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 | 594,102,696 | 9,464 | alexey-milovidov | `ClickHouse Integration Tests (release)` - that's suspicious.
I will merge if I find the changes totally unrelated... | 2020-03-03T18:38:48 | 2020-03-03T18:38:48 | {} | MEMBER |
ClickHouse | ClickHouse | 594,115,630 | 9,442 | alexey-milovidov | @danlark1 Thank you! I will try to get rid of boost::iterator. | 2020-03-03T19:06:26 | 2020-03-03T19:06:26 | {} | MEMBER |
ClickHouse | ClickHouse | 594,856,935 | 9,484 | alexey-milovidov | > We can eather copy&paste the same change there (sounds quite stupid), or make yandex/clickhouse-stateless-with-coverage-test use yandex/clickhouse-stateless-test as parent, can do that change if is't ok for you. (It will give some more advantages, like less images to download on CI).
Both options are Ok, let's try... | 2020-03-04T21:28:18 | 2020-03-04T21:28:18 | {} | MEMBER |
ClickHouse | ClickHouse | 594,980,919 | 9,506 | templarzq | any way,i'm not worried about brute force attack, just to predict that if there are collision occured in a eb/tb string type column . | 2020-03-05T01:33:19 | 2020-03-05T01:33:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,988,466 | 9,506 | templarzq | LOL | 2020-03-05T02:03:24 | 2020-03-05T02:03:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,113,925 | 9,464 | alexey-milovidov | @KochetovNicolai Please pay attention, because we can face the same issue when you will rewrite Replacing/Collapsing streams. | 2020-03-03T19:02:51 | 2020-03-03T19:02:51 | {} | MEMBER |
ClickHouse | ClickHouse | 594,817,343 | 4,132 | alexey-milovidov | Most likely everything will work file.
And it gives very high chance that we will add ppc64 build in our CI process.
Another examples of minimal testing after build:
```
clickhouse local --query "SELECT 1"
```
Check that the exception handling works:
```
clickhouse local --query "SELECT throwIf(1)"
```
| 2020-03-04T20:29:03 | 2020-03-04T20:29:03 | {} | MEMBER |
ClickHouse | ClickHouse | 594,982,414 | 9,506 | templarzq |
> > So what if we use sip64 instead of sip128?
>
> Then result will be affected by collisions and you will certainly get a few collisions on a set of size about 10 billion. For larger sets, result will be worse.
so if sip64 collision exists, how to ensure that sip128 has no collison in very large datasets? | 2020-03-05T01:39:43 | 2020-03-05T01:40:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 593,940,552 | 9,373 | qoega | OK. Do you plan to add English version as well? | 2020-03-03T13:06:40 | 2020-03-03T13:06:40 | {} | MEMBER |
ClickHouse | ClickHouse | 594,022,628 | 9,485 | traceon | > cp clickhouse-odbc-1.1.4-Linux/lib64/*.so /usr/local/lib/
BTW, it's always a good idea to install the driver to the dedicated prefix, like under `/opt/clickhouse-odbc/1.1.4/` | 2020-03-03T15:54:09 | 2020-03-03T15:54:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,147,947 | 9,484 | alexey-milovidov | > unbundled - 501 Not Implemented.
It is expected:
```
#if USE_POCO_SQLODBC || USE_POCO_DATAODBC
```
I will add an exception. | 2020-03-03T20:17:55 | 2020-03-03T20:17:55 | {} | MEMBER |
ClickHouse | ClickHouse | 594,148,144 | 9,484 | alexey-milovidov | > Actually it can't be tested in CI.
What do you mean? | 2020-03-03T20:18:25 | 2020-03-03T20:18:25 | {} | MEMBER |
ClickHouse | ClickHouse | 594,330,505 | 9,506 | templarzq | Does this lead to fault uniq result? | 2020-03-04T05:10:33 | 2020-03-04T05:10:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,437,223 | 9,503 | 4ertus2 | Files like Server.cpp and Settigns.h were moved to different directories | 2020-03-04T10:18:22 | 2020-03-04T10:18:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,736,031 | 4,132 | alexey-milovidov | https://developer.ibm.com/linuxonpower/advance-toolchain/advtool-compiler/ | 2020-03-04T18:35:31 | 2020-03-04T18:35:31 | {} | MEMBER |
ClickHouse | ClickHouse | 594,751,618 | 4,132 | dmkoltsov | Alexey (@alexey-milovidov ), thanks.
With two "disable"-operations for libhdfs3 and libcpuid modules I have finally build Clickhouse on ppc64le just now. Clickhouse shows help message:
```
# ./dbms/programs/clickhouse local --help
usage: clickhouse-local [initial table definition] [--query <query>]
clickhous... | 2020-03-04T18:56:47 | 2020-03-04T19:05:45 | {} | NONE |
ClickHouse | ClickHouse | 594,982,219 | 9,506 | alexey-milovidov | Yes. | 2020-03-05T01:38:51 | 2020-03-05T01:38:51 | {} | MEMBER |
ClickHouse | ClickHouse | 594,150,812 | 9,484 | alexey-milovidov | > Debug, stateless
```
2020-03-03 19:22:05 [35ab94dc7135] 2020.03.03 19:22:05.111320 [ 215 ] {3c7ac337-8893-4279-9e81-1fdc3fe89ca5} <Error> executeQuery: Code: 86, e.displayText() = DB::Exception: Received error from remote server /columns_info?connection_string=DSN%3D%7BClickHouse%20DSN%20%28ANSI%29%7D&schema=syst... | 2020-03-03T20:25:00 | 2020-03-03T20:25:00 | {} | MEMBER |
ClickHouse | ClickHouse | 594,256,224 | 9,504 | onine007 | OK ,thank you! | 2020-03-04T01:20:42 | 2020-03-04T01:20:42 | {} | NONE |
ClickHouse | ClickHouse | 594,373,261 | 9,176 | onine007 | @den-crane
version 20.2.1
Delete query: ALTER TABLE my_sdap.dm_user_behavior_events DELETE WHERE (event_dt >= 20200229) AND (event_dt <= 20200302) AND (dt = 20200302) AND (event_id IN (36, 37)) AND (client_id IN (1, 2))
[delete.log](https://github.com/ClickHouse/ClickHouse/files/4285848/delete.log)
MemoryTrack... | 2020-03-04T07:45:11 | 2020-03-04T07:57:51 | {} | NONE |
ClickHouse | ClickHouse | 594,651,058 | 9,495 | azat | Progress metrics after query execution (they were incomplete, i.e. number of written/read rows/bytes was less then processed) | 2020-03-04T16:41:17 | 2020-03-04T16:41:17 | {} | MEMBER |
ClickHouse | ClickHouse | 594,051,326 | 9,494 | alexeydubinin | fixed problem by reinstalling clickhouse-server and dropping all data files | 2020-03-03T16:49:29 | 2020-03-03T16:49:29 | {} | NONE |
ClickHouse | ClickHouse | 594,238,607 | 9,501 | den-crane | I don't think such mutations expected to work. And probably it would be super slow because sub-query will be executed for each row.
```
DELETE WHERE IN ( SELECT )
```
I would create a table
```create table TMP engine = MergeTree as
SELECT field1, field2, field9, field8, field7, max(field12)
FROM tm... | 2020-03-04T00:15:51 | 2020-03-04T00:15:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,643,242 | 4,132 | dmkoltsov | Thanks, Alexei, for quick answer!
As one can see, there is required dependency in libhdfs3 module on HWCrc32c code, which is represented in Clickhouse github project as implementation only for x86 platforms.
1) Is there a branch of Clickhouse on github with ppc64le version of this code segment ?
(e.g. accordin... | 2020-03-04T16:26:31 | 2020-03-04T16:28:14 | {} | NONE |
ClickHouse | ClickHouse | 594,706,047 | 4,132 | dmkoltsov | There is another problem - with libcpuid on Ubuntu 18.04 ppc64le. There is no such lib for ppc64le. Did you use some stub instead of original lib or there is a way to disable libcpuid component in Clickhouse build process ? | 2020-03-04T17:55:13 | 2020-03-04T18:00:35 | {} | NONE |
ClickHouse | ClickHouse | 593,942,046 | 9,373 | den-crane | Right after a merge of this one. Because I don't want to spend time on translation of incorrect version.
| 2020-03-03T13:10:19 | 2020-03-03T15:31:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,074,230 | 9,494 | filimonov | Thanks for the nice report. I will check that. | 2020-03-03T17:35:49 | 2020-03-03T17:35:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,115,440 | 9,442 | alexey-milovidov | It's clearly better.
```
SELECT count() FROM (SELECT materialize(1) AS x1, materialize(1) AS x2, materialize(1) AS x3, materialize(1) AS x4, materialize(1) AS x5, materialize(1) AS x6, materialize(1) AS x7, materialize(1) AS x8, materialize(1) AS x9, materialize(1) AS x10 FROM numbers(500000000)) WHERE NOT ignore(xor... | 2020-03-03T19:06:00 | 2020-03-03T19:06:00 | {} | MEMBER |
ClickHouse | ClickHouse | 594,148,915 | 9,484 | alexey-milovidov | > Address Sanitizer: No message received
It's also expected, because we dlopen third-party library that is not instrumented. But it allocates memory. This memory range is not accounted by ASan and ODBC Bridge fails.
I will add another exception. | 2020-03-03T20:20:21 | 2020-03-03T20:20:21 | {} | MEMBER |
ClickHouse | ClickHouse | 594,149,572 | 9,484 | alexey-milovidov | > Memory Sanitizer (stateless)
```
2020.03.03 19:43:29.982882 [ 201 ] {} <Error> Application: DB::Exception: Listen [::]:8443 failed: Code: 344, e.displayText() = DB::Exception: HTTPS protocol is disabled because Poco library was built without NetSSL support. (version 20.3.1.2552)
```
It is completely broken in... | 2020-03-03T20:21:53 | 2020-03-03T20:21:53 | {} | MEMBER |
ClickHouse | ClickHouse | 594,182,354 | 8,765 | alexey-milovidov | You forgot to mention about backward incompatible changes in `system.trace_log`. | 2020-03-03T21:37:07 | 2020-03-03T21:37:07 | {} | MEMBER |
ClickHouse | ClickHouse | 594,399,879 | 9,489 | filimonov | Not clear what with complicated aliases / defaults which depend on other aliases / defaults etc, maybe test for that?
I.e. something like
```
create table test (
a1 UInt64 DEFAULT a+1,
a1 UInt64 DEFAULT a+1,
a2 UInt64 DEFAULT a3+a4,
a3 UInt64 DEFAULT a2+1,
a4 UInt64 ALIAS a3+1
) Engine=Log;
... | 2020-03-04T08:55:44 | 2020-03-04T08:56:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,856,059 | 9,484 | alexey-milovidov | But we don't need it for "stateless with coverage". | 2020-03-04T21:27:04 | 2020-03-04T21:27:04 | {} | MEMBER |
ClickHouse | ClickHouse | 594,895,847 | 9,501 | Tri0L | With dict it's not working too.
Just stuck mutation.
In log I see only one message:
```
2020.03.05 01:14:04.721032 [ 176 ] {0c0e802c-c016-4775-a3b7-d4e6c225ff44} <Information> tmp.tbl: Added mutation: mutation_153.txt
```
system.mutations
```
┌─database─┬─table────────────┬─mutation_id──────┬─command─────... | 2020-03-04T22:28:05 | 2020-03-04T22:28:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,392,195 | 9,489 | qoega | I think we should write in docs explicitly what we support in default and what we don't.
https://github.com/ClickHouse/ClickHouse/blame/master/docs/en/query_language/create.md#L78-L88
| 2020-03-04T08:36:51 | 2020-03-04T08:36:51 | {} | MEMBER |
ClickHouse | ClickHouse | 594,631,700 | 4,132 | alexey-milovidov | We have successfully built and run performance tests of ClickHouse on ppc64le more than one year ago. As we did not implement the support for this type of build in our CI, the build is most likely broken. | 2020-03-04T16:07:39 | 2020-03-04T16:07:39 | {} | MEMBER |
ClickHouse | ClickHouse | 595,175,787 | 4,567 | filimonov | Любопытно что MySQL / Postgres при использовании USING успешно разрешают конфликт про "двузначность" колонок самостоятельно, а если заменить на ON - нет.
http://sqlfiddle.com/#!17/3d069/4
| 2020-03-05T11:16:17 | 2020-03-05T11:16:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 593,941,988 | 8,160 | alexeydubinin | +1. We at hh.ru want to use clickhouse with our kafka based event system and can't undetstand how to extract info from complex jsons. I think there must be solution to map kafka json event to string and extract required info with materialized view and json processing functions (like JSONExtractUInt)
Example json:
```... | 2020-03-03T13:10:12 | 2020-03-03T14:54:47 | {
"laugh": 2
} | NONE |
ClickHouse | ClickHouse | 594,025,091 | 9,480 | den-crane | Works as expected. Yes, collapsingMT replaces rows with the same sign.
```
create table c (A Int64, B String, S Int8) Engine=CollapsingMergeTree(S) order by A;
insert into c values(1, '1', 1);
insert into c values(1, '2', 1);
select * from c final;
┌─A─┬─B─┬─S─┐
│ 1 │ 2 │ 1 │
└───┴───┴───┘
optimize table... | 2020-03-03T15:58:55 | 2020-03-03T16:02:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,075,433 | 9,462 | den-crane | minimal example
```
CREATE TABLE default.test_table
(i Int64,
a_Map Nested(x UInt32,y Double,s String),
b Int32
)ENGINE = SummingMergeTree()
ORDER BY (i);
INSERT into default.test_table values (1,[1],[0.1],['x'],0);
select * from default.test_table;
┌─i─┬─a_Map.x─┬─a_Map.y─┬─a_Map.s─┬─b─┐
│ 1 │ [1] │... | 2020-03-03T17:38:44 | 2020-03-03T17:40:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,123,762 | 9,458 | alexey-milovidov | We decided to replace
```
300 14pt/200% -apple-system, BlinkMacSystemFont, Segoe UI, Helvetica, Arial, sans-serif, Apple Color Emoji,Segoe UI Emoji
```
to
```
300 14pt/200% -apple-system, BlinkMacSystemFont, Segoe UI, Noto Sans, Helvetica, Arial, sans-serif, Apple Color Emoji, Segoe UI Emoji
```
Please also n... | 2020-03-03T19:24:30 | 2020-03-03T19:24:30 | {} | MEMBER |
ClickHouse | ClickHouse | 594,583,621 | 9,506 | alexey-milovidov | > What if the hash code collision happened?
You write an article about the first found collision in SipHash128.
We know in advance that there are collisions in SipHash128 and the probabilities of collisions with random values are:
- that the random value will collide with the fixed value S: 1 / 2 ^ 128;
- that ... | 2020-03-04T15:06:23 | 2020-03-04T15:08:37 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 594,648,570 | 4,132 | dmkoltsov | I see. Thanks! | 2020-03-04T16:36:37 | 2020-03-04T16:36:37 | {} | NONE |
ClickHouse | ClickHouse | 594,979,962 | 9,506 | templarzq | > > What if the hash code collision happened?
>
> You write an article about the first found collision in SipHash128.
>
> We know in advance that there are collisions in SipHash128 and the probabilities of collisions with random values are:
>
> * that the random value will collide with the fixed value S: 1 / 2... | 2020-03-05T01:29:33 | 2020-03-05T01:29:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 593,979,521 | 9,489 | alesapin | Merge after https://github.com/ClickHouse/ClickHouse/pull/9481. | 2020-03-03T14:30:11 | 2020-03-03T14:30:11 | {} | MEMBER |
ClickHouse | ClickHouse | 594,639,642 | 9,495 | alexey-milovidov | Now Ok. | 2020-03-04T16:19:51 | 2020-03-04T16:19:51 | {} | MEMBER |
ClickHouse | ClickHouse | 594,731,124 | 4,132 | alexey-milovidov | If we want to add PPC build to CI, we need to get "toolchain" for cross compilation.
Please look at the build logs for AArch64 for example:
https://clickhouse-builds.s3.yandex.net/0/06c8e0499826c525c0846dccf18ca9b35ce12304/build_log_623232378_1583340152.txt
Also look here: https://clickhouse.tech/docs/en/develop... | 2020-03-04T18:28:54 | 2020-03-04T18:29:48 | {} | MEMBER |
ClickHouse | ClickHouse | 594,854,371 | 9,484 | alexey-milovidov | @filimonov We don't use system LLVM even for "unbundled" build.
https://clickhouse-builds.s3.yandex.net/9484/add201401da7ee7f90929ba383971da7bd93b0f6/build_log_622335845_1583258392.txt
Installation of LLVM in Docker images is pointless. | 2020-03-04T21:24:43 | 2020-03-04T21:24:43 | {} | MEMBER |
ClickHouse | ClickHouse | 594,987,321 | 9,506 | amosbird | I understand the concern of @templarzq (I was there before). I'll post the chat history with @alexey-milovidov here and hope it would help clarify this idea.
```
Amos Bird, [14.03.19 09:34]
I've also noticed that we use the cryptic hash functions for giant keys instead of pointers ... not sure if that's really val... | 2020-03-05T01:59:37 | 2020-03-05T01:59:37 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 595,080,587 | 9,511 | Slach | @alexey-milovidov done, I didn't know about /ping http handler, please look again | 2020-03-05T07:56:49 | 2020-03-05T07:56:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 593,942,912 | 9,373 | den-crane | Or I can add EN translation if this RU is correct and valid and ready to
merge. | 2020-03-03T13:12:14 | 2020-03-03T15:32:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 590,635,684 | 9,243 | onine007 | @filimonov @alesapin
If delete query affect rows then will OOM .
| 2020-02-25T01:33:01 | 2020-03-04T06:59:33 | {} | NONE |
ClickHouse | ClickHouse | 594,377,031 | 9,503 | qoega | Probably it is better to add cub as submodule as it is available here https://github.com/NVlabs/cub ? | 2020-03-04T07:56:46 | 2020-03-04T07:56:46 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 594,770,826 | 9,507 | filimonov | test failure 00910_zookeeper_test_alter_compression_codecs seems unrelated | 2020-03-04T19:21:19 | 2020-03-04T19:21:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,981,702 | 9,506 | templarzq | are there any essential differences between sip64/sip128? Just mapping space is larger in sip128？ | 2020-03-05T01:36:34 | 2020-03-05T01:38:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,983,795 | 9,506 | templarzq | what about to store original string value in the hashset insead of sip128 code? because in normal way string value length is smaller than 16byte(128bit code).this method can ensure no error in uniqExact function on string column. | 2020-03-05T01:45:10 | 2020-03-05T01:47:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,991,285 | 9,142 | alexey-milovidov | Can be squashed before merge. | 2020-03-05T02:13:13 | 2020-03-05T02:13:13 | {} | MEMBER |
ClickHouse | ClickHouse | 593,966,434 | 9,485 | filimonov | So you need to intsall packages using yum instead of apt-get
```
sudo yum install unixODBC
```
I don't remember the location of libs folder on centos, but i think `/usr/local/lib/` should work, ie. no other changes will be needed on centos. | 2020-03-03T14:03:48 | 2020-03-03T14:04:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 593,931,670 | 9,485 | Vibhooti123 | I am working on Redhat Centos8
On Tue, 3 Mar, 2020, 6:09 PM filimonov, <notifications@github.com> wrote:
> You can install clickhouse-odbc on ubuntu like that
>
> wget "https://github.com/ClickHouse/clickhouse-odbc/releases/download/v1.1.4.20200302/clickhouse-odbc-1.1.4-Linux.tar.gz"
> tar xvzf clickhouse-odbc-1.1.4-... | 2020-03-03T12:42:59 | 2020-03-03T12:42:59 | {} | NONE |
ClickHouse | ClickHouse | 594,060,346 | 9,462 | den-crane | The same unexpected?? (not sure maybe because of zeros in ua_is_mobile ) behaviour in 18.14.19. | 2020-03-03T17:06:44 | 2020-03-03T17:06:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,106,648 | 9,491 | alexey-milovidov | It's about padding. The correct result is `TEcgT3B0aW11cw==`
Probably the library forget to fill it with `=`. It should be easily reproducible under MSan. | 2020-03-03T18:47:03 | 2020-03-03T18:47:36 | {} | MEMBER |
ClickHouse | ClickHouse | 594,255,624 | 9,504 | alexey-milovidov | This is normal system behaviour. | 2020-03-04T01:18:41 | 2020-03-04T01:18:41 | {} | MEMBER |
ClickHouse | ClickHouse | 594,257,181 | 9,483 | alexey-milovidov | Performance comparison Ok.
https://clickhouse-test-reports.s3.yandex.net/9483/4f5a2ee6d11e61a1196ebc53d189bb48eb452b64/performance_comparison/report.html | 2020-03-04T01:22:12 | 2020-03-04T01:22:12 | {} | MEMBER |
ClickHouse | ClickHouse | 594,467,398 | 8,594 | excitoon | @alexey-milovidov I did not use MultiVersion because `shared_ptr` looks simpler thing to me. | 2020-03-04T11:27:15 | 2020-03-04T11:27:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,769,771 | 8,639 | nikitamikhaylov | `test_parts_delete_zookeeper/test.py::test_merge_doesnt_work_without_zookeeper` - something new for this PR, but it is known as a flappy test
`test_storage_kafka` errors (asan) are probably a flap, because logs contain nothing. i think it is a problem with kafka's containers startup. | 2020-03-04T19:20:05 | 2020-03-04T20:56:30 | {} | MEMBER |
ClickHouse | ClickHouse | 594,855,661 | 9,484 | alexey-milovidov | Actually it is used for symbolization of messages from sanitizers:
```
ln -s /usr/lib/llvm-8/bin/llvm-symbolizer /usr/bin/llvm-symbolizer
``` | 2020-03-04T21:26:28 | 2020-03-04T21:26:28 | {} | MEMBER |
ClickHouse | ClickHouse | 594,900,948 | 9,501 | den-crane | >MergeTree(field5, (field5, field2, field3), 8192);
Probably mutations are broken for tables with old syntax?
| 2020-03-04T22:33:43 | 2020-03-04T22:33:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 595,004,745 | 9,518 | den-crane | No, mutations make a list of parts which already exist and process this list only. Mutations does not see parts (data) inserted after and does not intersect with inserts. | 2020-03-05T03:06:22 | 2020-03-05T03:06:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 593,992,832 | 8,639 | alesapin | `01083_expressions_in_engine_arguments` -- OK
`01076_parallel_alter_replicated_zookeeper` -- non flapper, fixed in master. | 2020-03-03T14:56:47 | 2020-03-03T14:56:47 | {} | MEMBER |
ClickHouse | ClickHouse | 593,996,939 | 9,484 | filimonov | > Actually it can't be tested in CI.
Docker image will be rebuild only after merge in master. Did you mean that? | 2020-03-03T15:04:54 | 2020-03-03T15:04:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,064,456 | 8,701 | alesapin | Test are OK, will perform post-merge code review. | 2020-03-03T17:15:04 | 2020-03-03T17:15:04 | {} | MEMBER |
ClickHouse | ClickHouse | 594,249,962 | 9,501 | Tri0L | >I would create a table
>created a dictionary TMPDICT
Great way. I will try
>will be executed for each row.
OMG, I don't know it before. Thanks!
But I think the above example is another problem.
9 rows and 3 rows in `IN(...)`, the execution for each row should not take a lot of time.
| 2020-03-04T00:57:50 | 2020-03-04T00:57:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,265,718 | 9,495 | alexey-milovidov | Tests fail: `Unexpected packet from server (expected Exception, EndOfStream or Log, got Progress)` | 2020-03-04T01:29:00 | 2020-03-04T01:29:00 | {} | MEMBER |
ClickHouse | ClickHouse | 594,483,029 | 9,498 | Akazz | > > Functional stateless tests (address)
>
> ```
> 2020-03-04 00:46:32 00910_zookeeper_test_alter_compression_codecs: [ FAIL ] 0.49 sec. - return code 217
> 2020-03-04 00:46:32 Received exception from server (version 20.3.1):
> 2020-03-04 00:46:32 Code: 473. DB::Exception: Received from l... | 2020-03-04T12:06:25 | 2020-03-04T12:06:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,494,659 | 8,639 | CurtizJ | Failures in `test_ttl_move`/`test_multiple_disks` - ok. These tests are ~~fixed~~ skipped in master.
`test_distributed_ddl` fails with ```Data type have to be specified for column `s` to add```. Backward compatabity is broken #9227. | 2020-03-04T12:34:43 | 2020-03-04T12:54:04 | {} | MEMBER |
ClickHouse | ClickHouse | 594,639,854 | 9,495 | alexey-milovidov | But what exactly does it fix? What scenario? | 2020-03-04T16:20:13 | 2020-03-04T16:20:13 | {} | MEMBER |
ClickHouse | ClickHouse | 594,672,377 | 9,484 | filimonov | > Do not understand the reason. @filimonov Please try to reproduce.
It's because "Debug, stateless" uses another docker image `yandex/clickhouse-stateless-with-coverage-test ` which is not inherited for some reason from `yandex/clickhouse-stateless-test`.
We can eather copy&paste the same change there (sounds qu... | 2020-03-04T17:14:02 | 2020-03-04T17:57:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,730,177 | 4,132 | alexey-milovidov | Yes, cpuid is also optional.
It is automatically disabled on ARM: `cmake/find/cpuid.cmake`
It can be also disabled for PPC in the same way.
Or, just change the condition from `NOT ARM` to `AMD64`.
| 2020-03-04T18:27:33 | 2020-03-04T18:27:33 | {} | MEMBER |
ClickHouse | ClickHouse | 594,981,204 | 9,506 | alexey-milovidov | > So what if we use sip64 instead of sip128?
Then result will be affected by collisions and you will certainly get a few collisions on a set of size about 10 billion. For larger sets, result will be worse. | 2020-03-05T01:34:30 | 2020-03-05T01:34:30 | {} | MEMBER |
ClickHouse | ClickHouse | 595,031,603 | 9,518 | Inasayang | Is it possible that the failed mutation caused dirty data? | 2020-03-05T05:08:23 | 2020-03-05T05:08:23 | {} | NONE |
ClickHouse | ClickHouse | 593,969,189 | 9,485 | Vibhooti123 | Ohkk thanks for your help. I will try this.
On Tue, 3 Mar, 2020, 7:33 PM filimonov, <notifications@github.com> wrote:
> So you need to intsall packages using yum instead of apt-get
>
> sudo yum install unixODBC
>
> I don't remember the location of libs folder on centos, but i think
> /usr/local/lib/ should work.
>
> ... | 2020-03-03T14:09:36 | 2020-03-03T14:09:36 | {} | NONE |
ClickHouse | ClickHouse | 594,026,069 | 3,700 | Tri0L | Поймал такое же поведение на версии CH `20.1.2.4`.
Помогло конвертирование даты в timestamp.
Для словарей используем вот такой грязный хак:
```
<expression>cast(my_date_param as timestamp)</expression>
``` | 2020-03-03T16:00:47 | 2020-03-03T16:00:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,255,751 | 9,504 | alexey-milovidov | And this is not an error but `postpone_reason`. | 2020-03-04T01:19:07 | 2020-03-04T01:19:07 | {} | MEMBER |
ClickHouse | ClickHouse | 594,607,462 | 4,132 | dmkoltsov | @danlark1, @alexey-milovidov - Hello, colleagues. With these commits Out-of-box Clickhouse (from github) have no successful build on ppc64le with gcc-9 or advance toolchain 13 on Ubuntu 18.04. Are these commits enough to build on ppc64le ? Were there any successful trials of clickhouse build on ppc64le platform ? Could... | 2020-03-04T15:36:44 | 2020-03-04T15:37:07 | {} | NONE |
ClickHouse | ClickHouse | 594,646,382 | 4,132 | alexey-milovidov | > Is libhdfs3 module mandatory
No. It's only used for interaction with HFDS (rare use case).
And you may notice that we already have one non-x86 (AArch64) build in our CI that is run for every commit.
ppc64le build can be added in a similar fashion.
| 2020-03-04T16:32:33 | 2020-03-04T16:32:33 | {} | MEMBER |
ClickHouse | ClickHouse | 594,733,693 | 4,132 | alexey-milovidov | https://developer.ibm.com/linuxonpower/advance-toolchain/ | 2020-03-04T18:32:17 | 2020-03-04T18:32:17 | {} | MEMBER |
ClickHouse | ClickHouse | 594,988,216 | 9,506 | amosbird | Anyway, if one day you found a collision when using `uniqExact`, try writing a blog and get famous in cryptographic communities. | 2020-03-05T02:02:40 | 2020-03-05T02:02:40 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 594,129,076 | 9,495 | alexey-milovidov | I have labeled it as "improvement" because I don't see how the previous behaviour can be considered a bug. It is visible only in tools that use the native protocol and receive progress and use the progress not only for visualization purposes (otherwise it's indistinguishable). | 2020-03-03T19:35:33 | 2020-03-03T19:35:49 | {} | MEMBER |
ClickHouse | ClickHouse | 594,254,230 | 9,498 | alexey-milovidov | > Functional stateless tests (address)
```
2020-03-04 00:46:32 00910_zookeeper_test_alter_compression_codecs: [ FAIL ] 0.49 sec. - return code 217
2020-03-04 00:46:32 Received exception from server (version 20.3.1):
2020-03-04 00:46:32 Code: 473. DB::Exception: Received from localhost:900... | 2020-03-04T01:13:43 | 2020-03-04T01:13:43 | {} | MEMBER |
ClickHouse | ClickHouse | 594,373,324 | 7,546 | choleraehyq | @alexey-milovidov Still meet `-Wno-ctad-maybe-unsupported` and other errors using both `cmake ..` and
```
cmake .. -DCMAKE_CXX_COMPILER=`which clang++` -DCMAKE_C_COMPILER=`which clang`
```
on 26c74d679f350c482c4e69a0992b323f2f954e94, which is the current master.
cmake error log:
[CMakeError.log](https://githu... | 2020-03-04T07:45:22 | 2020-03-04T07:45:22 | {} | NONE |
ClickHouse | ClickHouse | 594,909,023 | 9,501 | Tri0L | same problem:
```
CREATE TABLE tmp.tbl
(
`field1` Int64,
`field2` UInt64,
`field3` UInt32,
`field4` String,
`field5` Date,
`field6` DateTime,
`field7` Float32,
`field8` Float32,
`field9` Float32,
`field10` Float32,
`field11` UInt32,
`field12` D... | 2020-03-04T22:48:05 | 2020-03-04T22:48:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,979,128 | 9,513 | filimonov | this time 00834_kill_mutation_replicated_zookeeper failed (have no idea why, looks unrelated). | 2020-03-05T01:26:06 | 2020-03-05T01:26:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,983,089 | 9,506 | alexey-milovidov | Both SipHash64 and SipHash128 have collisions, but to get a collision of 128bit variant you have to observe quadratic amount of values comparing to SipHash64 (e.g. billions of billions instead of billions).
| 2020-03-05T01:42:23 | 2020-03-05T01:42:55 | {
"heart": 1
} | MEMBER |
ClickHouse | ClickHouse | 595,047,224 | 7,589 | onine007 | I got same issue in 20.1.* | 2020-03-05T06:12:03 | 2020-03-05T06:15:05 | {} | NONE |
ClickHouse | ClickHouse | 594,020,469 | 9,482 | den-crane | The main problem of optimize final (table or partition does not matter) that it re-writes/re-merges a partition fully even if partition have only 1 part which is excessive in 99.9999% occasions!!!! It re-merges old data which was finally merged already!!!
It needed because sometimes one needs to collapse rows (dupli... | 2020-03-03T15:49:55 | 2020-03-03T15:49:55 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,104,539 | 9,464 | alexey-milovidov | I see that the same test has failed in another PR: #9469
test_ttl_replicated/test.py::test_ttl_many_columns
| 2020-03-03T18:42:43 | 2020-03-03T18:42:43 | {} | MEMBER |
ClickHouse | ClickHouse | 594,143,453 | 9,459 | alexey-milovidov | `Compatibility check — Compatibility check passed` | 2020-03-03T20:07:22 | 2020-03-03T20:07:22 | {} | MEMBER |
ClickHouse | ClickHouse | 594,447,462 | 9,488 | 4ertus2 | Tests with tsan killed for unknown reason. I can't reproduce the kill locally. Perf tests look like flaps. | 2020-03-04T10:41:51 | 2020-03-04T10:41:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 594,590,611 | 9,509 | alexey-milovidov | This bug/feature was implemented in https://github.com/ClickHouse/ClickHouse/pull/4215 | 2020-03-04T15:15:15 | 2020-03-04T15:15:15 | {} | MEMBER |
ClickHouse | ClickHouse | 581,903,811 | 8,968 | liwanxue | version: 19.14.7.15
it's huge log,put it here link: https://pan.baidu.com/s/1EaOSMU6u6riLu2Rob2MnNw verification code: 339s
restart time is "2020.01.20 05:33" and recover time is "2020.01.20 07:13" | 2020-02-04T13:14:15 | 2020-02-04T13:14:15 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.