fdb-record-layer

fdb-record-layer copied to clipboard

Sometimes the test times out: ``` 2024-09-06T07:50:36.194640Z FDBLuceneQueryTest > threadedLuceneScanDoesntBreakPlannerAndSearch-PoolThreadCount=1 STARTED 2024-09-06T07:55:41.248346Z FDBLuceneQueryTest > threadedLuceneScanDoesntBreakPlannerAndSearch-PoolThreadCount=1 FAILURE (305,006ms) FDBLuceneQueryTest > threadedLuceneScanDoesntBreakPlannerAndSearch(Integer) > com.apple.foundationdb.record.lucene.FDBLuceneQueryTest.threadedLuceneScanDoesntBreakPlannerAndSearch(Integer)[1] FAILED java.util.concurrent.TimeoutException: threadedLuceneScanDoesntBreakPlannerAndSearch(java.lang.Integer) timed out after 5 minutes...

ScottDugas

This issue is to exclusively use `IndexKeyValueToPartialRecord` in the heuristic planner and use a `RecordConstructorValue` or similar to compute the partial record that the `RecordQueryCoveringIndexScanPlan` can then return.

normen662

In #2858 we kept some code in `ScanWithMatchCandidate` that allowed us to create compatible plans when not planning mixed order ordering requirements. This should be removed once downstream is adapted.

normen662

For example: ``` LuceneIndexMaintenanceTest > concurrentStoreTest(boolean, boolean, boolean, int, long) > com.apple.foundationdb.record.lucene.LuceneIndexMaintenanceTest.concurrentStoreTest(boolean, boolean, boolean, int, long)[2] FAILED java.util.concurrent.CompletionException: java.lang.RuntimeException: Failed merge at iteration 16 at java.base/java.util.concurrent.CompletableFuture.encodeThrowable(CompletableFuture.java:314) at java.base/java.util.concurrent.CompletableFuture.completeThrowable(CompletableFuture.java:319) at java.base/java.util.concurrent.CompletableFuture$AsyncSupply.run(CompletableFuture.java:1702)...

ScottDugas

Add comparison keys to plan string for in-union

1

The `comparisonKeyFunction` is currently not included in the plan representation of `InUnion` plans. This can make it hard to inspect bugs that could be a result of an in-union improperly...

alecgrieser

LuceneOptimizedStoredFieldsFormatTest > testNumericField cannot find stored fields sometimes

2

This happens sometimes (with the same seed): ``` LuceneOptimizedStoredFieldsFormatTest > testNumericField FAILED com.apple.foundationdb.record.RecordCoreStorageException: Could not find stored fields at __randomizedtesting.SeedInfo.seed([C185081D42F0F43C:91AB1306856D114E]:0) at app//com.apple.foundationdb.record.lucene.directory.FDBDirectory.readStoredFields(FDBDirectory.java:539) at app//com.apple.foundationdb.record.lucene.codec.LuceneOptimizedStoredFieldsReader.visitDocument(LuceneOptimizedStoredFieldsReader.java:76) at app//org.apache.lucene.index.CodecReader.document(CodecReader.java:84) at app//org.apache.lucene.index.IndexReader.document(IndexReader.java:352) at app//org.apache.lucene.index.BaseStoredFieldsFormatTestCase.testNumericField(BaseStoredFieldsFormatTestCase.java:297)...

ScottDugas

I ran into a planner `VerifyException` in a test query when trying to determine the ordering property on an intersection on values plan. I was able to reproduce it fairly...

alecgrieser

Here is an example of a self-join query that reuses the `TypeFilter` twice (because it scans the same table twice): ```sql select * from t1 as x, t1 as y...

hatyo

Currently, the rebalancing code uses a single transaction to iterate through the grouping keys and rebalance the first partition in need in each group. This may cause a transaction timeout,...

jjezra

We generally support read-your-writes semantics, and so uncommitted data is visible within the context of the transaction that writes it. However, this is broken for version indexes, where uncommitted data...

alecgrieser