ClickHouse: Segmentation fault when a query to a distributed table is merging data from shards

we should have a cluster with both localhost and remote nodes;

test_cluster_two_shards is enough

we should have LowCardinality on initiator and different String and LowCardinality(String) on remote nodes;

yes

two level group by should be activated;

not necessary

distributed_aggregation_memory_efficient should be set.

yes

Here is what I got:

create table data (key String) Engine=Memory();
create table dist (key LowCardinality(String)) engine=Distributed(test_cluster_two_shards, currentDatabase(), data);
insert into data values ('foo');
set distributed_aggregation_memory_efficient=1;
select * from dist group by key;

2020.05.30 05:58:23.367133 [ 36 ] {2446f1e1-bbce-484d-8b9a-68efb7d4ec6a} <Trace> Aggregator: Aggregating
2020.05.30 05:58:23.367455 [ 36 ] {2446f1e1-bbce-484d-8b9a-68efb7d4ec6a} <Trace> Aggregator: Aggregation method: key_string
2020.05.30 05:58:23.367654 [ 36 ] {2446f1e1-bbce-484d-8b9a-68efb7d4ec6a} <Trace> Aggregator: Aggregated. 1 to 1 rows (from 0.000 MiB) in 0.000 sec. (3957.669 rows/sec., 0.045 MiB/sec.)
2020.05.30 05:58:23.367914 [ 36 ] {2446f1e1-bbce-484d-8b9a-68efb7d4ec6a} <Trace> Aggregator: Merging aggregated data
2020.05.30 05:58:23.368707 [ 28 ] {2446f1e1-bbce-484d-8b9a-68efb7d4ec6a} <Information> executeQuery: Read 1 rows, 12.00 B in 0.003 sec., 325 rows/sec., 3.81 KiB/sec.
2020.05.30 05:58:23.368748 [ 32 ] {7d5ca935-0738-470c-a4a0-edd0b0114af7} <Trace> Aggregator: Merging partially aggregated blocks (bucket = -1).
2020.05.30 05:58:23.369045 [ 28 ] {2446f1e1-bbce-484d-8b9a-68efb7d4ec6a} <Debug> MemoryTracker: Peak memory usage (for query): 5.55 KiB.
2020.05.30 05:58:23.369600 [ 28 ] {2446f1e1-bbce-484d-8b9a-68efb7d4ec6a} <Trace> virtual DB::MergingAndConvertingBlockInputStream::~MergingAndConvertingBlockInputStream(): Waiting for threads to finish
2020.05.30 05:58:23.369831 [ 28 ] {2446f1e1-bbce-484d-8b9a-68efb7d4ec6a} <Information> TCPHandler: Processed in 0.005 sec.
2020.05.30 05:58:23.370575 [ 37 ] {} <Fatal> BaseDaemon: ########################################
2020.05.30 05:58:23.370918 [ 37 ] {} <Fatal> BaseDaemon: (version 19.14.9.12 (official build)) (from thread 32) Received signal Segmentation fault (11).
2020.05.30 05:58:23.371139 [ 37 ] {} <Fatal> BaseDaemon: Address: 0x1 Access: read. Address not mapped to object.
2020.05.30 05:58:23.371417 [ 37 ] {} <Fatal> BaseDaemon: Stack trace: 0x55555e8148ca 0x55555cd4a15e 0x55555ccbbbd3 0x55555cc56cac 0x55555cc5763d 0x55555940d0ce 0x55555940d6de 0x55555940ab7c 0x55555f078d20 0x7ffff7bbd6db 0x7ffff74da88f
2020.05.30 05:58:23.406432 [ 37 ] {} <Fatal> BaseDaemon: 3. 0x55555e8148ca CityHash_v1_0_2::CityHash64(char const*, unsigned long) /usr/bin/clickhouse
2020.05.30 05:58:23.406997 [ 37 ] {} <Fatal> BaseDaemon: 4. 0x55555cd4a15e void DB::Aggregator::mergeStreamsImplCase<false, DB::AggregationMethodString<HashMapTable<StringRef, HashMapCellWithSavedHash<StringRef, char*, StringRefHash64, HashTableNoState>, StringRefHash64, HashTableGrower<8ul>, Allocator<true, true> > >, HashMapTable<StringRef, HashMapCellWithSavedHash<StringRef, char*, StringRefHash64, HashTableNoState>, StringRefHash64, HashTableGrower<8ul>, Allocator<true, true> > >(DB::Block&, DB::Arena*, DB::AggregationMethodString<HashMapTable<StringRef, HashMapCellWithSavedHash<StringRef, char*, StringRefHash64, HashTableNoState>, StringRefHash64, HashTableGrower<8ul>, Allocator<true, true> > >&, HashMapTable<StringRef, HashMapCellWithSavedHash<StringRef, char*, StringRefHash64, HashTableNoState>, StringRefHash64, HashTableGrower<8ul>, Allocator<true, true> >&, char*) const /usr/bin/clickhouse
2020.05.30 05:58:23.407371 [ 37 ] {} <Fatal> BaseDaemon: 5. 0x55555ccbbbd3 DB::Aggregator::mergeBlocks(std::__cxx11::list<DB::Block, std::allocator<DB::Block> >&, bool) /usr/bin/clickhouse
2020.05.30 05:58:23.407495 [ 37 ] {} <Fatal> BaseDaemon: 6. 0x55555cc56cac DB::MergingAggregatedMemoryEfficientBlockInputStream::mergeThread(std::shared_ptr<DB::ThreadGroupStatus>) /usr/bin/clickhouse
2020.05.30 05:58:23.407529 [ 37 ] {} <Fatal> BaseDaemon: 7. 0x55555cc5763d ? /usr/bin/clickhouse
2020.05.30 05:58:23.407565 [ 37 ] {} <Fatal> BaseDaemon: 8. 0x55555940d0ce ThreadPoolImpl<ThreadFromGlobalPool>::worker(std::_List_iterator<ThreadFromGlobalPool>) /usr/bin/clickhouse
2020.05.30 05:58:23.407629 [ 37 ] {} <Fatal> BaseDaemon: 9. 0x55555940d6de ThreadFromGlobalPool::ThreadFromGlobalPool<ThreadPoolImpl<ThreadFromGlobalPool>::scheduleImpl<void>(std::function<void ()>, int, std::optional<unsigned long>)::{lambda()#3}>(ThreadPoolImpl<ThreadFromGlobalPool>::scheduleImpl<void>(std::function<void ()>, int, std::optional<unsigned long>)::{lambda()#3}&&)::{lambda()#1}::operator()() const /usr/bin/clickhouse
2020.05.30 05:58:23.407674 [ 37 ] {} <Fatal> BaseDaemon: 10. 0x55555940ab7c ThreadPoolImpl<std::thread>::worker(std::_List_iterator<std::thread>) /usr/bin/clickhouse
2020.05.30 05:58:23.407716 [ 37 ] {} <Fatal> BaseDaemon: 11. 0x55555f078d20 ? /usr/bin/clickhouse
2020.05.30 05:58:23.407765 [ 37 ] {} <Fatal> BaseDaemon: 12. 0x7ffff7bbd6db start_thread /lib/x86_64-linux-gnu/libpthread-2.27.so
2020.05.30 05:58:23.407806 [ 37 ] {} <Fatal> BaseDaemon: 13. 0x7ffff74da88f __clone /lib/x86_64-linux-gnu/libc-2.27.so
2020.05.30 05:58:23.425761 [ 4 ] {} <Trace> SystemLog (system.trace_log): Flushing system log
2020.05.30 05:58:23.425912 [ 4 ] {} <Debug> SystemLog (system.trace_log): Creating new table system.trace_log for TraceLog
2020.05.30 05:58:23.426733 [ 4 ] {} <Information> BackgroundProcessingPool: Create BackgroundProcessingPool with 16 threads
2020.05.30 05:58:23.427539 [ 4 ] {} <Debug> system.trace_log: Loading data parts
2020.05.30 05:58:23.427753 [ 4 ] {} <Debug> system.trace_log: Loaded data parts (0 items)
2020.05.30 05:58:23.441144 [ 4 ] {} <Trace> system.trace_log: Renaming temporary part tmp_insert_202005_1_1_0 to 202005_1_1_0.

I think that there may be incorrect types for LowCardinality, i.e. missing getDictionaryType().get(), but need to verify, I looked only briefly for now