owner stringclasses 14
values | repo stringclasses 14
values | id int64 116k 4.61B | issue_number int32 1 181k | author stringlengths 1 39 | body stringlengths 1 262k | created_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:16:18 | updated_at timestamp[us]date 2000-06-06 02:40:44 2026-06-02 20:22:56 | reactions unknown | author_association stringclasses 5
values |
|---|---|---|---|---|---|---|---|---|---|
ClickHouse | ClickHouse | 705,558,489 | 15,731 | CurtizJ | Test failures are not related to changes. | 2020-10-08T13:12:32 | 2020-10-08T13:12:48 | {} | MEMBER |
ClickHouse | ClickHouse | 705,558,891 | 5,293 | alexey-milovidov | @heyciao We cannot use notice because it goes to the error log, see the details in #15531
What we can do instead is to lower the level of these messages:
```
<Information> TCPHandler: Processed in 0.025 sec.
``` | 2020-10-08T13:13:19 | 2020-10-08T13:13:19 | {} | MEMBER |
ClickHouse | ClickHouse | 705,918,914 | 15,702 | liangli1213 | @BohuTANG , my test has different result, an VIP is floating between mysql1 and mysql2 (mysql1 and mysql1 has bi-direction replication), I made ClickHouse replication through VIP, at first, VIP reside on mysql1, then replication built up, then I switch VIP to mysql2, so do you think ClickHouse replication will continue... | 2020-10-09T01:46:22 | 2020-10-09T01:46:22 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 705,920,194 | 15,265 | liangli1213 | @BohuTANG , the backup user works well for MySQL's slave replication, and I also try manually access master by using mysql -h bdcmysqldev2 -ubackup -p**** from ClickHouse server, it work, I don't know why it raise Access denied issue when ClickHouse use the account, I think you can have a try on your side, it is easy t... | 2020-10-09T01:51:35 | 2020-10-09T01:51:35 | {} | NONE |
ClickHouse | ClickHouse | 705,963,224 | 9,434 | ucasfl | Hi, I'm not very clear about what these two functions do? @alexey-milovidov | 2020-10-09T04:38:34 | 2020-10-09T04:38:34 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 706,153,073 | 15,782 | den-crane | And your question is ? | 2020-10-09T12:30:49 | 2020-10-09T12:30:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,790,279 | 12,136 | vivarum | > @vivarum Hi! Could you please share some cases of usage for this function?
You can have a look on the original issue, probably it can give some background: https://github.com/ClickHouse/ClickHouse/issues/4023 | 2020-10-07T08:48:08 | 2020-10-07T08:48:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 705,137,085 | 13,956 | Akazz | @excitoon
I've looked into `Integration tests flaky check` and there seems to be a problem with cleaning up between tests in `test_multiple_disks` - storage policy `small_jbod_with_external` starts moving parts around before tests expect it to. | 2020-10-07T19:09:05 | 2020-10-07T19:09:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,004,158 | 15,775 | yisany | ch version is updated to 20.3.19.4, but this error is still there.
And I found an interesting phenomenon, although the error exists, but the table of each node has actually been modified successfully | 2020-10-09T06:48:04 | 2020-10-09T06:48:04 | {} | NONE |
ClickHouse | ClickHouse | 706,062,184 | 11,594 | crakjie | @den-crane I'm not sure to understand what you are proposing.
What is `GROUP BY CUBE` compared to `WITH CUBE` ?
In your exemple totals are not differenciated from than null column ? How can you know if it's a totals or a row ? | 2020-10-09T09:01:39 | 2020-10-09T09:01:39 | {} | NONE |
ClickHouse | ClickHouse | 706,148,807 | 15,777 | tavplubix | Fixed in #14887, backported to 20.6.7 | 2020-10-09T12:21:08 | 2020-10-09T12:21:08 | {} | MEMBER |
ClickHouse | ClickHouse | 704,879,007 | 15,231 | alexey-milovidov | There were issues with "unbundled", "split" and "freebsd" builds, all of them are unrelated to the fix and don't matter for release.
Merged. | 2020-10-07T11:43:05 | 2020-10-07T11:43:05 | {} | MEMBER |
ClickHouse | ClickHouse | 705,479,592 | 9,703 | smedoji | Facing the same problem. Also observed that this is happening with only one specific distributed table. The zookeeper is good and I don't see any errors at that instance. | 2020-10-08T10:26:44 | 2020-10-08T10:26:44 | {} | NONE |
ClickHouse | ClickHouse | 705,523,907 | 15,685 | Avogar | I guess it would be better, if you add `SELECT * FROM <table>` in tests and write output in .reference file. To ensure that table contains right data after INSERT | 2020-10-08T12:06:24 | 2020-10-08T12:06:24 | {} | MEMBER |
ClickHouse | ClickHouse | 704,761,546 | 15,473 | filimonov | It maybe related to some issues in concrete librdkafka version, according to the reporter the issue goes away after update to 20.8 | 2020-10-07T07:51:32 | 2020-10-07T07:51:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,844,570 | 15,596 | OmarBazaraa | > it will not impact inserts. It's related to only selects from engine=Distributed
What about the INSERTs to a distributed table?
Wouldn't the distributed table try to choose some healthy replicas as well to insert into? | 2020-10-07T10:26:38 | 2020-10-07T10:39:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,981,485 | 15,237 | KochetovNicolai | looks similar, but does not reproduces easily : https://clickhouse-test-reports.s3.yandex.net/15717/cbf43a870de1733b15edf43178af9e38532f49f1/fuzzer/report.html#fail1
```
2020.10.07 16:21:12.480994 [ 98 ] {952ca530-fa8b-413e-9548-6d5b74b02475} <Debug> executeQuery: (from [::1]:33264) SELECT '922337203685477580.7' IN (... | 2020-10-07T14:38:26 | 2020-10-07T14:38:26 | {} | MEMBER |
ClickHouse | ClickHouse | 705,381,343 | 15,730 | abykuruvilla-dev | I meant you guys have so some kind off caching or marking to make the subsequent identical queries faster. I have seen it happen. But it works if I query it directly and not via a distributed table.
This is not the exact DDL, but resembles our setup. The original table have about 25 columns
The distributed table i... | 2020-10-08T07:16:42 | 2020-10-08T07:21:12 | {} | NONE |
ClickHouse | ClickHouse | 705,481,460 | 15,643 | alexey-milovidov | > 01194_http_query_id
Unrelated but we have to investigate it... | 2020-10-08T10:30:32 | 2020-10-08T10:30:32 | {} | MEMBER |
ClickHouse | ClickHouse | 705,550,862 | 15,581 | alexey-milovidov | Not relevant. | 2020-10-08T12:59:54 | 2020-10-08T12:59:54 | {} | MEMBER |
ClickHouse | ClickHouse | 705,551,514 | 15,531 | alexey-milovidov | > test_match_process_uid_against_data_owner/test.py::test_different_user
Integration test has failed due to changes. Need to be updated. | 2020-10-08T13:00:58 | 2020-10-08T13:00:58 | {} | MEMBER |
ClickHouse | ClickHouse | 704,847,262 | 15,702 | bohutang | A1:
Drop the MaterializeMySQL database and re-create will work again, this perhaps the only consistent way.
A2:
If you do `reset master` on MySQL side, it breaks the master-slave synchronization consistency design, even for mysql.
Think a case:
the Slave has many GTIDs behinds the Master, if we reset master sla... | 2020-10-07T10:32:24 | 2020-10-07T10:33:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 705,236,618 | 15,740 | azat | >Functional stateless tests (release, DatabaseOrdinary) — fail: 1, passed: 2315, skipped: 4
01514_distributed_cancel_query_on_error - #15745 | 2020-10-07T22:59:24 | 2020-10-07T22:59:24 | {} | MEMBER |
ClickHouse | ClickHouse | 705,512,954 | 15,741 | CurtizJ | Found 2 non-critical bugs.
`01508_partition_pruning` - same as #12298
`01482_move_to_prewhere_and_cast` - same as #10694
Both are connected with `PREWHERE`. | 2020-10-08T11:40:57 | 2020-10-08T11:40:57 | {} | MEMBER |
ClickHouse | ClickHouse | 705,548,816 | 15,602 | alexey-milovidov | > Docs check passed | 2020-10-08T12:56:05 | 2020-10-08T12:56:05 | {} | MEMBER |
ClickHouse | ClickHouse | 705,938,762 | 15,775 | den-crane | 19.14.6.12 is not supported anymore.
Check CH logs at shards. Check that DDLWorker is running.
https://github.com/ClickHouse/ClickHouse/issues/8282
https://github.com/ClickHouse/ClickHouse/issues/6639 | 2020-10-09T02:57:04 | 2020-10-09T02:57:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,069,346 | 15,124 | lexand | almost same here
its should be OK
```
CREATE TABLE test_local ON CLUSTER default_cluster
(
date Date,
c1 Int64
)
ENGINE = ReplicatedMergeTree('/clickhouse/tables/{shard}/default/test_local', '{replica}')
PARTITION BY toStartOfMonth(date)
ORDER BY (date);
```
try to make i... | 2020-10-09T09:16:25 | 2020-10-09T09:16:25 | {} | NONE |
ClickHouse | ClickHouse | 706,147,496 | 15,086 | filimonov | ```
SELECT * FROM system.parts WHERE database = 'system' and table='query_log' and part='202009_1_16_3';
```
Also - please grep the logs for: '202009_1_16_3' | 2020-10-09T12:18:07 | 2020-10-09T12:18:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 705,560,448 | 15,497 | alexey-milovidov | @filipecaixeta PVS-Studio has found an error:
https://clickhouse-test-reports.s3.yandex.net/15497/0436e0f8f145c331dcec224d478ac6d80ef60680/pvs_studio_report/pvs-studio-html-report/sources/FunctionsFormatting.h_77.html#ln460 | 2020-10-08T13:16:00 | 2020-10-08T13:16:00 | {} | MEMBER |
ClickHouse | ClickHouse | 705,576,041 | 15,199 | filimonov | Integration tests flaky check (asan) - see: https://github.com/ClickHouse/ClickHouse/pull/15757 | 2020-10-08T13:40:49 | 2020-10-08T13:40:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,054,024 | 14,157 | lostsnow | `create database planet_prod engine=MaterializeMySQL(...)`
> ClickHouse exception, code: 390, host: xxx, port: 8123; Code: 390, e.displayText() = DB::Exception: There is no SHOW CREATE TABLE query for DatabaseMaterializeMySQL (version 20.9.3.45 (official build))
https://paste.ubuntu.com/p/t8W95QkDZp/
`SELECT *... | 2020-10-09T08:44:35 | 2020-10-09T08:53:14 | {} | NONE |
ClickHouse | ClickHouse | 706,148,322 | 13,993 | filimonov | https://github.com/ClickHouse/ClickHouse/pull/15756 ? | 2020-10-09T12:20:01 | 2020-10-09T12:20:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,159,055 | 15,757 | filimonov | Something strange:
```
2020.10.09 13:47:24.263212 Before drop_instance_zk_connections
2020.10.09 13:51:54.287981 After drop_instance_zk_connections
```
4 minutes!
https://github.com/ClickHouse/ClickHouse/blob/f800844da281c6c79f9a721dfffa887a1db0fffd/tests/integration/test_storage_kafka/test.py#L1848-L1850
| 2020-10-09T12:44:15 | 2020-10-09T12:44:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,841,182 | 15,706 | ne1r0n | I also can't drop `mysql` database when it not accessible.
```
DROP DATABASE mysql
Received exception from server (version 20.9.2):
Code: 1000. DB::Exception: Received from clickhouse-dev209:9000. DB::Exception: mysqlxx::ConnectionFailed: Unknown MySQL server host 'mysql' (-2) ((nullptr):0).
```
| 2020-10-07T10:19:57 | 2020-10-07T10:19:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,871,003 | 15,537 | tavplubix | AST fuzzer — Received signal 6 - reported
Performance - unrelated | 2020-10-07T11:25:13 | 2020-10-07T11:25:13 | {} | MEMBER |
ClickHouse | ClickHouse | 704,951,774 | 15,704 | alexey-milovidov | It's expected that only small portion of data is in compact parts - so reading less than one granule is not worth development for compact parts.
Also now we are using "adaptive granularity" - if table rows are heavy, granules will have less amount of rows automatically (it makes reading less than one granule partially... | 2020-10-07T13:53:42 | 2020-10-07T13:53:42 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 705,550,817 | 15,582 | alexey-milovidov | Not relevant. | 2020-10-08T12:59:50 | 2020-10-08T12:59:50 | {} | MEMBER |
ClickHouse | ClickHouse | 705,744,552 | 15,163 | alexey-milovidov | We want to ship this feature https://github.com/ClickHouse/ClickHouse/pull/15601 in 20.10. | 2020-10-08T18:24:45 | 2020-10-08T18:24:53 | {} | MEMBER |
ClickHouse | ClickHouse | 703,162,050 | 15,548 | azat | >Yandex third-party checks (only for Yandex employees)
But this also breaks this check, so there is no point in a separate PR, only if it is ...
>Need manual editing of third-party code, I will do...
@alexey-milovidov is it simpler to fix the code for this change then for #14731 ? | 2020-10-03T20:40:37 | 2020-10-08T21:31:00 | {} | MEMBER |
ClickHouse | ClickHouse | 706,002,357 | 15,763 | alesapin | `test_multiple_disks/test.py::test_start_stop_moves` flaky test | 2020-10-09T06:43:22 | 2020-10-09T06:43:22 | {} | MEMBER |
ClickHouse | ClickHouse | 706,148,539 | 15,756 | filimonov | Does it also fix #13993 ? | 2020-10-09T12:20:31 | 2020-10-09T12:20:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,644,912 | 15,497 | filipecaixeta | @Slach @alexey-milovidov Any more changes needed ? | 2020-10-07T01:50:40 | 2020-10-07T01:50:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,775,833 | 15,684 | alexey-milovidov | @filimonov Let's remove these metrics. | 2020-10-07T08:20:35 | 2020-10-07T08:20:35 | {} | MEMBER |
ClickHouse | ClickHouse | 705,009,959 | 15,670 | den-crane | >2020.09.28 18:59:52.864637 [ 267261 ] {} <Debug> Table B: Removing part from filesystem 1601305200_29_29_0
so the part was removed 2020.09.28
When did you restart ZK ? October 6 ? | 2020-10-07T15:22:18 | 2020-10-07T15:22:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 705,509,089 | 15,127 | ucasfl | > > If implement multiversion storage, do we still need table-level locks for storagememory? It doesn't seem to be needed anymore?
>
> Yes, there will be no table-level locks inside StorageMemory.
> We only need a mutex, under which we will copy the list of blocks.
It seems that we still need a table-level lock ... | 2020-10-08T11:31:53 | 2020-10-08T11:31:53 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 705,553,701 | 15,531 | alexey-milovidov | The proper solution is not to use `Notice` level, but lower the level of some information messages instead. | 2020-10-08T13:04:29 | 2020-10-08T13:04:29 | {} | MEMBER |
ClickHouse | ClickHouse | 705,565,368 | 15,730 | den-crane | >I meant you guys have so some kind off caching or marking to make the subsequent identical queries faster.
>I have seen it happen. But it works if I query it directly and not via a distributed table.
CH does not have a query cache.
You are observing something else. Maybe query plan is different in these cases.... | 2020-10-08T13:23:48 | 2020-10-08T13:23:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 705,834,631 | 15,548 | azat | >@alexey-milovidov is it simpler to fix the code for this change then for #14731 ?
I guess the answer is no, ok, closing | 2020-10-08T21:31:03 | 2020-10-08T21:31:03 | {} | MEMBER |
ClickHouse | ClickHouse | 706,098,900 | 9,434 | alexey-milovidov | `tid()` - simply returns current thread id.
`logTrace('message')` - emits trace log message for each block (`LOG_TRACE`), message argument must be constant expression. | 2020-10-09T10:19:51 | 2020-10-09T10:19:51 | {} | MEMBER |
ClickHouse | ClickHouse | 705,033,051 | 15,715 | KochetovNicolai | Failed build in `concat.cpp`, OOM | 2020-10-07T15:58:36 | 2020-10-07T15:58:36 | {} | MEMBER |
ClickHouse | ClickHouse | 705,815,346 | 15,471 | tavplubix | I have another solution (#15767) for this issue, but it's a bit tricky and probably requires changes in some other places (like `StorageMerge`), so I don't know which one is better. Do you have any suggestions? | 2020-10-08T20:49:05 | 2020-10-08T20:49:05 | {} | MEMBER |
ClickHouse | ClickHouse | 704,510,916 | 15,675 | alesapin | Maybe we can use gcc-10 for unbundled build. | 2020-10-06T19:41:18 | 2020-10-06T19:41:18 | {} | MEMBER |
ClickHouse | ClickHouse | 704,573,643 | 11,594 | den-crane | @crakjie Then nothing. What do you think?
```
psql> \pset null '<null>'
psql> SELECT a,b, count(*)
FROM (select 1 a, 2 b union all select 1 a, 1 b union all select null a, 1 b) t
GROUP BY CUBE (a, b)
ORDER BY a,b;
a | b | count
--------+--------+-------
1 | 1 | 1
1 | 2... | 2020-10-06T21:54:19 | 2020-10-06T21:58:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,952,701 | 15,700 | den-crane | Т.е. наоборот, то что сейчас работает вот такое:
```
create table test(id1 UInt64,id2 UInt64)engine = MergeTree order by tuple();
select * from test where (id1, 0) in (select (1, 0));
Ok.
0 rows in set. Elapsed: 0.002 sec.
```
это баг, **ожидаемое поведение: DB::Exception: Number of columns in section IN does... | 2020-10-07T13:55:00 | 2020-10-07T13:56:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 705,277,054 | 11,568 | bookeezhou | thank you watched the issue , i found you build aarch64 v20.3.9.70 is work right.
but another version is not ok.
so, i close it
在 2020-09-29 15:01:24，"alexey-milovidov" <notifications@github.com> 写道：
@bookeezhou Why close? Is it solved?
—
You are receiving this because you were mentioned.
Reply to this email dir... | 2020-10-08T01:26:11 | 2020-10-08T01:26:11 | {} | NONE |
ClickHouse | ClickHouse | 705,800,819 | 15,701 | tavplubix | AST fuzzer - known issue
Docs check - unrelated
Performance - unrelated | 2020-10-08T20:18:53 | 2020-10-08T20:18:53 | {} | MEMBER |
ClickHouse | ClickHouse | 706,158,247 | 12,513 | filimonov | Stateless example:
```
SELECT min(number) as min_number
FROM numbers(10)
GROUP BY number
having 1
ORDER BY min_number
``` | 2020-10-09T12:42:29 | 2020-10-09T12:42:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,590,696 | 14,414 | azat | 00933_test_fix_extra_seek_on_compressed_cache - https://clickhouse-test-reports.s3.yandex.net/15642/e00008e61dc13cedb8e9f4c8afd6501f9f7b6a6e/fast_test.html#fail1 (refs #13747) | 2020-10-06T22:40:25 | 2020-10-06T22:40:25 | {} | MEMBER |
ClickHouse | ClickHouse | 704,850,257 | 15,704 | alexey-milovidov | Reading less than one granule is needed when granule is large and we don't want to create large blocks in memory.
As far as I remember, it's controlled by `preferred_block_size_bytes` setting.
Set it to lower value, make a table with wide rows, do SELECT *, and you will notice reading less than one granule. | 2020-10-07T10:39:04 | 2020-10-07T10:39:04 | {} | MEMBER |
ClickHouse | ClickHouse | 704,874,460 | 15,691 | 4ertus2 | Perf test errors are not related to PR | 2020-10-07T11:32:50 | 2020-10-07T11:32:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,880,623 | 12,136 | AnaUvarova | > > @vivarum Hi! Could you please share some cases of usage for this function?
>
> You can have a look on the original issue, probably it can give some background: #4023
Thank you. | 2020-10-07T11:46:49 | 2020-10-07T11:46:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 705,010,970 | 15,670 | keliss | Correct. | 2020-10-07T15:23:40 | 2020-10-07T15:23:40 | {} | NONE |
ClickHouse | ClickHouse | 705,549,080 | 15,603 | alexey-milovidov | > Docs check passed | 2020-10-08T12:56:38 | 2020-10-08T12:56:38 | {} | MEMBER |
ClickHouse | ClickHouse | 705,558,060 | 15,646 | alexey-milovidov | What is the scenario - is it when globs expand to different regions?
Is it possible to provide a test? | 2020-10-08T13:11:45 | 2020-10-08T13:11:45 | {} | MEMBER |
ClickHouse | ClickHouse | 705,558,327 | 15,646 | alexey-milovidov | The rest of the code LGTM. | 2020-10-08T13:12:14 | 2020-10-08T13:12:14 | {} | MEMBER |
ClickHouse | ClickHouse | 705,629,579 | 15,702 | bohutang | The `Binlog File` and `Binlog Position` in the .metadata is not used, we only use the `Executed GTID` info.
MySQL master or slave is rebuild, the replication channel will not be affected, this is what we use gtid for.
WARNING:
You can't do anything related to gtid, such as `set global gtid_purged`, or the replciat... | 2020-10-08T15:00:11 | 2020-10-08T15:04:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,087,821 | 14,157 | tavplubix | > `create database planet_prod engine=MaterializeMySQL(...)`
>
> > ClickHouse exception, code: 390, host: xxx, port: 8123; Code: 390, e.displayText() = DB::Exception: There is no SHOW CREATE TABLE query for DatabaseMaterializeMySQL (version 20.9.3.45 (official build))
Fixed in #14849, not merged yet
> `SELECT ... | 2020-10-09T09:55:47 | 2020-10-09T09:55:47 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 704,620,811 | 15,700 | den-crane | Though, may be it's not a bug after all.
select * from test where (id1, 0) in (select toUInt64(1), toUInt64(0));
0 rows in set. Elapsed: 0.001 sec.
Couple years ago we discussed that
( , ) in (select **tuple()**) -- must be forbidden, because it's incorrect. | 2020-10-07T00:22:02 | 2020-10-07T00:22:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,856,971 | 15,717 | KochetovNicolai | Reasons of error are:
1. `StorageValues`, used for `MV`, returned more columns then required in `IStorage::read()`
2. Optimisation for `ARRAY JOIN` added `PROJECT` action for input columns
As long as set of columns known by `ExpressionActions` was less then set of columns used by `ArrayJoinStep` from query plan, `... | 2020-10-07T10:53:58 | 2020-10-07T10:54:39 | {} | MEMBER |
ClickHouse | ClickHouse | 705,091,952 | 15,705 | alexey-milovidov | Waiting only builds. | 2020-10-07T17:42:53 | 2020-10-07T17:42:53 | {} | MEMBER |
ClickHouse | ClickHouse | 705,106,471 | 15,306 | den-crane | fixed: https://github.com/ClickHouse/ClickHouse/pull/15717
clickhouse-client -mn <t.sql
2020-09-25 11:00:00 40
2020-09-25 11:00:00 12
2020-09-25 11:00:00 54
2020-09-25 11:00:00 13
ClickHouse client version 20.10.1.4853.
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 2... | 2020-10-07T18:10:21 | 2020-10-07T18:11:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 706,152,619 | 15,778 | den-crane | duplicate: https://github.com/ClickHouse/ClickHouse/issues/13634
fixed: https://github.com/ClickHouse/ClickHouse/pull/14646
20.3 is out support.
This bug is fixed in the latest stable versions 20.7 20.8 20.9. | 2020-10-09T12:29:51 | 2020-10-09T12:29:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,726,595 | 15,684 | filimonov | I think we either need to obsolete / remove those metrics, either split them in
`BackgroundProcessingPoolBytesAllocated` & `BackgroundProcessingPoolByteFreed` etc. | 2020-10-07T06:33:59 | 2020-10-07T06:33:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,901,817 | 15,701 | tavplubix | Integration tests (release) - test_always_fetch_merged/test.py::test_replica_always_download - flacky
Performance - unrelated
Stress test (undefined) — Server failed to start - not a bug, seems like we need more retries in alive check | 2020-10-07T12:29:23 | 2020-10-07T12:29:23 | {} | MEMBER |
ClickHouse | ClickHouse | 704,967,124 | 15,702 | liangli1213 | @BohuTANG , let me explain more clearly, here is the topology of the replication chain:
`MySQL Master -> MySQL Slave -> ClickHouse`
What I did on MySQL Slave is:
1. reset master
2. set global gtid_purged = 'xxxxxx:1-xxxx'
After this, MySQL Master to Slave's replication can work, but I notice MySQL Slave to C... | 2020-10-07T14:16:53 | 2020-10-07T14:16:53 | {} | NONE |
ClickHouse | ClickHouse | 705,271,923 | 8,417 | n87 | Another, shorter way to reproduce the error message:
```
CREATE TABLE j (id UInt8, val UInt8) Engine = Join(ALL, INNER, id);
SELECT * FROM (SELECT 0 id, 1 val) _ JOIN j USING id;
Received exception from server (version 20.9.2):
Code: 10. DB::Exception: Received from localhost:9000. DB::Exception: Not found colum... | 2020-10-08T01:05:19 | 2020-10-08T01:05:19 | {} | NONE |
ClickHouse | ClickHouse | 705,844,262 | 15,768 | tavplubix | `max_replicated_logs_to_keep` is 10000 by default in 20.6 | 2020-10-08T21:55:27 | 2020-10-08T21:55:27 | {} | MEMBER |
ClickHouse | ClickHouse | 706,065,155 | 10,601 | compasses | > Honest MergeJoin logic removed in favour of #11596. Current prototype shows that it's hard to parallelyze MergeJoin of big tables without buckets.
@4ertus2 Hi, what's your mean with 'Honest MergeJoin logic removed in favour of #11596'? The merge join algo will be removed when the grace hash join algo ready? thanks... | 2020-10-09T09:07:44 | 2020-10-09T09:07:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,779,374 | 15,700 | Fallenyasha | Указание пар значений без кортежа работает без ошибок
select * from test where (id1, id2) in (select filter1, filter2 from filter)
Просто странное поведение получается, что позволяет указывать кортежи, но только в случае, если не используется индексированное поле, причем не в правой части, а в левой. | 2020-10-07T08:27:20 | 2020-10-07T08:27:20 | {} | NONE |
ClickHouse | ClickHouse | 705,003,537 | 15,199 | filimonov | Test failures are relevant, will fix them soon | 2020-10-07T15:12:24 | 2020-10-07T15:12:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 705,253,431 | 15,726 | tavplubix | Builds are broken in 20.8, tests of this bugfix in other branches are ok | 2020-10-07T23:57:23 | 2020-10-07T23:57:23 | {} | MEMBER |
ClickHouse | ClickHouse | 705,464,726 | 15,745 | alexey-milovidov | > Integration tests (asan)
Unrelated. | 2020-10-08T09:57:57 | 2020-10-08T09:57:57 | {} | MEMBER |
ClickHouse | ClickHouse | 705,951,441 | 15,767 | zhang2014 | 👍 I prefer this solution. But it cause unexpected behavior when only string types exist in the table(because string type `haveMaximumSizeOfValue` method returns false)? | 2020-10-09T03:51:17 | 2020-10-09T03:51:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,909,495 | 15,703 | dependabot-preview[bot] | OK, I won't notify you again about this release, but will get in touch when a new version is available. If you'd rather skip all updates until the next major or minor version, let me know by commenting `@dependabot ignore this major version` or `@dependabot ignore this minor version`.
If you change your mind, just re-... | 2020-10-07T12:43:33 | 2020-10-07T12:43:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,949,559 | 15,700 | den-crane | Я понял, правильное поведение именно выдавать ошибку **"DB::Exception: Number of columns in section IN doesn't match. 2 at left, 1 at right."**
Такой запрос должен падать, начиная с
https://github.com/ClickHouse/ClickHouse/blob/master/docs/en/whats-new/changelog/2018.md#backward-incompatible-changes-backward-inc... | 2020-10-07T13:50:22 | 2020-10-07T13:50:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,981,863 | 15,717 | KochetovNicolai | Crash in fuzzer the same as in #15237 | 2020-10-07T14:39:02 | 2020-10-07T14:39:02 | {} | MEMBER |
ClickHouse | ClickHouse | 705,320,703 | 15,702 | liangli1213 | @BohuTANG @zhang2014 , I have done a test, mysqldb is a virtual IP, mysql1 and mysql2 are 2 physical server, at first mysqldb is on mysql2, then I have made MySQL to ClickHouse replication by using mysqldb as MySQL name, then I failover mysqldb to mysql1, obviously the Binlog File and Binlog Position has changed, it tr... | 2020-10-08T04:25:38 | 2020-10-08T04:25:38 | {} | NONE |
ClickHouse | ClickHouse | 705,509,895 | 15,755 | alesapin | I've checked tests locally, so I'll merge. | 2020-10-08T11:33:47 | 2020-10-08T11:33:47 | {} | MEMBER |
ClickHouse | ClickHouse | 705,550,920 | 15,580 | alexey-milovidov | Not relevant. | 2020-10-08T13:00:01 | 2020-10-08T13:00:01 | {} | MEMBER |
ClickHouse | ClickHouse | 705,946,063 | 15,610 | chenkaiyin1201 | Hello,
I have kind of the same issue with Clickhouse-Copier in version 20.9.2.20
| 2020-10-09T03:26:51 | 2020-10-09T03:26:51 | {} | NONE |
ClickHouse | ClickHouse | 706,112,821 | 14,113 | johnskopis | I *think* I have addressed all the feedback. Please take a look.
Also, I'm not sure what to do about:
flake: Just delete it?
broken clang11 build -- it doesn't like the naming convention what should I do instead? | 2020-10-09T10:53:25 | 2020-10-09T10:53:25 | {} | NONE |
ClickHouse | ClickHouse | 704,769,078 | 15,700 | Fallenyasha | Такой синтаксис используется для проверки пар значений, например
select * from test where (id1, id2) in (select (filter1, filter2) from filter) | 2020-10-07T08:07:06 | 2020-10-07T08:07:06 | {} | NONE |
ClickHouse | ClickHouse | 704,814,608 | 15,685 | vivarum | - Fixed issue description
- Made style fixes for failed style checks
- Added one more checked statement into tests (to cover all places in code changed) | 2020-10-07T09:31:28 | 2020-10-07T09:31:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 704,897,356 | 15,704 | hustnn | @alexey-milovidov May I know why some reader (like `MergeTreeReaderCompact`) does not support reading of incomplete granule? | 2020-10-07T12:20:37 | 2020-10-07T12:20:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 705,422,853 | 11,755 | xmariachi | Hi @4ertus2 ,
Thanks.
Is the error confirmed to still happen in a version that already contains the #15242 fix?
| 2020-10-08T08:41:33 | 2020-10-08T08:41:33 | {} | NONE |
ClickHouse | ClickHouse | 705,603,509 | 15,172 | vdimir | Does it require setting appropriate ranges for `Aws::S3::Model::GetObjectRequest` and downloading, them in parallel and concatenating? Should it be implemented inside `ReadBufferFromS3` or on top of it?
https://github.com/ClickHouse/ClickHouse/blob/bf36ffde33f52855a315cc777bfc861e3b67e1bc/src/IO/ReadBufferFromS3.cpp... | 2020-10-08T14:20:21 | 2020-10-08T14:21:14 | {} | MEMBER |
ClickHouse | ClickHouse | 705,656,214 | 15,757 | filimonov | Hm... Strange, lot of unrelated things blew up, timing looks ok now:
```
2020.10.08 18:14:36.836185 [ 37 ] {2c98244f-06a5-4c1f-b421-cd105f954725} <Debug> executeQuery: (from 172.18.0.1:48722) CREATE TABLE test.view (key UInt64, value String) ENGINE = ReplicatedMergeTree('/clickhouse/kafkatest/tables/no_holes_when_wri... | 2020-10-08T15:43:09 | 2020-10-08T15:43:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 708,625,663 | 14,985 | filimonov | Next iteration: #15986 | 2020-10-14T19:54:19 | 2020-10-14T19:54:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 708,157,696 | 15,951 | bohutang | It seems like a GTID Sets shrinking issue, we will try to fix it. | 2020-10-14T05:01:45 | 2020-10-14T05:01:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 708,179,267 | 15,793 | alex-zaitsev | TSV is already done, I guess: https://github.com/ClickHouse/ClickHouse/pull/14586 | 2020-10-14T06:06:29 | 2020-10-14T06:06:29 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.