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 | 683,185,972 | 14,093 | filimonov | > `min_merge_bytes_to_use_direct_io` in MergeTreeSettings.
OK. So actually aio = direct io here (was messed by naming). BTW - did you ever try/consider `posix_fadvise` instead o_direct? | 2020-08-28T23:17:36 | 2020-08-28T23:17:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,292,479 | 14,227 | amosbird | Your query is invalid (without a where clause). And I cannot reproduce it with current master. | 2020-08-29T13:48:25 | 2020-08-29T13:48:25 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 683,355,055 | 14,235 | bohutang | what's the MySQL side SQL looks like?
| 2020-08-29T23:46:02 | 2020-08-29T23:46:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,405,297 | 14,251 | bohutang | From cnf look like nothing weird.
`4. DB::MySQLReplication::TableMapEvent::parseImpl(DB::ReadBuffer&) @ 0x1737bac9 in /usr/bin/clickhouse`
means there is some writes on the MySQL side, TableMap is the prefix event for writing/updating.
Maybe you can repeat your operation and check.
| 2020-08-30T10:51:12 | 2020-08-30T10:51:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,487,055 | 14,253 | bohutang | Ping @alexey-milovidov for review | 2020-08-30T23:45:23 | 2020-08-30T23:45:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,582,621 | 14,232 | zhang2014 | Compatible with the MySQL Identifier(https://dev.mysql.com/doc/refman/8.0/en/identifiers.html) | 2020-08-31T06:17:59 | 2020-08-31T06:17:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,644,820 | 14,260 | Slach | @azat thanks a lo for reproduce test case | 2020-08-31T08:32:30 | 2020-08-31T08:32:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,684,721 | 14,207 | jjtjiang | yes, it was killed by oomkiller. thanks | 2020-08-31T09:58:52 | 2020-08-31T09:58:52 | {} | NONE |
ClickHouse | ClickHouse | 683,680,538 | 14,045 | radiatorius | Here is a log from version with asan (this is sufficient to identify the issue) ?
`alter table lake.items update state='test3' where id=1 ` - mutation that caused the issue.
**clickhouse-server.log**
```
> 2020.08.31 08:48:02.647391 [ 13711 ] {} <Fatal> BaseDaemon: 18. /build/obj-x86_64-linux-gnu/../src/Data... | 2020-08-31T09:49:26 | 2020-08-31T10:42:13 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 683,724,965 | 14,261 | alexey-milovidov | See #13252 | 2020-08-31T11:34:32 | 2020-08-31T11:34:32 | {} | MEMBER |
ClickHouse | ClickHouse | 683,187,752 | 14,201 | ramazanpolat | Let's say there are more than 50 clients each has a different software inserting data to my table (KafkaEngine).
In this scenario, each and every one of the client software has to notify me about unsuccessful inserts and I have to ask for the data which they couldn't insert to see what is wrong with the data. Or mayb... | 2020-08-28T23:24:20 | 2020-08-28T23:24:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,189,972 | 14,201 | filimonov | > Let's say there are more than 50 clients each has a different software inserting data to my table (KafkaEngine).
> In this scenario, each and every one of the client software has to notify me about unsuccessful inserts and I have to ask for the data which they couldn't insert to see what is wrong with the data. Or m... | 2020-08-28T23:33:23 | 2020-08-28T23:33:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,209,983 | 14,231 | liangli1213 | @zhang2014 @BohuTANG | 2020-08-29T01:09:18 | 2020-08-29T01:09:18 | {} | NONE |
ClickHouse | ClickHouse | 683,275,518 | 14,234 | amosbird | > so, I suspect that the ReplicatedMergeTree engine doesn't support multi-disk policies.
It supports storage policy perfectly.
Did you setup `/data3/clickhouse/` yet? Does it have the right access permission? | 2020-08-29T11:10:36 | 2020-08-29T11:10:36 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 683,310,922 | 14,227 | amosbird | Please share a minimal reproduceable case starting with `create table; insert into; ...` | 2020-08-29T16:17:04 | 2020-08-29T16:17:04 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 683,326,068 | 14,227 | den-crane | it's actually duplicate of https://github.com/ClickHouse/ClickHouse/issues/13634 | 2020-08-29T18:25:31 | 2020-08-29T18:25:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,726,523 | 13,891 | BayoNet | Тесты почему-то не запустились. Локальная сборка ошибок не показала, поэтому мёржу. | 2020-08-31T11:38:28 | 2020-08-31T11:38:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,188,490 | 14,208 | filimonov | `appears to be already active. If you're sure it's not, try again in a minute or remove znode` it may mean some misconfiguratio (check the macros) | 2020-08-28T23:27:14 | 2020-08-28T23:27:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,305,051 | 14,225 | alexey-milovidov | @Akazz has been reviewed yesterday and said that it's all good. | 2020-08-29T15:26:24 | 2020-08-29T15:26:24 | {} | MEMBER |
ClickHouse | ClickHouse | 683,319,040 | 14,227 | calebeaires | Here it is
### ClickHouse version = 20.6.4.44
1. Create
```sql
drop table my_table;
create table my_table (
dt Nullable(Date),
name Nullable(String)
) engine MergeTree order by tuple();
```
---
2. Insert data: dt field will get TODAY
```sql
insert into my_table (name, dt) values ('car', today());... | 2020-08-29T17:24:48 | 2020-08-29T17:25:40 | {} | NONE |
ClickHouse | ClickHouse | 683,681,897 | 14,232 | zhang2014 | @alexey-milovidov
How to get `AST fuzzer — Received signal 6` server logs? | 2020-08-31T09:52:33 | 2020-08-31T09:52:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,724,111 | 14,261 | alexey-milovidov | The data in parts is cleared according to the TTL logic but parts theirself are not removed even if they are empty.
This is the current behaviour. | 2020-08-31T11:32:23 | 2020-08-31T11:32:23 | {} | MEMBER |
ClickHouse | ClickHouse | 683,403,153 | 14,251 | liangli1213 | Do you think it has some relationship I set slave's data dir on /nfs drive? I don't know which sql cause the issue, this is not manually triggered.
Master my.cnf:
```
[root@bdcmysqldev2 mysql]# cat my.cnf
[mysqld]
basedir = /opt/mysql
datadir = /data/mysql5.7.21_db
socket = /var/lib/mysql/mysql.sock
pid-file =... | 2020-08-30T10:27:18 | 2020-08-30T10:27:18 | {} | NONE |
ClickHouse | ClickHouse | 683,437,944 | 14,255 | dmthuc | I've made the pull request to fix this problem in here https://github.com/ClickHouse/ClickHouse/pull/14256 | 2020-08-30T16:01:48 | 2020-08-30T16:01:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,486,646 | 14,251 | bohutang | > @BohuTANG my slave server crash when I want to reproduce the issue, and here it comes another problem, I found a piece of transaction are missing in slave, so how can I fix it?
>
> Master:
>
> ```
> root@bdcmysqldev1:(none)>show master status\G
> *************************** 1. row ***************************
... | 2020-08-30T23:42:04 | 2020-08-30T23:42:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,694,350 | 13,691 | KochetovNicolai | So, now I can't copy result from client? I've expected that it will be fixed ... | 2020-08-31T10:20:30 | 2020-08-31T10:20:30 | {} | MEMBER |
ClickHouse | ClickHouse | 683,728,193 | 14,040 | alexey-milovidov | Ok, perfect.
PS. AFAIK, MySQL does not have significant performance decrease if you use VARCHAR(65535).
Two bytes are used for length instead of one but it barely affects anything. | 2020-08-31T11:42:36 | 2020-08-31T11:42:36 | {} | MEMBER |
ClickHouse | ClickHouse | 683,251,185 | 13,597 | fsanano | @yanllearnn Thank you! | 2020-08-29T07:35:04 | 2020-08-29T07:35:04 | {} | NONE |
ClickHouse | ClickHouse | 683,291,198 | 14,220 | filimonov | > it doesn't seem to be expected by users
For me it's unexpected that after optimize final it still may be needed to call optimize materialize ttl. Why second iteration / second column processing? (materialize ttl is much cheaper than final, so for me expected behaviour is to include it in final) | 2020-08-29T13:37:18 | 2020-08-29T13:37:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,355,777 | 14,225 | alexey-milovidov | I have reviewed the diff in performance test and found it unrelated. | 2020-08-29T23:58:18 | 2020-08-29T23:58:18 | {} | MEMBER |
ClickHouse | ClickHouse | 683,379,555 | 14,235 | bohutang | MySQL master not write empty transactions to binlog in gereral case.
This is a slave work that to populate the ignore event with special empty transaction, and this empty transaction should be processed.
| 2020-08-30T05:44:53 | 2020-08-30T05:44:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,617,420 | 13,333 | alesapin | Closed in favor to #14225. | 2020-08-31T07:34:32 | 2020-08-31T07:34:32 | {} | MEMBER |
ClickHouse | ClickHouse | 683,631,135 | 14,262 | filimonov | Duplicate of https://github.com/ClickHouse/ClickHouse/issues/14209 | 2020-08-31T08:04:52 | 2020-08-31T08:04:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,733,766 | 14,260 | Slach | @azat error 159 returns only for `remote()` table function, but doesn't return when try to INSERT into real ENGINE=Distributed() table
` kubectl -n test exec chi-test-cluster-for-alerts-default-0-0-0 -n test -- clickhouse-client -mn -h 127.0.0.1 --port=9000 -u default --query="CREATE TABLE default.test ON CLUSTER ... | 2020-08-31T11:56:17 | 2020-08-31T11:56:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,162,798 | 5,237 | filimonov | > @Onlinehead
>
> > Also, as you see in logs, query has been processed in 75.310 sec, but in logs on Sever2 and Server3 I see that each server processed data much faster:
>
> Can you attach full logs (with timestamps)?
Report from may of 2019, there is full log in that message - https://github.com/ClickHouse/C... | 2020-08-28T21:53:59 | 2020-08-28T21:53:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,174,698 | 13,939 | filimonov | >>True, the background schedule pool is overused, we can use other thread pool.
> How about creating separate pool for kafka engine? (this idea was around for awhile
+1
That need to be done. Or scheduled pool can just run the job by the thread from the main pool, and be released immediately (that way threads fro... | 2020-08-28T22:34:11 | 2020-08-28T22:34:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,223,737 | 11,512 | dengc367 | @Enmk when conflicting files resolved, this pr can be merged? I need this PR for the MaterializeMySQL case testing. thanks. | 2020-08-29T02:36:35 | 2020-08-29T02:37:34 | {} | NONE |
ClickHouse | ClickHouse | 683,304,923 | 14,223 | alexey-milovidov | > 2020-08-28 16:39:44 g++-9: fatal error: Killed signal terminated program cc1plus
OOM in build. | 2020-08-29T15:25:22 | 2020-08-29T15:25:22 | {} | MEMBER |
ClickHouse | ClickHouse | 683,374,039 | 14,251 | liangli1213 | MaterializeMySQL stop sync with below error message:
```
2020.08.30 12:06:14.714991 [ 104873 ] {} <Error> MaterializeMySQLSyncThread: Code: 32, e.displayText() = DB::Exception: Attempt to read after eof, Stack trace (when copying this message, always include the lines below):
0. Poco::Exception::Exception(std::_... | 2020-08-30T04:15:52 | 2020-08-30T04:15:52 | {} | NONE |
ClickHouse | ClickHouse | 683,433,917 | 14,251 | liangli1213 | @BohuTANG my slave server crash when I want to reproduce the issue, and here it comes another problem, I found a piece of transaction are missing in slave, so how can I fix it?
Master:
```
root@bdcmysqldev1:(none)>show master status\G
*************************** 1. row ***************************
Fi... | 2020-08-30T15:27:13 | 2020-08-30T15:27:13 | {} | NONE |
ClickHouse | ClickHouse | 683,463,746 | 14,258 | azat | This also fixed in:
- #14210
- #14237 | 2020-08-30T20:01:12 | 2020-08-30T20:01:12 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 683,472,750 | 11,512 | Enmk | > @Enmk when conflicting files resolved, this pr can be merged? I need this PR for the MaterializeMySQL case testing. thanks.
Sure, I hope it will be merged soon. | 2020-08-30T21:26:16 | 2020-08-30T21:26:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,624,239 | 14,203 | donge | test case for this issue is verified by adding the script, please comment. @filimonov | 2020-08-31T07:49:38 | 2020-08-31T07:49:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,713,897 | 14,237 | qoega | I'll fix stateless in separate PR. | 2020-08-31T11:08:37 | 2020-08-31T11:08:37 | {} | MEMBER |
ClickHouse | ClickHouse | 683,726,024 | 6,943 | alexey-milovidov | @hzruandd It's not a hardcoded restriction, it depends on the actual memory amount and needs. | 2020-08-31T11:37:15 | 2020-08-31T11:37:15 | {} | MEMBER |
ClickHouse | ClickHouse | 683,729,537 | 14,256 | alexey-milovidov | It is possible to disable `-fchar8_t` flag only for CapNProto rather than globally. | 2020-08-31T11:45:53 | 2020-08-31T11:45:53 | {} | MEMBER |
ClickHouse | ClickHouse | 683,350,463 | 14,201 | azat | >When Kafka engine consumes an error, I can skip the error message and continue to consume, prevent the message backlog, and record the error message body and the cause of the error to facilitate tracking of the problem.
As a workaround **for now** the following can be used:
```sql
select * from system.text_log ... | 2020-08-29T22:56:16 | 2020-08-29T22:56:16 | {} | MEMBER |
ClickHouse | ClickHouse | 683,389,536 | 14,251 | bohutang | Please provide:
1. What's the SQL that the MySQL side executed
2. How to produce this issue
| 2020-08-30T07:59:16 | 2020-08-30T07:59:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,527,621 | 14,253 | bohutang | > The `COMMIT` query will not have any effect on MaterializeMySQL engine. but maybe this fix would be better.
Yes, it is a `disturbing` event, we should filte it.
| 2020-08-31T03:15:01 | 2020-08-31T03:15:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,529,737 | 13,879 | ZonghaoWang | Thank you anyway. Well-known perproty filter system like postgres/mysql, ES, Clickhouse, and feature-based search system also serival like [milvus](https://github.com/milvus-io/milvus), [faiss](https://github.com/facebookresearch/faiss) etc. But they can not union always (feature-first-perproty-last or perproty-first-... | 2020-08-31T03:24:19 | 2020-08-31T03:25:40 | {} | NONE |
ClickHouse | ClickHouse | 683,616,862 | 14,260 | azat | >has been broken a long time ago
Didn't noticed that it checks each block, so everything works there (more or less).
Reproducer is simple:
```sql
create table data engine=Null() as system.numbers;
set insert_distributed_timeout=1;
set insert_distributed_sync=1;
insert into function remote('127.1', currentD... | 2020-08-31T07:33:23 | 2020-08-31T07:40:36 | {} | MEMBER |
ClickHouse | ClickHouse | 683,680,249 | 14,251 | liangli1213 | @BohuTANG I still get Packet payload not fully read problem, after:
1. Reinstall ClickHouse
2. Increase server memory
And I'm trying to find out which SQL cause the problem.
```
2020.08.31 16:01:41.201741 [ 13110 ] {} <Debug> MaterializeMySQLSyncThread: Skip MySQL event:
=== TableMapEvent ===
Timestamp: ... | 2020-08-31T09:48:44 | 2020-08-31T09:48:44 | {} | NONE |
ClickHouse | ClickHouse | 683,725,263 | 10,004 | alexey-milovidov | @azat is it fixed by #12982 ? | 2020-08-31T11:35:19 | 2020-08-31T11:35:19 | {} | MEMBER |
ClickHouse | ClickHouse | 683,166,910 | 13,939 | filimonov | > Kafka engine consumes data in the single thread.
Are you sure?
https://github.com/ClickHouse/ClickHouse/pull/11388#issuecomment-638893203
Multithreading is inside UnionBlockInputStream passed to copyData
> They can use single threaded Kafka engine by setting the `kafka_num_consuers = 1
They want to co... | 2020-08-28T22:07:24 | 2020-08-28T22:21:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,260,969 | 14,235 | liangli1213 | @zhang2014 @BohuTANG | 2020-08-29T08:59:27 | 2020-08-29T08:59:27 | {} | NONE |
ClickHouse | ClickHouse | 683,526,622 | 14,235 | zhang2014 | The `COMMIT` query will not have any effect on MaterializeMySQL engine(this is just a warning, and I will change this log to waring log level) | 2020-08-31T03:09:57 | 2020-08-31T03:10:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,548,461 | 14,203 | donge | Just get integration test running. not easy for new guy.
will write a case
```
.Set up TSKV
Set up JSONCompactEachRowWithNamesAndTypes
Set up JSONEachRow
Set up Regexp
Set up TSVWithNames
Set up CSVWithNames
Set up CapnProto
Set up RowBinary
Set up MsgPack
Set up JSONCompactEachRow
Set up Protobuf
Set u... | 2020-08-31T04:56:29 | 2020-08-31T04:56:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,187,983 | 14,208 | filimonov | You are using the version which is not supported anymore. Please update to one of the recent versions, and if the issue reproduces - open a new issue with details. | 2020-08-28T23:25:16 | 2020-08-28T23:25:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,251,831 | 14,220 | alesapin | > Shouldn't final do all needed stuff at once? I.e. remerge all parts (even if it's a single part), calculate all indexes, materialize ttl, materialize missing columns etc?
>
> Ps. I did't check the code, may be i just misleaded by PR description
No, it shouldn't. At least it's not stated in the documentation. An... | 2020-08-29T07:39:53 | 2020-08-29T07:40:30 | {} | MEMBER |
ClickHouse | ClickHouse | 683,399,595 | 14,251 | bohutang | > The replication relationship has been established between MySQL and Clickhouse, after a while, I got below error message, then the replication stopped.
>
> ```
> 2020.08.30 16:38:32.314629 [ 57592 ] {} <Error> MaterializeMySQLSyncThread: Code: 99, e.displayText() = DB::Exception: Packet payload is not fully read.... | 2020-08-30T09:47:46 | 2020-08-30T09:47:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,656,578 | 13,963 | doom369 | @filimonov sorry for the late reply. Looks like some tables were corrupted during the CH update. Table truncate for corrupted tables solved the issue. Funny thing is - only materialized view tables were broken. I'll pay attention during the next updates, as tables corruption doesn't look like the correct flow as well. | 2020-08-31T08:57:55 | 2020-08-31T08:58:21 | {} | NONE |
ClickHouse | ClickHouse | 683,727,531 | 4,761 | alexey-milovidov | In-memory data parts are implemented in #10697 but they are experimental and not enabled by default. | 2020-08-31T11:40:48 | 2020-08-31T11:40:48 | {} | MEMBER |
ClickHouse | ClickHouse | 683,733,442 | 13,939 | fastio | @filimonov Thanks.
> Mybe we can add new opiton, i.e. --enable-multi-threads, which enables users to use multi-threaded Kafka engine, otherwise, original one.
>
> Yes, that is acceptable, but name is bad, as both options are multithreaded. The difference is 'squash data from many threads and flush as a single blo... | 2020-08-31T11:55:29 | 2020-08-31T11:55:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,177,018 | 14,201 | filimonov | > How do we differ between valid and invalid messages?
```
-- MV collecting errors, usual columns exists, but empty
Select _error_message, _topic, _partition, _offset, _raw_message from kafka_table where _error_code <> 0;
-- MV collecting data
Select * from kafka_table where _error_code=0;
```
Or smth similar | 2020-08-28T22:43:12 | 2020-08-28T22:44:54 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,228,460 | 14,231 | l1t1 | many databaes do not support numberleading column names | 2020-08-29T03:18:42 | 2020-08-29T03:18:42 | {} | NONE |
ClickHouse | ClickHouse | 683,326,475 | 14,227 | den-crane | I can reproduce it with 20.8.1.4497 (gcc) | 2020-08-29T18:28:46 | 2020-08-29T18:28:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,525,876 | 14,253 | zhang2014 | The `COMMIT` query will not have any effect on MaterializeMySQL engine(this is just a warning, and I will change this log to waring log level). but maybe this fix would be better. | 2020-08-31T03:06:07 | 2020-08-31T03:09:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,709,619 | 12,552 | KochetovNicolai | I think that it is not very difficult to separate serialization format and data type.
We may postulate that name which is used for serialization is a "serialization format" name, not "data type" name.
There will be mapping "serialization format" -> "data type" (many to one), and currently this mapping is identical.... | 2020-08-31T10:58:59 | 2020-08-31T10:58:59 | {} | MEMBER |
ClickHouse | ClickHouse | 683,729,780 | 14,256 | alexey-milovidov | See `contrib/capnproto-cmake/CMakeLists.txt` | 2020-08-31T11:46:25 | 2020-08-31T11:46:25 | {} | MEMBER |
ClickHouse | ClickHouse | 683,730,090 | 14,256 | alexey-milovidov | With similar `target_compile_options` that we have at the bottom of this file. | 2020-08-31T11:47:08 | 2020-08-31T11:47:08 | {} | MEMBER |
ClickHouse | ClickHouse | 683,734,045 | 14,247 | alexey-milovidov | The relevant code does not exist in this release. | 2020-08-31T11:57:01 | 2020-08-31T11:57:01 | {} | MEMBER |
ClickHouse | ClickHouse | 683,322,157 | 14,227 | den-crane | It seems all CH versions affected. And not only Date type.
```
create table my_table ( i Nullable(UInt64),name Nullable(String)
) engine MergeTree order by tuple();
insert into my_table (name, i) values ('car', 111);
alter table my_table update i = toUInt64OrNull('333') where name = 'car';
select * from ... | 2020-08-29T17:50:52 | 2020-08-29T17:51:09 | {
"hooray": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,323,966 | 14,227 | amosbird | I cannot reproduce with current master. Some bisecting is needed... | 2020-08-29T18:06:42 | 2020-08-29T18:06:42 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 683,372,076 | 14,235 | liangli1213 | @BohuTANG , I find the root cause of empty transaction, I use replicate-ignore-table to filter no needed tables on the slave side, which cause the empty transaction.
So it is better to recognize the empty transaction to previent exception. | 2020-08-30T03:43:08 | 2020-08-30T03:43:08 | {} | NONE |
ClickHouse | ClickHouse | 683,393,551 | 14,251 | liangli1213 | The replication relationship has been established between MySQL and Clickhouse, after a while, I got below error message, then the replication stopped.
```
2020.08.30 16:38:32.314629 [ 57592 ] {} <Error> MaterializeMySQLSyncThread: Code: 99, e.displayText() = DB::Exception: Packet payload is not fully read. Stopped... | 2020-08-30T08:42:58 | 2020-08-30T08:42:58 | {} | NONE |
ClickHouse | ClickHouse | 683,593,826 | 14,218 | markpapadakis | This apparently failed because of PR description quality control issues(1) and because of the following build error which I don't really think it has anything at all to do with the proposed change(2)
Re: (1)... | 2020-08-31T06:44:37 | 2020-08-31T06:44:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,610,420 | 12,551 | alesapin | `> When serializing data types with multiple streams, codecs should be transformed in the following way: specialized codec is dropped; for multiple codecs, every specialized codec is dropped.`
Only for streams which shouldn't be serialized with specialized codecs (array sizes, null mask, etc. )? | 2020-08-31T07:19:39 | 2020-08-31T07:19:39 | {} | MEMBER |
ClickHouse | ClickHouse | 683,432,669 | 14,255 | dmthuc | I can see that problem is from capnproto library that has been described in here https://github.com/capnproto/capnproto/issues/815 | 2020-08-30T15:16:10 | 2020-08-31T08:38:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,670,837 | 14,110 | 4ertus2 | Code in PR should not affect the queries from perf tests that became faster or slower. | 2020-08-31T09:27:36 | 2020-08-31T09:27:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,166,370 | 5,237 | azat | >Report from may of 2019, there is full log in that message - #5237 (comment)
I saw this comment, but I don't see problem here, I was asking about the following query
>Also, as you see in logs, query has been processed in 75.310 sec, but in logs on Sever2 and Server3 I see that each server processed data much fas... | 2020-08-28T22:05:41 | 2020-08-28T22:05:41 | {} | MEMBER |
ClickHouse | ClickHouse | 683,275,114 | 14,235 | liangli1213 | I have dig more and found that there's continuous empty transaction generated, like follows,
```
# at 25921966
#200829 18:10:39 server id 936 end_log_pos 25922034 CRC32 0x7615ebf4 Query thread_id=948 exec_time=0 error_code=0
SET TIMESTAMP=1598695839/*!*/;
BEGIN
/*!*/;
# at 25922034
#200829 18:10:3... | 2020-08-29T11:07:33 | 2020-08-29T11:07:33 | {} | NONE |
ClickHouse | ClickHouse | 683,372,875 | 14,251 | liangli1213 | @BohuTANG @zhang2014 | 2020-08-30T03:55:43 | 2020-08-30T03:55:43 | {} | NONE |
ClickHouse | ClickHouse | 683,691,872 | 13,994 | KochetovNicolai | I expect that such a pipeline will never stop execution (or, probably, will stuck).
In general, it's possible that ill-formed pipeline will stuck or go into infinite loop. Pipeline may also depend on data it reads. Code which creates pipeline is responsible creating correct pipelines. | 2020-08-31T10:14:40 | 2020-08-31T10:14:40 | {} | MEMBER |
ClickHouse | ClickHouse | 683,705,790 | 12,508 | 4ertus2 | 00693_max_block_size_system_tables_columns flapped
https://clickhouse-test-reports.s3.yandex.net/14219/f51e0bfeb0d7b93d8e0da9ceae0e4013d7d57927/fast_test.html#fail1
```
2020-08-28 17:27:07 00693_max_block_size_system_tables_columns: [ FAIL ] - result differs with reference:
2020-08-28 ... | 2020-08-31T10:48:49 | 2020-08-31T10:50:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,167,659 | 14,201 | ramazanpolat | @filimonov putting unparsed rows into KafkaEngine is a great idea but how come a MV can know if it is an unparsed message or not? How do we differ between valid and invalid messages?
For example, `MV1` watches the KafkaEnfine table and tries to parse the message and if its is a valid message, `MV1` puts it into the ... | 2020-08-28T22:10:01 | 2020-08-28T22:10:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,173,367 | 13,939 | filimonov | > Mybe we can add new opiton, i.e. --enable-multi-threads, which enables users to use multi-threaded Kafka engine, otherwise, original one.
Yes, that is acceptable, but name is bad, as both options are multithreaded. The difference is 'squash data from many threads and flush as a single block' vs 'every consumer fl... | 2020-08-28T22:29:00 | 2020-08-28T22:35:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,527,142 | 13,879 | ZonghaoWang | We don't need an high-concurrency, high-performance feature retrieval system, but need a feature + property union search system. HNSW is an graph-based feature-retrieval algrithom, I think it difficult to suit Clickhouse.
No metter what algrithom we use, we must read all feature and property to complish a query. S... | 2020-08-31T03:12:42 | 2020-08-31T06:41:56 | {} | NONE |
ClickHouse | ClickHouse | 683,645,578 | 13,498 | emakarov | @blinkov looks like original documentation was wrongly understood (by me), and functionality with `DateTime64` actually not supported
see here https://t.me/clickhouse_ru/178887
https://t.me/clickhouse_ru/178892 | 2020-08-31T08:34:02 | 2020-08-31T09:03:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,723,562 | 14,268 | pavitrakumar78 | @alexeyerm Thanks for the suggestion. I'm using `groupArrayMovingAvg` for simple moving average ([SMA](https://www.investopedia.com/terms/s/sma.asp#:~:text=A%20simple%20moving%20average%20(SMA)%20is%20an%20arithmetic%20moving%20average,periods%20in%20the%20calculation%20average.)) calculation, but EMA calculation requi... | 2020-08-31T11:31:13 | 2020-08-31T11:31:13 | {} | NONE |
ClickHouse | ClickHouse | 683,165,839 | 5,237 | azat | @den-crane
```sql
-- Seems initiator waits for query results from shards and then starts to solve the query locally
CREATE TABLE events_test1 on cluster test (A Int64) ENGINE = MergeTree() ORDER BY tuple()
CREATE TABLE events_test1_d on cluster test as events_test1 ENGINE = Distributed(test, currentDatabase(),... | 2020-08-28T22:03:55 | 2020-08-28T22:03:55 | {} | MEMBER |
ClickHouse | ClickHouse | 683,310,531 | 14,227 | calebeaires | hey @amosbird , thanks for quick answer... about the query: I've updated the query. | 2020-08-29T16:13:49 | 2020-08-29T16:13:49 | {} | NONE |
ClickHouse | ClickHouse | 683,334,126 | 12,508 | alexey-milovidov | The `01442_merge_detach_attach` test highlights issues if we enable Thread Fuzzer:
https://clickhouse-test-reports.s3.yandex.net/9813/e412017200ef3300d1b36bcafdf46410dd5f4cfc/fast_test/runlog.out.log | 2020-08-29T19:43:45 | 2020-08-29T19:43:45 | {} | MEMBER |
ClickHouse | ClickHouse | 683,714,141 | 14,210 | qoega | fixed in #14237 | 2020-08-31T11:09:15 | 2020-08-31T11:09:15 | {} | MEMBER |
ClickHouse | ClickHouse | 683,722,934 | 14,218 | alexey-milovidov | This is an "exotic" build type. Unfortunately, it was broken in master at this specific moment. See https://github.com/ClickHouse/ClickHouse/pull/13995 for the details. | 2020-08-31T11:29:46 | 2020-08-31T11:30:22 | {} | MEMBER |
ClickHouse | ClickHouse | 683,355,248 | 14,235 | bohutang | To get the sql what's slave executed, you can enable the general log:
```
mysql> set global log_output='table';
Query OK, 0 rows affected (0.00 sec)
mysql> set global general_log=1;
Query OK, 0 rows affected (0.05 sec)
mysql> begin;
Query OK, 0 rows affected (0.00 sec)
mysql> commit;
Query OK, 0 rows aff... | 2020-08-29T23:49:20 | 2020-08-29T23:49:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,529,834 | 14,040 | etah000 | @alexey-milovidov I plan to implement the feature, but just as "den-crane" suggested, I think I can use a rather large number for the default length constraint. Maybe there is no obvious performance difference between VARCH(32) and VARCHAR(65536) if the actually stored data length is smaller than 32 as some RDBMS, s... | 2020-08-31T03:24:48 | 2020-08-31T03:24:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 683,601,798 | 14,045 | alesapin | > > Maybe you can try to install the same version with address sanitizer? You can download it from:
> > https://clickhouse-builds.s3.yandex.net/12470/9c642ff38f8d3180b88caa75228fa2052ddf0036/clickhouse_build_check/report.html
> > ![image](https://user-images.githubusercontent.com/3684697/91466780-48117c80-e898-11ea-9... | 2020-08-31T07:01:51 | 2020-08-31T07:01:51 | {} | MEMBER |
ClickHouse | ClickHouse | 683,721,090 | 14,268 | alexeyerm | There is a chance to get your results via `groupArrayMovingAvg` function.
| 2020-08-31T11:25:37 | 2020-08-31T11:25:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 661,804,580 | 8,678 | alexey-milovidov | We have updated simdjson. The issue is most likely to be fixed.
But I did not validate it because it is an alternative build that we don't have in our CI.
Please let me know if the issue still exisits. | 2020-07-21T11:40:15 | 2020-07-21T11:40:27 | {} | MEMBER |
ClickHouse | ClickHouse | 661,927,736 | 12,625 | alesapin | `2020.07.21 16:36:09.101660 [ 254 ] {} <Error> : Logical error: 'Sort order of blocks violated for column number 6, left: UInt64_255, right: UInt64_0.'.`
Hmmmmmmmm | 2020-07-21T15:23:30 | 2020-07-21T15:23:30 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.