* plan consistently with either UnionDataSource or UnionQuery for decoupled mode
* expose errors
* move decoupled related setting from PlannerConfig to QueryContexts
This method was missing some required synchronization. This patch also
adds GuardedBy annotations to historicalServers and realtimeServers, which
would have caught it.
This patch supports sorting segments by non-time columns (added in #16849) to MSQ compaction.
Specifically, if `forceSegmentSortByTime` is set in the data schema, either via the user-supplied
compaction config or in the inferred schema, the following steps are taken:
- Skip adding `__time` explicitly as the first column to the dimension schema since it already comes
as part of the schema
- Ensure column mappings propagate `__time` in the order specified by the schema
- Set `forceSegmentSortByTime` in the MSQ context.
Changes
---------
- Add Overlord runtime property `druid.indexer.tasklock.batchAllocationReduceMetadataIO`
- Setting this flag to true (default value) allows the Overlord to fetch only necessary segment
payloads during segment allocation
- Setting this flag to false restores original segment allocation behaviour
This change is to emit following metrics as part of GroupByStatsMonitor monitor,
mergeBuffer/used -> Number of merge buffers used.
mergeBuffer/acquisitionTimeNs -> Total time required to acquire merge buffer.
mergeBuffer/acquisition -> Number of queries that acquired a batch of merge buffers.
groupBy/spilledQueries -> Number of queries that spilled onto the disk.
groupBy/spilledBytes-> Spilled bytes on the disk.
groupBy/mergeDictionarySize -> Size of the merging dictionary.
All JDK 8 based CI checks have been removed.
Images used in Dockerfile(s) have been updated to Java 17 based images.
Documentation has been updated accordingly.
* SQL syntax error should target USER persona
* * revert change to queryHandler and related tests, based on review comments
* * add test
* * add `ingest/notices/queueSize` and `ingest/pause/time` to statsd emitter
* * add taskStatus dimension to `service/heartbeat` metric
* Revert "* add taskStatus dimension to `service/heartbeat` metric"
This reverts commit cfb02a2813.
changes:
* fix issue when merging projections from multiple-incremental persists which was hoping that some 'dim conversion' buffers were not closed, but they already were (by the merging iterator). fix involves selectively persisting these conversion buffers to temp files in the segment write out directory and mapping them and tying them to the segment level closer so that they are available after the lifetime of the parent merger
* modify auto column serializers to use segment write out directory for temp files instead of java.io.tmpdir
* fix queryable index projection to not put the time-like column as a dimension, instead only adding it as __time
* use smoosh for temp files so can safely write any Serializer to a temp smoosh
Zoltan Haindrich
Vadim Ogievetsky
Charles Smith
Vishesh Garg
Gian Merlino
Clint Wylie
Akshat Jain
Kashif Faraz
Rishabh Singh
Adarsh Sanjeev
Katya Macedo
Vivek Dhiman
Adithya Chakilam
zachjsh