Подскажите пожалуйста, кликхаус периодически производит дисконект. Через минуту снова работает

нормально.
В логах вот это:

2021.09.01 09:50:30.125841 [ 2792599 ] {} <Fatal> BaseDaemon: ########################################
2021.09.01 09:50:30.125870 [ 2792599 ] {} <Fatal> BaseDaemon: (version 21.8.4.51 (official build), build id: F8BD60273E32A369A2F51EADA624A8E2B522CDF0) (from thread 2789598) (query_id: ad26bd21-bdcf-4d65-830f-ef71df44b21d) Received signal Segmentation fault (11)
2021.09.01 09:50:30.125888 [ 2792599 ] {} <Fatal> BaseDaemon: Address: NULL pointer. Access: read. Address not mapped to object.
2021.09.01 09:50:30.125911 [ 2792599 ] {} <Fatal> BaseDaemon: Stack trace: 0x8f8fc10 0x1082597c 0x10618c7c 0x10c6f5d0 0x10c77818 0x10c77db1 0x10c77db1 0x10c77db1 0x10c6d7b0 0x10d59cd4 0x10d6c6fb 0x10d72147 0x8fd8f18 0x8fdaabf 0x8fd61ff 0x8fd9ae3 0x7efebaaf9609 0x7efebaa0f103
2021.09.01 09:50:30.125941 [ 2792599 ] {} <Fatal> BaseDaemon: 1. memcpy @ 0x8f8fc10 in /usr/bin/clickhouse
2021.09.01 09:50:30.125966 [ 2792599 ] {} <Fatal> BaseDaemon: 2. DB::ColumnString::insert(DB::Field const&) @ 0x1082597c in /usr/bin/clickhouse
2021.09.01 09:50:30.125985 [ 2792599 ] {} <Fatal> BaseDaemon: 3. DB::MergeTreeSetIndex::checkInRange(std::__1::vector<DB::Range, std::__1::allocator<DB::Range> > const&, std::__1::vector<std::__1::shared_ptr<DB::IDataType const>, std::__1::allocator<std::__1::shared_ptr<DB::IDataType const> > > const&) const @ 0x10618c7c in /usr/bin/clickhouse
2021.09.01 09:50:30.126009 [ 2792599 ] {} <Fatal> BaseDaemon: 4. DB::KeyCondition::checkInHyperrectangle(std::__1::vector<DB::Range, std::__1::allocator<DB::Range> > const&, std::__1::vector<std::__1::shared_ptr<DB::IDataType const>, std::__1::allocator<std::__1::shared_ptr<DB::IDataType const> > > const&) const @ 0x10c6f5d0 in /usr/bin/clickhouse
2021.09.01 09:50:30.126021 [ 2792599 ] {} <Fatal> BaseDaemon: 5. ? @ 0x10c77818 in /usr/bin/clickhouse
2021.09.01 09:50:30.126040 [ 2792599 ] {} <Fatal> BaseDaemon: 6. ? @ 0x10c77db1 in /usr/bin/clickhouse
2021.09.01 09:50:30.126076 [ 2792599 ] {} <Fatal> BaseDaemon: 7. ? @ 0x10c77db1 in /usr/bin/clickhouse
2021.09.01 09:50:30.126082 [ 2792599 ] {} <Fatal> BaseDaemon: 8. ? @ 0x10c77db1 in /usr/bin/clickhouse
2021.09.01 09:50:30.126104 [ 2792599 ] {} <Fatal> BaseDaemon: 9. DB::KeyCondition::checkInRange(unsigned long, DB::FieldRef const*, DB::FieldRef const*, std::__1::vector<std::__1::shared_ptr<DB::IDataType const>, std::__1::allocator<std::__1::shared_ptr<DB::IDataType const> > > const&, bool, BoolMask) const @ 0x10c6d7b0 in /usr/bin/clickhouse
2021.09.01 09:50:30.126117 [ 2792599 ] {} <Fatal> BaseDaemon: 10. DB::MergeTreeDataSelectExecutor::markRangesFromPKRange(std::__1::shared_ptr<DB::IMergeTreeDataPart const> const&, std::__1::shared_ptr<DB::StorageInMemoryMetadata const> const&, DB::KeyCondition const&, DB::Settings const&, Poco::Logger*) @ 0x10d59cd4 in /usr/bin/clickhouse
2021.09.01 09:50:30.126139 [ 2792599 ] {} <Fatal> BaseDaemon: 11. ? @ 0x10d6c6fb in /usr/bin/clickhouse
2021.09.01 09:50:30.126146 [ 2792599 ] {} <Fatal> BaseDaemon: 12. ? @ 0x10d72147 in /usr/bin/clickhouse
2021.09.01 09:50:30.126169 [ 2792599 ] {} <Fatal> BaseDaemon: 13. ThreadPoolImpl<ThreadFromGlobalPool>::worker(std::__1::__list_iterator<ThreadFromGlobalPool, void*>) @ 0x8fd8f18 in /usr/bin/clickhouse
2021.09.01 09:50:30.126185 [ 2792599 ] {} <Fatal> BaseDaemon: 14. ThreadFromGlobalPool::ThreadFromGlobalPool<void ThreadPoolImpl<ThreadFromGlobalPool>::scheduleImpl<void>(std::__1::function<void ()>, int, std::__1::optional<unsigned long>)::'lambda0'()>(void&&, void ThreadPoolImpl<ThreadFromGlobalPool>::scheduleImpl<void>(std::__1::function<void ()>, int, std::__1::optional<unsigned long>)::'lambda0'()&&...)::'lambda'()::operator()() @ 0x8fdaabf in /usr/bin/clickhouse
2021.09.01 09:50:30.126195 [ 2792599 ] {} <Fatal> BaseDaemon: 15. ThreadPoolImpl<std::__1::thread>::worker(std::__1::__list_iterator<std::__1::thread, void*>) @ 0x8fd61ff in /usr/bin/clickhouse

и как часто? раз в пару часов?

примерно да

линукс наверно сам убивает долговисящие коннекты. максимум часа 4 может провисеть по идее при дефолтных настройках. зависит от дистриба

падает КХ. и рестартует через минуту кроном. что делал ad26bd21-bdcf-4d65-830f-ef71df44b21d, грепайте лог

Спасибо за наводку

лучше сразу в github открыть issue. 21.8 очень стабильная и всем будет интересно починить это