Removes the coordinator sanity check that prevents it from dropping all
segments. It's useful to get rid of this, since the behavior is
unintuitive for dev/testing clusters where users might regularly want
to drop all their data to get back to a clean slate.
But the sanity check was there for a reason: to prevent a race condition
where the coordinator might drop all segments if it ran before the
first metadata store poll finished. This patch addresses that concern
differently, by allowing methods in MetadataSegmentManager to return
null if a poll has not happened yet, and canceling coordinator runs
in that case.
This patch also makes the "dataSources" reference in
SQLMetadataSegmentManager volatile. I'm not sure why it wasn't volatile
before, but it seems necessary to me: it's not final, and it's dereferenced
from multiple threads without synchronization.
* refactor lookups to be more chill to router
* remove accidental change
* fix and combine LookupIntrospectionResourceTest
* fix inspection
* rename RouterLookupModule to LookupSerdeModule and RouterLookupExtractorFactoryContainerProvider to NoopLookupExtractorFactoryContainerProvider
* make comment generic
* use ConfigResourceFilter instead of StateResourceFilter
* fix indentation
* unused import
* another unused import
* refactor some stuff into processing module, split up LookupModule.java classes into their own files
* Fix two issues with Coordinator -> Overlord communication.
1) ClientCompactQuery needs to recognize the potential for 'intervals'
to be set instead of 'segments'. The lack of this led to a
NullPointerException on DruidCoordinatorSegmentCompactor.java:102.
2) In two locations (DruidCoordinatorSegmentCompactor,
DruidCoordinatorCleanupPendingSegments) tasks were being retrieved
using waiting/pending/running tasks in the wrong order: by checking
'running' first and then 'pending', tasks could be missed if they
moved from 'pending' to 'running' in between the two calls. Replaced
these methods with calls to 'getActiveTasks', a new method that does
the calls in the right order.
* Remove unused import.
* Make IngestSegmentFirehoseFactory splittable for parallel ingestion
* Code review feedback
- Get rid of WindowedSegment
- Don't document 'segments' parameter or support splitting firehoses that use it
- Require 'intervals' in WindowedSegmentId (since it won't be written by hand)
* Add missing @JsonProperty
* Integration test passes
* Add unit test
* Remove two FIXME comments from CompactionTask
I'd like to leave this PR in a potentially mergeable state, but I still would
appreciate reviewer eyes on the questions I'm removing here.
* Updates from code review
* Update init
Fix bin/init to source from proper directory.
* Fix for Proposal #6518: Shutdown druid processes upon complete loss of ZK connectivity
* Zookeeper Loss:
- Add feature documentation
- Cosmetic refactors
- Variable extractions
- Remove getter
* - Change config key name and reword documentation
- Switch from Function<Void,Void> to Runnable/Lambda
- try { … } finally { … }
* Fix line length too long
* - change to formatted string for logging
- use System.err.println after lifecycle stops
* commenting on makeEnsembleProvider()-created Zookeeper termination
* Add javadoc
* added java doc reference back to apache discussion thread.
* move comment to other class
* favor two-slash comments instead of multiline comments
* maxTotalRows should be checked in DataSourceCompactionConfig before setting targetCompactionSizeBytes
* remove unnecessary default values
* remove flacky test
* fix build
* Add comments
* Moved Scan Builder to Druids class and started on Scan Benchmark setup
* Need to form queries
* It runs.
* Stuff for time-ordered scan query
* Move ScanResultValue timestamp comparator to a separate class for testing
* Licensing stuff
* Change benchmark
* Remove todos
* Added TimestampComparator tests
* Change number of benchmark iterations
* Added time ordering to the scan benchmark
* Changed benchmark params
* More param changes
* Benchmark param change
* Made Jon's changes and removed TODOs
* Broke some long lines into two lines
* nit
* Decrease segment size for less memory usage
* Wrote tests for heapsort scan result values and fixed bug where iterator
wasn't returning elements in correct order
* Wrote more tests for scan result value sort
* Committing a param change to kick teamcity
* Fixed codestyle and forbidden API errors
* .
* Improved conciseness
* nit
* Created an error message for when someone tries to time order a result
set > threshold limit
* Set to spaces over tabs
* Fixing tests WIP
* Fixed failing calcite tests
* Kicking travis with change to benchmark param
* added all query types to scan benchmark
* Fixed benchmark queries
* Renamed sort function
* Added javadoc on ScanResultValueTimestampComparator
* Unused import
* Added more javadoc
* improved doc
* Removed unused import to satisfy PMD check
* Small changes
* Changes based on Gian's comments
* Fixed failing test due to null resultFormat
* Added config and get # of segments
* Set up time ordering strategy decision tree
* Refactor and pQueue works
* Cleanup
* Ordering is correct on n-way merge -> still need to batch events into
ScanResultValues
* WIP
* Sequence stuff is so dirty :(
* Fixed bug introduced by replacing deque with list
* Wrote docs
* Multi-historical setup works
* WIP
* Change so batching only occurs on broker for time-ordered scans
Restricted batching to broker for time-ordered queries and adjusted
tests
Formatting
Cleanup
* Fixed mistakes in merge
* Fixed failing tests
* Reset config
* Wrote tests and added Javadoc
* Nit-change on javadoc
* Checkstyle fix
* Improved test and appeased TeamCity
* Sorry, checkstyle
* Applied Jon's recommended changes
* Checkstyle fix
* Optimization
* Fixed tests
* Updated error message
* Added error message for UOE
* Renaming
* Finish rename
* Smarter limiting for pQueue method
* Optimized n-way merge strategy
* Rename segment limit -> segment partitions limit
* Added a bit of docs
* More comments
* Fix checkstyle and test
* Nit comment
* Fixed failing tests -> allow usage of all types of segment spec
* Fixed failing tests -> allow usage of all types of segment spec
* Revert "Fixed failing tests -> allow usage of all types of segment spec"
This reverts commit ec470288c7.
* Revert "Merge branch '6088-Time-Ordering-On-Scans-N-Way-Merge' of github.com:justinborromeo/incubator-druid into 6088-Time-Ordering-On-Scans-N-Way-Merge"
This reverts commit 57033f36df, reversing
changes made to 8f01d8dd16.
* Check type of segment spec before using for time ordering
* Fix bug in numRowsScanned
* Fix bug messing up count of rows
* Fix docs and flipped boolean in ScanQueryLimitRowIterator
* Refactor n-way merge
* Added test for n-way merge
* Refixed regression
* Checkstyle and doc update
* Modified sequence limit to accept longs and added test for long limits
* doc fix
* Implemented Clint's recommendations
* Move GCP to a core extension
* Don't provide druid-core >.<
* Keep AWS and GCP modules separate
* Move AWSModule to its own module
* Add aws ec2 extension and more modules in more places
* Fix bad imports
* Fix test jackson module
* Include AWS and GCP core in server
* Add simple empty method comment
* Update version to 15
* One more 0.13.0-->0.15.0 change
* Fix multi-binding problem
* Grep for s3-extensions and update docs
* Update extensions.md
* Fix exclusivity for start offset in kinesis indexing service
* some adjustment
* Fix SeekableStreamDataSourceMetadata
* Add missing javadocs
* Add missing comments and unit test
* fix SeekableStreamStartSequenceNumbers.plus and add comments
* remove extra exclusivePartitions in KafkaIOConfig and fix downgrade issue
* Add javadocs
* fix compilation
* fix test
* remove unused variable
* Avoid many unnecessary materializations of collections of 'all segments in cluster' cardinality
* Fix DruidCoordinatorTest; Renamed DruidCoordinator.getReplicationStatus() to computeUnderReplicationCountsPerDataSourcePerTier()
* More Javadocs, typos, refactor DruidCoordinatorRuntimeParams.createAvailableSegmentsSet()
* Style
* typo
* Disable StaticPseudoFunctionalStyleMethod inspection because of too much false positives
* Fixes
* Throw caught exception.
* Throw caught exceptions.
* Related checkstyle rule is added to prevent further bugs.
* RuntimeException() is used instead of Throwables.propagate().
* Missing import is added.
* Throwables are propogated if possible.
* Throwables are propogated if possible.
* Throwables are propogated if possible.
* Throwables are propogated if possible.
* * Checkstyle definition is improved.
* Throwables.propagate() usages are removed.
* Checkstyle pattern is changed for only scanning "Throwables.propagate(" instead of checking lookbehind.
* Throwable is kept before firing a Runtime Exception.
* Fix unused assignments.
#### `EventReceiverFirehoseFactory`
Fixed several concurrency bugs in `EventReceiverFirehoseFactory`:
- Race condition over putting an entry into `producerSequences` in `checkProducerSequence()`.
- `Stopwatch` used to measure time across threads, but it's a non-thread-safe class.
- Use `System.nanoTime()` instead of `System.currentTimeMillis()` because the latter are [not suitable](https://stackoverflow.com/a/351571/648955) for measuring time intervals.
- `close()` was not synchronized by could be called from multiple threads concurrently.
Removed unnecessary `readLock` (protecting `hasMore()` and `nextRow()` which are always called from a single thread). Removed unnecessary `volatile` modifiers.
Documented threading model and concurrent control flow of `EventReceiverFirehose` instances.
**Important:** please read the updated Javadoc for `EventReceiverFirehose.addAll()`. It allows events from different requests (batches) to be interleaved in the buffer. Is this OK?
#### `TimedShutoffFirehoseFactory`
- Fixed a race condition that was possible because `close()` that was not properly synchronized.
Documented threading model and concurrent control flow of `TimedShutoffFirehose` instances.
#### `Firehose`
Refined concurrency contract of `Firehose` based on `EventReceiverFirehose` implementation. Importantly, now it states that `close()` doesn't affect `hasMore()` and `nextRow()` and could be called concurrently with them. In other words, specified that `close()` is for "row supply" side rather than "row consume" side. However, I didn't check that other `Firehose` implementatations adhere to this contract.
<hr>
This issue is the result of reviewing `EventReceiverFirehose` and `TimedShutoffFirehose` using [this checklist](https://medium.com/@leventov/code-review-checklist-java-concurrency-49398c326154).
* Added checkstyle for "Methods starting with Capital Letters" and changed the method names violating this.
* Un-abbreviate the method names in the calcite tests
* Fixed checkstyle errors
* Changed asserts position in the code
IndexTask had special-cased code to properly send a TaskToolbox to a
IngestSegmentFirehoseFactory that's nested inside a CombiningFirehoseFactory,
but ParallelIndexSubTask didn't.
This change refactors IngestSegmentFirehoseFactory so that it doesn't need a
TaskToolbox; it instead gets a CoordinatorClient and a SegmentLoaderFactory
directly injected into it.
This also refactors SegmentLoaderFactory so it doesn't depend on
an injectable SegmentLoaderConfig, since its only method always
replaces the preconfigured SegmentLoaderConfig anyway.
This makes it possible to use SegmentLoaderFactory without setting
druid.segmentCaches.locations to some dummy value.
Another goal of this PR is to make it possible for IngestSegmentFirehoseFactory
to list data segments outside of connect() --- specifically, to make it a
FiniteFirehoseFactory which can query the coordinator in order to calculate its
splits. See #7048.
This also adds missing datasource name URL-encoding to an API used by
CoordinatorBasedSegmentHandoffNotifier.
* Remove DataSegmentFinder, InsertSegmentToDb, and descriptor.json file
* delete descriptor.file when killing segments
* fix test
* Add doc for ha
* improve warning
* maintenance mode for Historical
forbidden api fix, config deserialization fix
logging fix, unit tests
* addressed comments
* addressed comments
* a style fix
* addressed comments
* a unit-test fix due to recent code-refactoring
* docs & refactoring
* addressed comments
* addressed a LoadRule drop flaw
* post merge cleaning up
* Prohibit assigning concurrent maps into Map-types variables and fields; Fix a race condition in CoordinatorRuleManager; improve logic in DirectDruidClient and ResourcePool
* Enforce that if compute(), computeIfAbsent(), computeIfPresent() or merge() is called on a ConcurrentHashMap, it's stored in a ConcurrentHashMap-typed variable, not ConcurrentMap; add comments explaining get()-before-computeIfAbsent() optimization; refactor Counters; fix a race condition in Intialization.java
* Remove unnecessary comment
* Checkstyle
* Fix getFromExtensions()
* Add a reference to the comment about guarded computeIfAbsent() optimization; IdentityHashMap optimization
* Fix UriCacheGeneratorTest
* Workaround issue with MaterializedViewQueryQueryToolChest
* Strengthen Appenderator's contract regarding concurrency
* Add published segment cache in broker
* Change the DataSegment interner so it's not based on DataSEgment's equals only and size is preserved if set
* Added a trueEquals to DataSegment class
* Use separate interner for realtime and historical segments
* Remove trueEquals as it's not used anymore, change log message
* PR comments
* PR comments
* Fix tests
* PR comments
* Few more modification to
* change the coordinator api
* removeall segments at once from MetadataSegmentView in order to serve a more consistent view of published segments
* Change the poll behaviour to avoid multiple poll execution at same time
* minor changes
* PR comments
* PR comments
* Make the segment cache in broker off by default
* Added a config to PlannerConfig
* Moved MetadataSegmentView to sql module
* Add doc for new planner config
* Update documentation
* PR comments
* some more changes
* PR comments
* fix test
* remove unintentional change, whether to synchronize on lifecycleLock is still in discussion in PR
* minor changes
* some changes to initialization
* use pollPeriodInMS
* Add boolean cachePopulated to check if first poll succeeds
* Remove poll from start()
* take the log message out of condition in stop()
* Adding new web console.
* fixed css
* fix form height
* fix typo
* do import custom react-table css
* added repo field so npm does not complain
* ask travis for node 10
* move indexing-service/src/main/resources/indexer_static into web-console
* fix resource names and paths
* add licenses
* fix exclude file
* add licenses to misc files and tidy up
* remove rebase marker
* fix link
* updated env variable name
* tidy up licenses and surface errors
* cleanup
* remove unused code, fix missing await
* TeamCity does not like the name aux
* add more links to tasks view
* rm pages
* update gitignore
* update readme to be accurate
* make clean script
* removed old console dependancy
* update Jetty routes
* add a comment for welcome files for coordinator
* do not show inital notifaction for now
* renamed overlord console back to console.html
* fix coordinator console
* rename coordinator-console.html to index.html
* * Add few methods about base64 into StringUtils
* Use `java.util.Base64` instead of others
* Add org.apache.commons.codec.binary.Base64 & com.google.common.io.BaseEncoding into druid-forbidden-apis
* Rename encodeBase64String & decodeBase64String
* Update druid-forbidden-apis
* use SqlLifecyle to manage sql execution, add sqlId
* add sql request logger
* fix UT
* rename sqlId to sqlQueryId, sql/time to sqlQuery/time, etc
* add docs and more sql request logger impls
* add UT for http and jdbc
* fix forbidden use of com.google.common.base.Charsets
* fix UT in QuantileSqlAggregatorTest, supressed unused warning of getSqlQueryId
* do not use default method in QueryMetrics interface
* capitalize 'sql' everywhere in the non-property parts of the docs
* use RequestLogger interface to log sql query
* minor bugfixes and add switching request logger
* add filePattern configs for FileRequestLogger
* address review comments, adjust sql request log format
* fix inspection error
* try SuppressWarnings("RedundantThrows") to fix inspection error on ComposingRequestLoggerProvider
* Fix num_replicas count from sys.segments
* Adjust unit test for num_replica > 1
* Pass named arguments instead of passing boolean constants
* Address PR comments
* PR comments
* Use multi-guava version friendly direct executor implementation
* Don't use a singleton
* Fix strict compliation complaints
* Copy Guava's DirectExecutor
* Fix javadoc
* Imports are the devil
* Handoff should ignore segments that are dropped by drop rules
* fix travis-ci
* fix tests
* address comments
* remove line added by accident
* address comments
* add javadoc and logging the full stack trace of exception
* add error message
* Fix issue that tasks failed because of no sink for identifier
* make find sinks to persist run in one callable together with the actual persist work
* Revert "make find sinks to persist run in one callable together with the actual persist work"
This reverts commit a24a2d80ae.
* Broker: Await initialization before finishing startup.
In particular, hold off on announcing the service and starting the
HTTP server until the server view and SQL metadata cache are finished
initializing. This closes a window of time where a Broker could return
partial results shortly after startup.
As part of this, some simplification of server-lifecycle service
announcements. This helps ensure that the two different kinds of
announcements we do (legacy and new-style) stay in sync.
* Remove unused imports.
* Fix NPE in ServerRunnable.
* make logs that are only useful for debugging be at debug level so log volume is much more chill
* info level messages for total merge buffer allocated/free
* more chill compaction logs
* FilteredRequestLogger: Fix start/stop, invalid delegate behavior.
Fixes two bugs:
1) FilteredRequestLogger did not start/stop the delegate.
2) FilteredRequestLogger would ignore an invalid delegate type, and
instead silently substitute the "noop" logger. This was due to a larger
problem with RequestLoggerProvider setup in general; the fix here is
to remove "defaultImpl" from the RequestLoggerProvider interface, and
instead have JsonConfigurator be responsible for creating the
default implementations. It is stricter about things than the old system
was, and is only willing to make a noop logger if it doesn't see any
request logger configs. Otherwise, it'll raise a provision error.
* Remove unneeded annotations.
* autosize processing buffers based on direct memory sizing
* remove oops, more test
* max 1gb autosize buffers, test, start of docs
* fix oops
* revert accidental change
* print buffer size in exception
* change the things
* remove AbstractResourceFilter.isApplicable because it is not, add tests for OverlordResource.doShutdown and OverlordResource.shutdownTasksForDatasource
* cleanup
* Fix missing default config in some calls to coordinator dynamic configs.
The lack of a default config meant that if someone called an API
_without_ a default config before one _with_ a default config, then
the default value would get stuck at null instead of the intended
default value. I noticed this in a cluster where calling /druid/coordinator/v1/config
before a coordinator had fully started up would lead to NPEs during
DruidCoordinatorRuleRunner.
This patch makes the default configs consistent across all calls.
* Remove unnecessary null check.
* Add checkstyle rules about imports and empty lines between members
* Add suppressions
* Update Eclipse import order
* Add empty line
* Fix StatsDEmitter
* Use current coordinator leader instead of cached one (#6551)
Check the response status and throw exception if not OK
* Modify tests
* PR comment
* Add the correct check for status of BytesAccumulatingResponseHandler
* Move the status check into JsonParserIterator so sql query outputs meaningful message on failure
* Fix tests
* Period load/drop/broadcast rules should include the future by default
* address comments
* adjust coordinator console and tweak docs
* address comments
* fix travis-ci
This PR allows to control the fields in `RequestLogEvent`, emitted in `EmittingRequestLogger`. In our case, we want to get rid of the `intervals` fields of the query objects that are a part of `DefaultRequestLogEvent`. They are enormous (thousands of segments) and not useful.
Related to #5522, FYI @a2l007.
* Prohibit some guava collection APIs and use JDK APIs directly
* reset files that changed by accident
* sort codestyle/druid-forbidden-apis.txt alphabetically
* 1. Mysql default transaction isolation is REPEATABLE_READ, treat it as READ_COMMITTED will reduce insert id conflict.
2. Add an index to 'dataSource used end' is work well for the most of scenarios(get recently segments), and it will speed up sync add pending segments in DB.
3. 'select and insert' is not need within transaction.
* Use TaskLockbox.doInCriticalSection instead of synchronized syntax to speed up insert pending segments.
* fix typo for NullPointerException
* Use NodeType enum instead of Strings
* Make NodeType constants uppercase
* Fix CommonCacheNotifier and NodeType/ServerType comments
* Reconsidering comment
* Fix import
* Add a comment to CommonCacheNotifier.NODE_TYPES
Gian Merlino
Justin Borromeo
Clint Wylie
David Glasser
Michael Trelinski
Jihoon Son
Charles Allen
Roman Leventov
Atul Mohan
Furkan KAMACI
Hongze Zhang
Samarth Jain
Himanshu Pandey
Mingming Qiu
Jonathan Wei
Egor Riashin
Surekha
Vadim Ogievetsky
Benedict Jin
Dayue Gao
elloooooo
dongyifeng
hate13
seoeun
David Lim
Faxian Zhao
patelh
