-count check in ITCompactionTaskTest to eliminate a
race condition (it was looking for 6 segments, which only exist
together briefly, until the older 4 are marked unused).
Javadoc updates:
1) AbstractBatchIndexTask: Added javadocs to determineLockGranularityXXX
that make it clear when taskLockHelper will be initialized as a side
effect. (Related to the second bug above.)
2) Task: Clarified that "isReady" is not guaranteed to be called before
"run". It was already implied, but now it's explicit.
3) ZkCoordinator: Clarified deprecation message.
4) DataSegmentServerAnnouncer: Clarified deprecation message.
* Fix stop_cluster script.
* Fix sanity check in script.
* Fix hashbang lines.
* Test and doc adjustments.
* Additional tests, and adjustments for tests.
* Split ITs back out.
* Revert change to druid_coordinator_period_indexingPeriod.
* Set Indexer capacity to match MM.
* Bump up Historical memory.
* Bump down coordinator, overlord memory.
* Bump up Broker memory.
* fix to allow customer storage location selector strategy
* add test cases to check instance of selector strategy
* update doc
* code format
* resolve code review comments
* inject StorageLocation
* fix CI
* fix mismatched license item reported by CI
* change property path from druid.segmentCache.locationSelectorStrategy.type to druid.segmentCache.locationSelector.strategy
* using a helper method to bind to correct property path
* better API escape
* fix escaping issue, bigints
* update licenses
* fix align
* do not show Query with SQL if no SQL
* add prettify script
* update dev readme
* add ordering to the datasource list
* add ordering to supervisor table
This is useful because it groups up segments for the same time chunk
into the same splits, which in turn is useful because it minimizes the
number of time chunks that each task will have to deal with.
* First draft of grouping_id function
* Add more tests and documentation
* Add calcite tests
* Fix travis failures
* bit of a change
* Add documentation
* Fix typos
* typo fix
1) Accelerate coordinator runs to speed up segment load after publishing.
2) For streaming ingestion tests, Instead of waiting 3 minutes for data to
load, wait until the expected number of rows is loaded.
Also updates segment-count check in ITCompactionTaskTest to eliminate a
race condition (it was looking for 6 segments, which only exist together
briefly, until the older 4 are marked unused).
* Two fixes related to encoding of % symbols.
1) TaskResourceFilter: Don't double-decode task ids. request.getPathSegments()
returns already-decoded strings. Applying StringUtils.urlDecode on
top of that causes erroneous behavior with '%' characters.
2) Update various ThreadFactoryBuilder name formats to escape '%'
characters. This fixes situations where substrings starting with '%'
are erroneously treated as format specifiers.
ITs are updated to include a '%' in extra.datasource.name.suffix.
* Avoid String.replace.
* Work around surefire bug.
* Fix xml encoding.
* Another try at the proper encoding.
* Give up on the emojis.
* Less ambitious testing.
* Fix an additional problem.
* Adjust encodeForFormat to return null if the input is null.
These caused certain APIs to not actually be properly forbidden.
Also removed two MoreExecutors entries for methods that don't exist in
our version of Guava.
* Move common methods that are used in HadoopTuningConfig and in AppenderatorConfig to TuningConfig
* Rename rowFlushBoundary in HadoopTuningConfig to maxRowsInMemory to match TuningConfig API
* Add context dimension to DefaultQueryMetrics
* remove redundant addition of context dimension from DruidMetrics now that QueryMetrics adds it by default
* update SearchQueryMetrics to reflect the same pattern as other default dimensions in QueryMetrics
* add PublicApi annotation for context in QueryMetrics Interface
Charles Smith
Xavier Léauté
Lucas Capistrant
Maytas Monsereenusorn
Clint Wylie
keefe roedersheimer
Gian Merlino
Abhishek Agarwal
sthetland
zhangyue19921010
Vadim Ogievetsky
Himanshu
Harini Rajendran
Jihoon Son
Harini Rajendran
Atul Mohan
Vyatcheslav Mogilevsky
frank chen
egor-ryashin
Liran Funaro
Valdemar
Suneet Saldanha