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 | 586,597,187 | 7,704 | guanlisheng | not quite sure what's the final conclusion after checking #8334, especially for the last comment saying Arrow supports Array/List already. | 2020-02-15T14:47:34 | 2020-02-15T14:47:34 | {} | NONE |
ClickHouse | ClickHouse | 586,967,846 | 9,101 | nvartolomei | Related (i think) https://github.com/ClickHouse/ClickHouse/issues/6349.
Workaround: set `prefer_localhost_replica` to 0. | 2020-02-17T12:17:17 | 2020-02-17T12:18:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 587,052,356 | 5,301 | kreuzerkrieg | To whom it may concern.
PCHed build RFC PR (which does not work)
https://github.com/ClickHouse/ClickHouse/pull/9162
| 2020-02-17T15:44:01 | 2020-02-17T15:44:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 587,132,295 | 9,165 | alexey-milovidov | Now it works as announced. | 2020-02-17T19:36:07 | 2020-02-17T19:36:07 | {} | MEMBER |
ClickHouse | ClickHouse | 587,138,534 | 9,032 | sakthi7 | Thank you !!! | 2020-02-17T19:57:27 | 2020-02-17T19:57:27 | {} | NONE |
ClickHouse | ClickHouse | 587,286,095 | 9,162 | kreuzerkrieg | Maybe @proller or @nvartolomei can see something wrong with this PCH usage? | 2020-02-18T05:31:07 | 2020-02-18T05:31:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 587,325,281 | 9,159 | vpanfilov | @nikitamikhaylov Here is anonymized stacktrace and table DDL. I don't have any materialized or alias columns.
clickhouse-copier stack trace:
```
2020.02.17 16:50:31.645519 [ 1 ] {} <Debug> ClusterCopier: Processing /clickhouse/copytasks/test-task/tables/destination_cluster.<db>.<table>/20201223/shards/1
2020.02.1... | 2020-02-18T07:47:41 | 2020-02-18T07:50:09 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 586,520,907 | 9,091 | alexey-milovidov | > (except the part you mentioned perf top, I would need to have an insight about what kind of information you need)
Is the function really slow as it looks for me?
Is it true that the compiler cannot vectorize the loop with UInt8? | 2020-02-14T23:40:50 | 2020-02-14T23:40:50 | {} | MEMBER |
ClickHouse | ClickHouse | 586,522,279 | 9,129 | alexey-milovidov | Build check:
```
collect2: fatal error: ld terminated with signal 9 [Killed]
```
This is due to an issue in CI. | 2020-02-14T23:47:10 | 2020-02-14T23:47:10 | {} | MEMBER |
ClickHouse | ClickHouse | 586,535,067 | 9,070 | den-crane | @urykhy at least you can emulate MV insert fail
https://gist.github.com/den-crane/bc35a15a8d71899fe15e5ed36668d387
As far as I understand with your setting the second insert (retry) should insert rows into test12345mv and `SELECT count() FROM test12345mv` should return rows.
You can implement two test cases with... | 2020-02-15T00:56:35 | 2020-02-15T01:07:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,811,317 | 9,049 | alexey-milovidov | @alesapin Suddenly, all the tests have passed. | 2020-02-17T04:22:46 | 2020-02-17T04:22:46 | {} | MEMBER |
ClickHouse | ClickHouse | 586,965,006 | 9,153 | CurtizJ | It should be fixed in #8914. The fix is available in version 20.1.4.14. | 2020-02-17T12:08:43 | 2020-02-17T12:08:43 | {} | MEMBER |
ClickHouse | ClickHouse | 587,234,936 | 9,176 | onine007 | > Does this number 5491 change? Basically everything works as expected.
> You should understand that each mutation could re-write whole dm_user_behavior_events table completely. It could take a lot of time, need to read, decompress, filter, compress, write rows in many or all parts (all rows in a table).
>
> Mutati... | 2020-02-18T01:46:19 | 2020-02-18T01:46:19 | {} | NONE |
ClickHouse | ClickHouse | 587,333,320 | 9,176 | onine007 | query:select database,table,is_done,count() from system.mutations group by database,table,is_done
result :
my_sdap crowd_bitmap_result_replicated 1 100
my_sdap crowd_bitmap_result_replicated 0 445
my_sdap dm_user_behavior_events 0 5491
my_sdap dm_user_behavior_events 1 100
my_sdap book... | 2020-02-18T08:12:02 | 2020-02-18T08:12:02 | {} | NONE |
ClickHouse | ClickHouse | 586,520,689 | 9,052 | alexey-milovidov | Waiting for the test. | 2020-02-14T23:39:46 | 2020-02-14T23:39:46 | {} | MEMBER |
ClickHouse | ClickHouse | 586,527,721 | 9,138 | alexey-milovidov | A test was not added because of low significance of this feature. | 2020-02-15T00:12:19 | 2020-02-16T03:14:19 | {} | MEMBER |
ClickHouse | ClickHouse | 586,676,437 | 9,036 | alexey-milovidov | Also good for every color scheme in Konsole. | 2020-02-16T07:12:21 | 2020-02-16T07:12:21 | {} | MEMBER |
ClickHouse | ClickHouse | 586,697,358 | 640 | George3d6 | @ramazanpolat No easy way to go about it that I found, easiest is, give the rows a uid if they don't have them, insert with `--input_format_allow_errors_num=999999999` then look which of the uids are missing from the table you inserted in.
Alternatively there might be a better command for printing the insert error, ... | 2020-02-16T11:51:36 | 2020-02-16T11:52:00 | {
"+1": 4
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,846,491 | 9,004 | alexey-milovidov | https://clickhouse-test-reports.s3.yandex.net/codebrowser/html_report///ClickHouse/contrib/libunwind/src/Registers.hpp.html#381
https://clickhouse-test-reports.s3.yandex.net/codebrowser/html_report///ClickHouse/contrib/libunwind/src/config.h.html#113
@pva Looks like you are building the code without NDEBUG (with as... | 2020-02-17T07:04:47 | 2020-02-17T07:05:27 | {} | MEMBER |
ClickHouse | ClickHouse | 586,880,549 | 9,151 | KochetovNicolai | > PartialSortingTransform should leave const columns as is
It's true. The problem should have been in pipeline itself. | 2020-02-17T08:49:22 | 2020-02-17T08:49:33 | {} | MEMBER |
ClickHouse | ClickHouse | 587,132,476 | 9,164 | alexey-milovidov | > Style check passed | 2020-02-17T19:36:43 | 2020-02-17T19:36:43 | {} | MEMBER |
ClickHouse | ClickHouse | 586,534,439 | 9,115 | den-crane | ext.dictionaries
```
CREATE TABLE test1
(
`col1` Int8,
`col2` String,
`col3` Date,
`col4` UInt16
)
ENGINE = MergeTree order by tuple();
INSERT INTO test1 VALUES(1,'001','2020-01-01', 1)(1,'002','2020-01-01', 1)
(2,'001','2020-01-01', 2)(2,'002','2020-01-02', 3)(2,'003','2020-01-04', 5)... | 2020-02-15T00:52:26 | 2020-02-15T00:52:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,564,734 | 32 | stigsb | Could this be doable as a macro instead? Like a preprocessing step before the actual query parsing? | 2020-02-15T07:43:39 | 2020-02-15T07:43:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,685,690 | 9,054 | alexey-milovidov | `-DENABLE_EMBEDDED_COMPILER:BOOL=OFF`
BTW, this option is outdated long time ago. | 2020-02-16T09:31:20 | 2020-02-16T09:31:20 | {} | MEMBER |
ClickHouse | ClickHouse | 586,811,422 | 9,049 | alexey-milovidov | The issue may be related to the wrong code of #8765 (it contains data race). | 2020-02-17T04:23:27 | 2020-02-17T04:23:27 | {} | MEMBER |
ClickHouse | ClickHouse | 587,113,272 | 9,162 | alexey-milovidov | I have tested pch long time ago (with compiled Aggregator) and it made only slight improvement in build time. As far as I understand, the main cost of build time is template instantiation rather than headers parsing. | 2020-02-17T18:29:01 | 2020-02-17T18:29:01 | {} | MEMBER |
ClickHouse | ClickHouse | 587,125,269 | 9,161 | alexey-milovidov | #9161 | 2020-02-17T19:10:48 | 2020-02-17T19:10:48 | {} | MEMBER |
ClickHouse | ClickHouse | 587,346,797 | 3,277 | filimonov | Fixed in #8903 and #8933
Will be released in 20.2 | 2020-02-18T08:49:59 | 2020-02-18T08:49:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,720,469 | 9,145 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/1562
check this https://github.com/ClickHouse/ClickHouse/issues/4411#issuecomment-464329721 | 2020-02-16T15:39:31 | 2020-02-16T15:39:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,855,770 | 8,994 | alexey-milovidov | > Also there is domain-aware random generators like
@filimonov The best available domain-specific generator is in `clickhouse-obfuscator` tool. | 2020-02-17T07:38:37 | 2020-02-17T07:38:37 | {} | MEMBER |
ClickHouse | ClickHouse | 587,230,810 | 9,176 | den-crane | Does this number 5491 change? Basically everything works as expected.
You should understand that each mutation could re-write whole dm_user_behavior_events table completely. It could take a lot of time, need to read, decompress, filter, compress, write rows in many or all parts (all rows in a table).
Mutation sho... | 2020-02-18T01:26:46 | 2020-02-18T01:27:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,541,330 | 8,467 | yuzhichang | @alexey-milovidov The failing case looks caused by unstable test environment. | 2020-02-15T01:49:56 | 2020-02-15T01:49:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,675,158 | 9,143 | alexey-milovidov | https://clickhouse-test-reports.s3.yandex.net/9076/aaaf4ef6a80cacf649a2d80a74c51865622a9f91/functional_stateless_tests_(ubsan)/stderr.log | 2020-02-16T06:48:25 | 2020-02-16T06:48:25 | {} | MEMBER |
ClickHouse | ClickHouse | 586,827,971 | 9,153 | onine007 | Error log :
2020.02.17 13:45:02.396495 [ 90 ] {ac085a34-77a5-43dc-8d6b-f878fc6d1c06} <Error> executeQuery: Code: 12, e.displayText() = DB::Exception: Parameters start = 17658, length = 8 are out of bound in ColumnVector<T>::insertRangeFrom method (data.size() = 13269). (version 20.1.3.7 (official build)) (from 10.0.1... | 2020-02-17T05:47:37 | 2020-02-17T05:47:37 | {} | NONE |
ClickHouse | ClickHouse | 587,315,286 | 8,568 | dingxiangfei2009 | @huang--yijian It was supposed to be #8569. Don't worry, that issue is totally irrelevant to this. :sweat_smile: | 2020-02-18T07:14:42 | 2020-02-18T07:15:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,522,451 | 9,129 | alexey-milovidov | FYI. If you don't end the last line in a file with LF, github will show ugly mark in diff. | 2020-02-14T23:47:57 | 2020-02-14T23:47:57 | {} | MEMBER |
ClickHouse | ClickHouse | 586,694,591 | 8,985 | alexey-milovidov | A bug:
```
│ 4204544 │ StackTrace::tryCapture(
StackTrace::StackTrace(
DB::TraceCollector::collect(unsigned long
MemoryTracker::alloc(long
MemoryTracker::alloc(long
CurrentMemoryTracker::alloc(long
CurrentMemoryTracker::realloc(long, long
``` | 2020-02-16T11:19:28 | 2020-02-16T11:19:28 | {} | MEMBER |
ClickHouse | ClickHouse | 586,824,998 | 8,568 | huang--yijian | > @huang--yijian Sorry, I type the wrong digit for my issue number.
What's the correct issue number you are referring to ? | 2020-02-17T05:33:20 | 2020-02-17T05:33:20 | {} | NONE |
ClickHouse | ClickHouse | 586,901,486 | 9,156 | azat | >I have added a comment with explanation. Please check if it's correct.
I think it is, thanks! | 2020-02-17T09:37:56 | 2020-02-17T09:37:56 | {} | MEMBER |
ClickHouse | ClickHouse | 586,984,119 | 9,154 | azat | >don't throw from the job function of ThreadFromGlobalPool.
Ok.
Does `onFinish*` should be wrapped with `try`/`catch`/`onException` anyway? (looks like it still worth adding) | 2020-02-17T13:04:36 | 2020-02-17T13:04:36 | {} | MEMBER |
ClickHouse | ClickHouse | 587,165,342 | 9,156 | azat | >ClickHouse build check — 16/16 builds are OK
Builds are ok | 2020-02-17T21:31:14 | 2020-02-17T21:31:14 | {} | MEMBER |
ClickHouse | ClickHouse | 587,174,811 | 9,174 | alexey-milovidov | We can also close files in a loop after `vfork` and before `exec` but it can be tricky. | 2020-02-17T22:12:32 | 2020-02-17T22:12:32 | {} | MEMBER |
ClickHouse | ClickHouse | 587,269,398 | 9,174 | azat | >We can also close files in a loop after vfork and before exec but it can be tricky.
Indeed, and not only tricky, but also very slow | 2020-02-18T04:11:59 | 2020-02-18T04:11:59 | {} | MEMBER |
ClickHouse | ClickHouse | 586,905,329 | 9,157 | alexey-milovidov | We can do it with a "fallback table" with a single String column. This table can be specified in settings. The default like `system.parse_errors` may be created. | 2020-02-17T09:47:09 | 2020-02-17T09:47:09 | {} | MEMBER |
ClickHouse | ClickHouse | 587,175,280 | 9,159 | nikitamikhaylov | Could you please provide us more informative stacktrace?
I think that you can't copy tables with different schema, because "custom" insert-select is implemented in this program. But this issue will be evaluated as a feature request :wink: | 2020-02-17T22:14:36 | 2020-02-17T22:14:36 | {} | MEMBER |
ClickHouse | ClickHouse | 587,233,641 | 9,176 | onine007 | @den-crane 5491 is not changed , 63 delete queries is requested per day . | 2020-02-18T01:39:45 | 2020-02-18T01:39:45 | {} | NONE |
ClickHouse | ClickHouse | 586,453,761 | 7,704 | filimonov | Check https://github.com/ClickHouse/ClickHouse/pull/8334 + comments | 2020-02-14T20:09:17 | 2020-02-14T20:09:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,520,422 | 9,100 | alexey-milovidov | > The trace indeed shows that it is called from shutdown thinking. I can try and reproduce it with this patch and without next week.
Yes, and shutdown is probably caused by another reason that you can find. | 2020-02-14T23:38:31 | 2020-02-14T23:38:31 | {} | MEMBER |
ClickHouse | ClickHouse | 586,520,615 | 9,052 | alexey-milovidov | Performance test - checked manually. | 2020-02-14T23:39:27 | 2020-02-14T23:39:27 | {} | MEMBER |
ClickHouse | ClickHouse | 586,673,720 | 9,094 | alexey-milovidov | Should be easy to fix. | 2020-02-16T06:19:13 | 2020-02-16T06:19:13 | {} | MEMBER |
ClickHouse | ClickHouse | 586,681,550 | 6,863 | filimonov | Altinity marks some version as stable after collecting feedback, analizing known issues, doing some extra tests. In general it's always behind the edge version (latest avaliable in from yandex repo).
Avoid using versions older than latest Altinity stable.
Use Yandex latest version when possible - it have all the ... | 2020-02-16T08:35:36 | 2020-02-16T08:36:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,954,111 | 9,154 | azat | >No logic in ThreadPool was changed. It should work with unhandled exceptions regardless to the change
In case of exception it will shutdown it, and in some cases the global pool can be terminated, which will lead to "cannot schedule task" permanently for every new connection/query
https://github.com/ClickHouse/C... | 2020-02-17T11:40:47 | 2020-02-17T12:34:39 | {} | MEMBER |
ClickHouse | ClickHouse | 587,300,033 | 9,176 | onine007 | @den-crane
It is partition by event_dt , not dt . and also be fine (dt = 20191125) AND (event_id IN (29) ?
yes, it is replicated table .
query: select count() from system.replication_queue
return 0
| 2020-02-18T06:27:48 | 2020-02-18T06:27:48 | {} | NONE |
ClickHouse | ClickHouse | 586,675,911 | 9,054 | alexey-milovidov | A test:
```
sudo docker run --rm --network host -it ubuntu:eoan /bin/bash
apt update
apt install git cmake ninja-build g++ python
git clone --recursive https://github.com/ClickHouse/ClickHouse.git
mkdir build && cd build
cmake ../ClickHouse
ninja
``` | 2020-02-16T07:02:54 | 2020-02-16T07:02:54 | {} | MEMBER |
ClickHouse | ClickHouse | 586,682,445 | 9,148 | alexey-milovidov | ```
sudo docker run --rm --network host -it ubuntu:eoan /bin/bash
apt update
apt install git cmake ninja-build g++ python
git clone --recursive https://github.com/ClickHouse/ClickHouse.git
mkdir build && cd build
cmake ../ClickHouse
ninja
```
```
/usr/bin/ld.gold: error: cannot find -lzlibstatic
``` | 2020-02-16T08:48:50 | 2020-02-16T08:48:50 | {} | MEMBER |
ClickHouse | ClickHouse | 586,683,156 | 9,144 | alexey-milovidov | I've checked the log manually in CI runs.
> Functional stateless tests (unbundled) — fail:1, passed:1572, skipped:26
- see https://github.com/ClickHouse/ClickHouse/pull/9149 | 2020-02-16T08:57:46 | 2020-02-16T08:57:46 | {} | MEMBER |
ClickHouse | ClickHouse | 587,035,292 | 9,160 | alexey-milovidov | If use `hyperscan` instead of `re2`, everything is Ok:
`select match('\0 key="v" ', 'key="(.*?)"')`
0
`select multiMatchAny('\0 key="v" ', ['key="(.*?)"'])`
1 | 2020-02-17T15:03:26 | 2020-02-17T15:03:26 | {} | MEMBER |
ClickHouse | ClickHouse | 587,301,068 | 9,176 | onine007 | May be re-write whole table
│ my_sdap │ dm_user_behavior_events │ 0000016363 │ DELETE WHERE (event_dt = 20191127) AND (dt = 20191129) AND (event_id IN (2228)) │ 2020-02-18 14:24:40 │ ['20190412','20190505','20190506','20190511','20190512','20190513','20190515','20190516','20190517','20190519','20190520','20190521... | 2020-02-18T06:31:50 | 2020-02-18T06:31:50 | {} | NONE |
ClickHouse | ClickHouse | 586,546,208 | 9,134 | den-crane | ```
STR one host 20.1.4.14:
drop table if exists da_memory_efficient_shard ;
create table da_memory_efficient_shard(A Int64, B Int64) Engine=MergeTree order by A;
insert into da_memory_efficient_shard select toString(rand()),number from numbers(100000);
set distributed_aggregation_memory_efficient = 1,
group... | 2020-02-15T02:51:03 | 2020-02-15T02:51:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,688,820 | 9,054 | alexey-milovidov | Another test:
```
sudo docker run --rm --network host -it opensuse/tumbleweed /bin/bash
zypper install git cmake ninja gcc-c++ python
git clone --recursive https://github.com/ClickHouse/ClickHouse.git
mkdir build && cd build
cmake ../ClickHouse
ninja
```
Yes, it was failed at the cmake configuration stage ... | 2020-02-16T10:12:30 | 2020-02-16T10:13:29 | {} | MEMBER |
ClickHouse | ClickHouse | 586,692,283 | 9,149 | alexey-milovidov | url_engine - flacky test (the server for URL engine did not start) | 2020-02-16T10:52:58 | 2020-02-16T10:52:58 | {} | MEMBER |
ClickHouse | ClickHouse | 586,796,453 | 4,061 | yyfrankyy | Same here, but different error message:
```
Cannot load data for command line suggestions: Code: 1000, e.displayText() = DB::Exception: Received from clickhouse.weread.oa.com:29000. DB::Exception: mysqlxx::ConnectionFailed: Can't connect to MySQL server on '10.208.145.225' (110) ((nullptr):0). (version 19.16.3.6 (o... | 2020-02-17T02:53:21 | 2020-02-17T02:53:21 | {} | NONE |
ClickHouse | ClickHouse | 586,804,042 | 9,153 | onine007 | > query :
> SELECT
> event_dt,
> mode,
> countDistinct(user_id) AS measure
> FROM events
> WHERE (events.event_dt = 20200216) AND (events.client_id = 1) AND (events.event_id = 4) AND ((toInt32OrNull(events.event_attr16) IN (2)) AND (events.client_ver >= '0757'))
> GROUP BY
> event_dt,
> mode
>
> Received ex... | 2020-02-17T03:38:33 | 2020-02-17T03:38:33 | {} | NONE |
ClickHouse | ClickHouse | 586,906,542 | 9,151 | qoega | @KochetovNicolai have you tried to build with with `CXXFLAGS='-ggdb -fno-omit-frame-pointer -fsanitize-memory-track-origins=2'`? | 2020-02-17T09:50:02 | 2020-02-17T09:50:02 | {} | MEMBER |
ClickHouse | ClickHouse | 586,947,634 | 9,143 | alexey-milovidov | 01017_uniqCombined_memory_usage - flacky test? | 2020-02-17T11:23:48 | 2020-02-17T11:23:48 | {} | MEMBER |
ClickHouse | ClickHouse | 586,516,101 | 9,136 | alexey-milovidov | Dependency inversion must be in all factories. | 2020-02-14T23:19:29 | 2020-02-14T23:19:29 | {} | MEMBER |
ClickHouse | ClickHouse | 586,640,702 | 640 | ramazanpolat | Hi @George3d6, how do you handle rows that has parse error?
AFAIK when you use `--input_format_allow_errors_num`, Clickhouse does not show the rows with parsing errors. So how do you find the problematic rows? | 2020-02-15T20:59:14 | 2020-02-15T20:59:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,698,309 | 9,014 | alexey-milovidov | Tests have failed. | 2020-02-16T12:01:48 | 2020-02-16T12:01:48 | {} | MEMBER |
ClickHouse | ClickHouse | 586,711,929 | 9,145 | lzy305 |
(package_num = 0 ? toDecimal128(0.0000, 4) : d_rt_qty) / (package_num = 0 ? toDecimal128(1.0000, 4) : package_num)
| 2020-02-16T14:11:48 | 2020-02-16T14:11:48 | {} | NONE |
ClickHouse | ClickHouse | 586,807,068 | 9,151 | alexey-milovidov | @KochetovNicolai I thought PartialSortingTransform should leave const columns as is. Otherwise we will have performance degradation. Also look at MSan test. | 2020-02-17T03:56:15 | 2020-02-17T03:56:15 | {} | MEMBER |
ClickHouse | ClickHouse | 586,881,576 | 9,151 | KochetovNicolai | Msan points to `DB::StringEqualsImpl<false>::string_vector_string_vector`. I think I've seen it before. Can't get the reason why it happened. | 2020-02-17T08:52:04 | 2020-02-17T08:52:04 | {} | MEMBER |
ClickHouse | ClickHouse | 586,978,973 | 9,154 | alexey-milovidov | The proper fix will be:
- don't throw from the job function of ThreadFromGlobalPool.
https://clickhouse-test-reports.s3.yandex.net/codebrowser/html_report/ClickHouse/dbms/src/Common/ThreadPool.h.html#164 | 2020-02-17T12:50:57 | 2020-02-17T12:50:57 | {} | MEMBER |
ClickHouse | ClickHouse | 587,055,292 | 9,162 | alexey-milovidov | Probably it won't work in our CI because we use old CMake.
Is it possible to enable it conditionally depending on the version? | 2020-02-17T15:50:48 | 2020-02-17T15:50:48 | {} | MEMBER |
ClickHouse | ClickHouse | 587,122,448 | 9,161 | alexey-milovidov | In `daemon` and `mysqlxx` the directories `include` and `src` remain.
Let's merge together? | 2020-02-17T19:01:12 | 2020-02-17T19:01:12 | {} | MEMBER |
ClickHouse | ClickHouse | 587,237,079 | 9,176 | onine007 | @den-crane
│ my_sdap │ dm_user_behavior_events │ 0000010881 │ DELETE WHERE (event_dt >= 20191123) AND (event_dt <= 20191125) AND (dt = 20191125) AND (event_id IN (29)) │ 2019-12-03 16:46:09 │ ['20190412','20190505','20190506','20190511','20190512','20190513','20190515','20190516','20190517','20190519','20190520',... | 2020-02-18T01:57:13 | 2020-02-18T01:57:13 | {} | NONE |
ClickHouse | ClickHouse | 587,259,904 | 9,176 | den-crane | @onine007 it should be fine `AND (dt = 20191125) AND (event_id IN (29)` it will read and process only `'20191125'` .
Is it replicated table? Check system.replication_queue how many mutations are processed right now. | 2020-02-18T03:21:02 | 2020-02-18T03:21:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 587,270,178 | 9,154 | azat | >I'm suggesting to not shutdown the loop on exceptions, for now this can done only if this is the global pool, but I would prefer not to add this condition
Decided to make it for global pool only (at least for now) | 2020-02-18T04:16:03 | 2020-02-18T04:16:03 | {} | MEMBER |
ClickHouse | ClickHouse | 587,285,800 | 9,162 | kreuzerkrieg | > I have tested pch long time ago (with compiled Aggregator) and it made only slight improvement in build time. As far as I understand, the main cost of build time is template instantiation rather than headers parsing.
Hm... interesting... Then what improvement was exhibited by `cotire`? For some reason the detected... | 2020-02-18T05:29:39 | 2020-02-18T05:29:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,525,313 | 9,079 | alexey-milovidov | Integration test has failed for reason? | 2020-02-15T00:00:23 | 2020-02-15T00:00:23 | {} | MEMBER |
ClickHouse | ClickHouse | 586,545,439 | 9,134 | den-crane | STR:
at 2 hosts (CH 20.1.4) :
```
drop table if exists da_memory_efficient_shard ;
drop table if exists da_memory_efficient;
create table da_memory_efficient_shard(A Int64, B Int64) Engine=MergeTree order by A;
insert into da_memory_efficient_shard select toString(rand()),number from numbers(100000);
cre... | 2020-02-15T02:39:27 | 2020-02-15T02:41:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,677,084 | 6,402 | filimonov | Fixed by #8911 | 2020-02-16T07:23:57 | 2020-02-16T07:23:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,685,619 | 9,054 | alexey-milovidov | The test has completed successfully. It means that the issue is not related to glibc 2.30. | 2020-02-16T09:30:34 | 2020-02-16T09:30:34 | {} | MEMBER |
ClickHouse | ClickHouse | 586,905,497 | 9,157 | alexey-milovidov | The number of rejected rows should be also tracked in ProfileEvents. | 2020-02-17T09:47:35 | 2020-02-17T09:47:35 | {} | MEMBER |
ClickHouse | ClickHouse | 586,951,284 | 9,154 | alexey-milovidov | No logic in ThreadPool was changed. It should work with unhandled exceptions regardless to the change. | 2020-02-17T11:33:22 | 2020-02-17T11:33:22 | {} | MEMBER |
ClickHouse | ClickHouse | 587,122,172 | 9,157 | ramazanpolat | > We can do it with a "fallback table" with a single String column. This table can be specified in settings. The default like `system.parse_errors` may be created.
"Fallback table" is a great idea but please consider also providing option to print to sysout or syserr. This will be useful for bash scripts. | 2020-02-17T19:00:16 | 2020-02-17T19:00:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 587,162,298 | 9,168 | alexey-milovidov | @blinkov ClickHouse is not a Python project. | 2020-02-17T21:18:40 | 2020-02-17T21:18:40 | {} | MEMBER |
ClickHouse | ClickHouse | 587,229,386 | 9,176 | onine007 | @den-crane
result is empty.
┌─latest_fail_reason─┐
│ │
│ │
│ │
└────────────────────┘
3 rows in set. Elapsed: 0.507 sec. Processed 5.79 thousand rows, 35.63 MB (11.43 thousand rows/s., 70.33 MB/s.) | 2020-02-18T01:19:17 | 2020-02-18T01:19:17 | {} | NONE |
ClickHouse | ClickHouse | 586,444,781 | 7,462 | den-crane | ```
18.14.18
┌─attr─┬─_id─┬─z─────┐
│ 1 │ 0 │ ['1'] │
│ 1 │ 1 │ ['1'] │
└──────┴─────┴───────┘
20.2.1.2362
DB::Exception: Cannot get value from Function: While executing MergeSortingTransform.
``` | 2020-02-14T19:43:17 | 2020-02-14T19:43:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,663,754 | 9,052 | alexey-milovidov | Thank you! | 2020-02-16T03:11:29 | 2020-02-16T03:11:29 | {} | MEMBER |
ClickHouse | ClickHouse | 587,034,056 | 9,160 | alexey-milovidov | It looks like the deficiency of `re2` library. **Upd: untrue.**
It also does not support zero bytes in regexp. | 2020-02-17T15:00:41 | 2020-02-17T15:27:33 | {} | MEMBER |
ClickHouse | ClickHouse | 587,228,847 | 9,176 | den-crane | select **latest_fail_reason** from system.mutations where is_done=0 limit 3; | 2020-02-18T01:16:23 | 2020-02-18T01:16:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 586,978,307 | 9,154 | alexey-milovidov | > In case of exception it will shutdown it, and in some cases the global pool can be terminated, which will lead to "cannot schedule task" permanently for every new connection/query
We don't use GlobalThreadPool directly, only via ThreadFromGlobalPool. | 2020-02-17T12:48:50 | 2020-02-17T12:48:50 | {} | MEMBER |
ClickHouse | ClickHouse | 586,570,141 | 9,115 | chu1070y | Thank for help. Thanks. | 2020-02-15T09:03:01 | 2020-02-15T09:03:01 | {} | NONE |
ClickHouse | ClickHouse | 586,844,453 | 9,004 | alexey-milovidov | > Also it's important to know why you build against local libc - we don't support this in our CI
@abyss7 You are missing something. | 2020-02-17T06:57:20 | 2020-02-17T06:57:20 | {} | MEMBER |
ClickHouse | ClickHouse | 586,905,291 | 9,154 | azat | > 00974_query_profiler | FAIL
Fails on upstream/master too | 2020-02-17T09:47:03 | 2020-02-17T09:47:03 | {} | MEMBER |
ClickHouse | ClickHouse | 586,954,745 | 8,594 | nvartolomei | This is exciting. Can you share more about the intention here and what use cases you want to support? Ie. will it be possible to just remove a disk and ClickHouse will run part checks/repair for replicated tables and missing parts? | 2020-02-17T11:42:19 | 2020-02-17T11:42:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 587,064,232 | 9,162 | kreuzerkrieg | > Probably it won't work in our CI because we use old CMake.
> Is it possible to enable it conditionally depending on the version?
you didnt read what I've wrote :)
It doesnt make any difference in build time, it is RFC PR, maybe you guys can see what I did (or didnt) wrong | 2020-02-17T16:09:27 | 2020-02-17T16:09:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 587,132,390 | 9,154 | azat | >don't throw from the job function of ThreadFromGlobalPool.
So the exception need to be re-thrown somehow, this can be done from the `join()`, but in this case the following should be done:
- modify each `join()` that called in destructor -- not a problem
- `ThreadPoolImpl` will not known that exception occurred u... | 2020-02-17T19:36:24 | 2020-02-17T19:36:24 | {} | MEMBER |
ClickHouse | ClickHouse | 587,235,306 | 9,176 | den-crane | >Re-write whole table ? not some partitions ?
it could. it depends on mutation WHERE and table partition by / order by | 2020-02-18T01:48:20 | 2020-02-18T01:48:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 575,661,670 | 8,696 | akuzm | Reproduces as simple as changing the log file name in the config. The server continues writing to the old file. The snippet I'm using to check:
```
sed -i "s/log>clickhouse-server[0-9]*/log>clickhouse-server10/g" db/config.d/zz-logs.xml
sleep 1 # this sleep is important -- idk why
rm *.log ||:
while ls -lath /proc... | 2020-01-17T15:03:05 | 2020-01-17T15:03:05 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.