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 | 797,360,292 | 10,935 | filimonov | Many health checks just open/close port to check the connectivity. That's definitely should not spam in clickhouse-server.err.log | 2021-03-12T09:27:19 | 2021-03-12T09:27:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,510,619 | 19,257 | nikitamikhaylov | Last commit contains only extra suppression for UBSan. Merge. | 2021-03-12T14:05:18 | 2021-03-12T14:05:18 | {} | MEMBER |
ClickHouse | ClickHouse | 797,672,067 | 21,613 | alexey-milovidov | > should I use pow(2, arraySum(i -> log2(i), x)) the algorithm to implement arrayProduct?
Just simply do the multiplication. | 2021-03-12T18:26:05 | 2021-03-12T18:26:05 | {} | MEMBER |
ClickHouse | ClickHouse | 797,085,535 | 19,645 | UnamedRus | Related https://github.com/ClickHouse/ClickHouse/issues/19288 | 2021-03-11T22:14:19 | 2021-03-11T22:14:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,395,963 | 20,213 | MaxJPowers | I've added `<query_wait_timeout_milliseconds>1000</query_wait_timeout_milliseconds>` to config.xml but it doesn't change the timeout value. | 2021-03-12T10:28:58 | 2021-03-12T10:28:58 | {} | NONE |
ClickHouse | ClickHouse | 797,512,604 | 18,720 | PHaroZ | I think this could lead to error with `ANY JOIN`. A query like
```sql
SELECT *
FROM
(SELECT 1 AS id) AS l
INNER ANY JOIN (SELECT 1 AS id, number as v FROM system.numbers LIMIT 2) AS r
ON l.id = r.id AND r.v = 1
```
is not the same as
```sql
SELECT *
FROM
(SELECT 1 AS id) AS l
INNER ANY JOI... | 2021-03-12T14:08:42 | 2021-03-12T14:13:56 | {} | NONE |
ClickHouse | ClickHouse | 797,526,753 | 21,650 | den-crane | The behavior of parsing incorrect dates is implementation specific. ClickHouse may return zero date, throw an exception or do “natural” overflow.
```
SELECT toDate('2020-09-31')
┌─toDate('2020-09-31')─┐
│ 2020-10-01 │
└──────────────────────┘
```
| 2021-03-12T14:32:41 | 2021-03-12T14:32:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,660,094 | 21,613 | UnamedRus | > assign to me, should I use pow(2, arraySum(i -> log2(i), x)) the algorithm to implement arrayProduct?
I don't think so, it would produce Float value from multiplying integers.
```
SELECT
[2, 3, 40] AS x,
pow(2, arraySum(i -> log2(i), x)) AS arrayProduct
Query id: 1fcc62b2-fd29-4cff-b9ed-f4e434c9... | 2021-03-12T18:01:59 | 2021-03-12T18:01:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,681,225 | 21,643 | azat | >Integration tests (thread) — fail: 1, passed: 1161, error: 0
#21676
| 2021-03-12T18:51:23 | 2021-03-12T18:51:23 | {} | MEMBER |
ClickHouse | ClickHouse | 797,717,666 | 21,613 | pp641 | It should be
accumulate(v.begin(), v.end(), 1, multiplies<int>())
This function in c++ multiplies all the element of the array in c++
vector<int>arr = {1,2,3,4};
int ans = accumulate(arr.begin(), arr.end(), 1, multiplies<int>())
ans = 24 | 2021-03-12T19:53:31 | 2021-03-12T19:54:03 | {} | NONE |
ClickHouse | ClickHouse | 796,577,621 | 21,364 | filimonov | The scenario of creating local tables is not clear for me.
1) How local tables will be mapped to particular shard / replica?
2) How they will be connected to the ImplicitDistributed?
3) How to keep set of local tables in sync with cluster definition.
I.e. imagine I want 6 shards x 3 replicas on 2 nodes. So I need... | 2021-03-11T08:56:05 | 2021-03-11T08:56:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,949,802 | 21,339 | vitlibar | https://github.com/ClickHouse/ClickHouse/pull/21641 will implement an alternative approach. | 2021-03-11T18:29:38 | 2021-03-11T18:29:38 | {} | MEMBER |
ClickHouse | ClickHouse | 797,045,348 | 21,643 | azat | >AST fuzzer (TSan) — Received signal 11
#21646 | 2021-03-11T21:00:17 | 2021-03-11T21:00:17 | {} | MEMBER |
ClickHouse | ClickHouse | 796,539,922 | 21,617 | filimonov | > binary mlock will be disabled
binary mlock = put whole clickhouse binary in ram and forbid paging.
To be enabled the Linux capability CAP_IPC_LOCK is needed. (The log message is relevant and help you to fix the issue).
When you don't have such a capability the feature can not be enabled, that potentially may... | 2021-03-11T07:50:26 | 2021-03-11T07:50:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,777,836 | 21,364 | amosbird | After some discussion with @filimonov , it looks better to eliminate the command of `create local table` and embed it into ImplicitDistributed table engine. We will have a statement like the following:
```
CREATE DISTRIBUTED TABLE table_name on cluster <> (...) Engine=ReplacingMergeTree ... Sharded by ...
```
... | 2021-03-11T14:30:37 | 2021-03-11T14:31:56 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 797,041,290 | 20,451 | siradjev | >This optimisation does make sense, though would break backward compatibility, since optimised case won't require having TTL defined per part.
It does not have to break backward compatibility as old format parts will still be usable with newer version of clickhouse that implements this improvement. | 2021-03-11T20:52:46 | 2021-03-11T20:55:12 | {} | NONE |
ClickHouse | ClickHouse | 797,193,370 | 21,268 | olgarev | @kssenii Hi! Maybe it's better to merge my PR first, and then merge the changes from the second PR into it? | 2021-03-12T02:38:01 | 2021-03-12T02:38:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,425,958 | 14,614 | vdimir | Will be fixed in https://github.com/ClickHouse/ClickHouse/pull/21640 | 2021-03-12T11:17:39 | 2021-03-12T11:17:39 | {} | MEMBER |
ClickHouse | ClickHouse | 797,554,614 | 18,720 | vdimir | @PHaroZ makes sense, thank you for your notice. We are going to fix it. | 2021-03-12T15:16:03 | 2021-03-12T15:16:03 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 797,651,778 | 9,624 | alexey-milovidov | EOL | 2021-03-12T17:47:48 | 2021-03-12T17:47:48 | {} | MEMBER |
ClickHouse | ClickHouse | 797,765,718 | 21,611 | chenziliang | Close it as we can't reproduce this issue. Will report back if it arises again. | 2021-03-12T21:28:47 | 2021-03-12T21:28:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,776,997 | 21,580 | akuzm | Sometimes we reorder these lists and we're going to get something bad if there is a concurrent modification, but probably this is not important so let's try.
Maybe the `test/queries/skip_list.json` as well?
| 2021-03-11T14:29:31 | 2021-03-11T14:29:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,886,674 | 21,580 | azat | >Maybe the test/queries/skip_list.json as well?
This is json that may be broken, because of trailing `,`, that's why I did not added union merge for it. | 2021-03-11T16:57:52 | 2021-03-11T16:57:52 | {} | MEMBER |
ClickHouse | ClickHouse | 797,027,968 | 21,579 | azat | >AST fuzzer (UBSan) — ../src/Interpreters/ExpressionActions.h:182:33: runtime error: member call on null pointer of type 'DB::ActionsDAG'
#21646 | 2021-03-11T20:29:55 | 2021-03-11T20:29:55 | {} | MEMBER |
ClickHouse | ClickHouse | 796,693,045 | 21,588 | maxistua | I have the same problem:
```
Code: 100, e.displayText() = DB::Exception: Unknown packet 0 from server XXX.XXX.XXX.XXX:9000 (version 21.2.5.5 (official build)) (from 127.0.0.1:43620) (in query: SELECT min(__version), max(__version) FROM (select __version FROM lake.storage_history WHERE __version > 161536539818018131... | 2021-03-11T12:13:14 | 2021-03-12T09:47:36 | {} | NONE |
ClickHouse | ClickHouse | 797,384,593 | 21,149 | nikitamikhaylov | @alexey-milovidov @akuzm Merge? | 2021-03-12T10:10:03 | 2021-03-12T10:10:03 | {} | MEMBER |
ClickHouse | ClickHouse | 797,445,938 | 20,451 | xjewer | >It does not have to break backward compatibility as old format parts will still be usable with newer version of clickhouse that implements this improvement.
that's forward, not backward.
I mean the use case, where optimisation took place on a new CH version (table metadata was changed, but parts weren't mutated)... | 2021-03-12T11:59:12 | 2021-03-12T11:59:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,469,678 | 21,497 | olgarev | @nikitamikhaylov Headers corrected. | 2021-03-12T12:49:16 | 2021-03-12T12:49:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,579,063 | 20,196 | vmihailenco | Gentle ping. The problem persists on 21.3.2. I begin to wonder whether this is a bug or new reality... | 2021-03-12T15:54:02 | 2021-03-12T15:54:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,662,898 | 21,613 | den-crane | It should calculate `2*3*4`
arrayProduct([2,3,4]) -> `2*3*4` = 24 | 2021-03-12T18:06:29 | 2021-03-12T18:24:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,711,190 | 21,622 | filimonov | Just remove system database from metadata/data/store folder. It will be recreated.
And Yes, just like @abyss7 said, sometimes newer versions use some newer features, which can prevent the possibility to downgrade.
But in most of cases it's possible to disable new functionality if you need to keep a possibility... | 2021-03-11T12:47:15 | 2021-03-11T12:47:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,321,344 | 21,642 | alexey-milovidov | What's wrong with the `test_replace_partition/test.py::test_replace_after_replace_failover`? | 2021-03-12T08:24:02 | 2021-03-12T08:24:02 | {} | MEMBER |
ClickHouse | ClickHouse | 797,420,812 | 20,334 | maehrlae | I contacted ridiculousfish - the author of the libdivide library. He was so kind to exchange the relevant code snippet:
"I have released libdivide v4.0.0, which no longer contains code from Stack Overflow, to the best of my knowledge. All of its code is written by me or another contributor, and should be covered un... | 2021-03-12T11:07:46 | 2021-03-12T14:17:30 | {} | NONE |
ClickHouse | ClickHouse | 786,607,243 | 21,080 | taylor12805 | > ORC & Parquet (not sure) allows Null instead of list.
> Let's implement these
>
> from ORC(list) Clickhouse back to ORC(list)
> [] [] []
> [null] [null] [null]
> null [] []
> ```python
> with open("./test4.orc", "wb") as data:
> with pyorc.Writer(data, "struct<col0:array<int>,col1:array<string>>") as wr... | 2021-02-26T12:04:54 | 2021-03-12T16:38:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,611,024 | 21,671 | UnamedRus |
```
WITH xa AS
(
SELECT *
FROM numbers(10)
)
SELECT *
FROM xa
INNER JOIN xa AS ya USING (number)
WHERE number = 1
```
Can be optimized as:
```
WITH xa AS
(
SELECT *
FROM numbers(10)
)
SELECT xa.number
FROM
(
SELECT number
FROM numbers... | 2021-03-12T16:42:23 | 2021-03-12T16:42:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,665,997 | 21,642 | azat | >What's wrong with the test_replace_partition/test.py::test_replace_after_replace_failover?
Let's try #21674 | 2021-03-12T18:11:56 | 2021-03-12T18:11:56 | {} | MEMBER |
ClickHouse | ClickHouse | 797,696,452 | 21,677 | alesapin | Where to look https://github.com/jepsen-io/etcd | 2021-03-12T19:15:59 | 2021-03-12T19:15:59 | {} | MEMBER |
ClickHouse | ClickHouse | 796,542,592 | 21,607 | filimonov | Also that remind me https://github.com/ClickHouse/ClickHouse/issues/1270 :) | 2021-03-11T07:54:48 | 2021-03-11T07:54:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,212,344 | 21,643 | azat | >Testflows check — Failed to process results
@vzakaznikov can you take a look please?
```
1 module (1 ok)
1340 suites (1338 ok, 2 xfail)
4914 scenarios (4346 ok, 568 xfail)
6981 examples (6835 ok, 80 xfail, 66 xerror)
295857 steps (294273 ok, 1443 failed, 66 errored, 75 xfail)
Total time 1h 14m
Execute... | 2021-03-12T03:37:21 | 2021-03-12T03:37:21 | {} | MEMBER |
ClickHouse | ClickHouse | 797,609,039 | 21,080 | filimonov | Try:
```
CREATE TABLE default.event_table
(
`id` String,
`extended_field` Array(Nullable(String))
) ENGINE = MergeTree()
ORDER BY tuple()
```
How the task is going? Any progress? Do you need any help? | 2021-03-12T16:39:17 | 2021-03-12T16:40:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,552,488 | 21,419 | qoega | https://clickhouse-test-reports.s3.yandex.net/0/b399f80b7e8b140ebcd552e443c8f537b5353bf3/functional_stateless_tests_(address).html
https://clickhouse-test-reports.s3.yandex.net/0/b399f80b7e8b140ebcd552e443c8f537b5353bf3/functional_stateless_tests_(address)/test_run.txt.out.log
```
2021-03-10 15:54:56 [5696bab96c6b... | 2021-03-11T08:11:28 | 2021-03-11T08:11:28 | {} | MEMBER |
ClickHouse | ClickHouse | 796,612,441 | 21,149 | nikitamikhaylov | #21612 | 2021-03-11T09:52:18 | 2021-03-11T09:52:18 | {} | MEMBER |
ClickHouse | ClickHouse | 796,816,855 | 19,857 | akuzm | By the way, it is going to be simpler to implement than I tought. Window functions can't be inside other window functions, so we can always topo-sort the execution graph into three parts:
1. Base expressions w/o window functions
2. Window functions that reference the base expressions
3. Expressions that reference ... | 2021-03-11T15:24:55 | 2021-03-11T15:24:55 | {
"hooray": 6
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,848,452 | 21,584 | amosbird | > I can, but that issue not about WITH element's it's more about JOIN rewriting. it would reproduce even without with clause.
Ok. I fail to reproduce with the above queries. Not sure why it's related to explain only. | 2021-03-11T16:07:37 | 2021-03-11T16:08:26 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 797,070,627 | 21,584 | UnamedRus | @amosbird
```
SELECT
*,
version()
FROM system.build_options
WHERE name = 'VERSION_GITHASH'
┌─name────────────┬─value────────────────────────────────────┬─version()───┐
│ VERSION_GITHASH │ 3b193363049590958e9c5b066f5f56ba392b51d7 │ 21.4.1.6210 │
└─────────────────┴──────────────────────────────────... | 2021-03-11T21:45:27 | 2021-03-11T21:45:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,315,739 | 20,058 | alexey-milovidov | Ok. I add "no backport" tag to avoid surprises in patch releases.
> if a user has their custom edited configuration
For new versions it's ok if we put a note in the changelog. | 2021-03-12T08:14:05 | 2021-03-12T08:14:05 | {} | MEMBER |
ClickHouse | ClickHouse | 797,394,895 | 10,935 | filimonov | https://github.com/ClickHouse-Extras/poco/blob/a71e0776531861873144e4c4081204ea1a1258fb/NetSSL_OpenSSL/src/SecureSocketImpl.cpp#L514-L528
https://github.com/ClickHouse/ClickHouse/blob/25f941020b578f2e9d479ae7df02d0148619fdb1/base/common/ErrorHandlers.h#L21-L37 | 2021-03-12T10:26:58 | 2021-03-12T10:26:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,645,299 | 21,643 | vzakaznikov | Try to run CI/CD again. The issue with TestFlows check will be fixed with https://github.com/ClickHouse/ClickHouse/pull/21673. | 2021-03-12T17:36:38 | 2021-03-12T17:36:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,649,616 | 10,278 | filimonov | ```
DROP TABLE IF EXISTS partition_overflow;
CREATE TABLE partition_overflow
(
date Date,
id UInt64
)
ENGINE MergeTree
PARTITION BY date
ORDER BY id;
INSERT INTO partition_overflow VALUES (toDate(49712), 49712); /* that creates 'bad' part in 21060207 partition */
INSERT INTO partition_overflow V... | 2021-03-11T10:54:34 | 2021-03-11T11:02:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,690,578 | 21,622 | abyss7 | Do you mean "downgrade to version" instead? We don't support downgrading by any means - the metadata, binary formats, etc may change without backward compatibility. | 2021-03-11T12:08:18 | 2021-03-11T12:08:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,974,512 | 21,529 | azat | >Testflows check — failed: 751, passed: 2612, other: 26
Broken in upstream too
> Functional stateless tests (memory) — Timeout :(
> Functional stateless tests (thread) — Timeout :(
Any clue? Links are under nda.
> Integration tests (asan) — fail: 3, passed: 1162, error: 0
> Integration tests (thread) — fa... | 2021-03-11T19:09:55 | 2021-03-11T19:09:55 | {} | MEMBER |
ClickHouse | ClickHouse | 797,008,243 | 21,579 | azat | >Testflows check — failed: 751, passed: 2688, other: 26
unrelated and broken | 2021-03-11T19:56:41 | 2021-03-11T19:56:41 | {} | MEMBER |
ClickHouse | ClickHouse | 797,443,792 | 14,878 | Avogar | Was fixed in https://github.com/ClickHouse/ClickHouse/pull/17365 | 2021-03-12T11:54:32 | 2021-03-12T11:54:32 | {} | MEMBER |
ClickHouse | ClickHouse | 796,612,359 | 14,995 | zvonand | @traceon @vitlibar @filimonov
Kerberos Testflows tests are disabled for now and will be added as a standalone PR later | 2021-03-11T09:52:10 | 2021-03-11T09:52:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,792,337 | 20,477 | akuzm | Maybe we can go the opposite way and split off the resource accounting instead? I.e. add a possibility to tell the local secondary query: use this resource accounting context and don't create your own. | 2021-03-11T14:51:15 | 2021-03-11T14:51:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,971,062 | 20,451 | xjewer | >When expression is monotonic function of column(s), columns min/max themselves are enough. They are already kept for partition key, primary key, so when ttl is added on one of these there is no need to do anything.
When ttl_only_drop_parts=0 parts that partially satisfy TTL expression will be mutated, when ttl_only_d... | 2021-03-11T19:04:18 | 2021-03-11T19:04:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,318,053 | 21,182 | alexey-milovidov | @kitaisreal As the issue is already fixed in another pull request, let's choose what is remaining here - maybe just merge the test? | 2021-03-12T08:17:57 | 2021-03-12T08:17:57 | {} | MEMBER |
ClickHouse | ClickHouse | 797,319,715 | 21,631 | Ash214 | UnamedRus, thanks, this is work | 2021-03-12T08:21:01 | 2021-03-12T08:21:01 | {} | NONE |
ClickHouse | ClickHouse | 797,356,956 | 19,627 | filimonov | > what should happen when we execute UPDATE on different nodes with different SET clauses but with the same conditions for WHERE clause which affect the same rows in ReplicatedMergeTree table?
Real updates are out of the scope for now (see last section in the top post). UPSERT will be executed in the order of appea... | 2021-03-12T09:22:15 | 2021-03-12T09:22:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,422,763 | 19,288 | vdimir | Third query
`SELECT *
FROM mysql.members AS m
INNER JOIN mysql.committees AS c ON m.name = c.name
WHERE `c.name` = 2`
should not work because `c.name` has type `Nullable(String)` and it cannot be compared with int.
Second query work in PR https://github.com/ClickHouse/ClickHouse/pull/21640
| 2021-03-12T11:11:31 | 2021-03-12T11:11:31 | {} | MEMBER |
ClickHouse | ClickHouse | 797,426,086 | 19,288 | UnamedRus | > should not work because c.name has type Nullable(String) and it cannot be compared with int.
Yes, it was added just to mention that CH know that c.name type is String
> ^ Clickhouse actually know that c.name type is String. | 2021-03-12T11:17:52 | 2021-03-12T11:17:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,588,586 | 21,593 | alesapin | UB reported https://github.com/ClickHouse/ClickHouse/issues/21619 | 2021-03-11T09:13:15 | 2021-03-11T09:13:15 | {} | MEMBER |
ClickHouse | ClickHouse | 796,847,237 | 21,584 | UnamedRus | I can, but that issue not about WITH element's it's more about JOIN rewriting. it would reproduce even without with clause. | 2021-03-11T16:05:59 | 2021-03-11T16:05:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,116,118 | 21,638 | den-crane | >can you create UpdatingMergeTree for update old value of record.
We cannot.
try this: https://gist.github.com/den-crane/d03524eadbbce0bafa528101afa8f794
| 2021-03-11T23:21:08 | 2021-03-11T23:21:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,455,239 | 21,643 | vzakaznikov | > > Testflows check — Failed to process results
>
> @vzakaznikov can you take a look please?
>
> ```
> 1 module (1 ok)
> 1340 suites (1338 ok, 2 xfail)
> 4914 scenarios (4346 ok, 568 xfail)
> 6981 examples (6835 ok, 80 xfail, 66 xerror)
> 295857 steps (294273 ok, 1443 failed, 66 errored, 75 xfail)
>
> Tot... | 2021-03-12T12:17:34 | 2021-03-12T12:17:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,522,501 | 21,651 | filimonov | Or just update clickhouse version. | 2021-03-12T14:25:56 | 2021-03-12T14:25:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,612,917 | 21,391 | ka1bi4 | @Avogar Hi, Pavel!
Thanks for your comments, i fixed them. | 2021-03-12T16:45:34 | 2021-03-12T16:45:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,636,631 | 21,630 | nikitamikhaylov | Too slow :( | 2021-03-12T17:21:41 | 2021-03-12T17:21:41 | {} | MEMBER |
ClickHouse | ClickHouse | 796,564,332 | 18,720 | vdimir | Closed via https://github.com/ClickHouse/ClickHouse/pull/19685 | 2021-03-11T08:32:46 | 2021-03-11T08:32:46 | {} | MEMBER |
ClickHouse | ClickHouse | 796,585,731 | 21,593 | alesapin | Database replicated `00682_empty_parts_merge` Reported to the maintainer of the feature :) | 2021-03-11T09:08:39 | 2021-03-11T09:08:48 | {} | MEMBER |
ClickHouse | ClickHouse | 796,785,744 | 21,584 | amosbird | Could you try constructing a case with this PR https://github.com/ClickHouse/ClickHouse/pull/21414 ? It doesn't suffer from this issue by removing useless WITH elements... | 2021-03-11T14:41:55 | 2021-03-11T14:41:55 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 796,789,981 | 21,505 | anmolarora | @Avogar Appreciate the detailed ellaboration. I understand that there's not bug.
Closing the issue. | 2021-03-11T14:47:54 | 2021-03-11T14:47:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,442,646 | 9,900 | Avogar | > @Avogar did you add that test?
Yes, I added a similar test:
https://github.com/ClickHouse/ClickHouse/blob/2e7f756d7c823438ff7821260f39f0f8f9aedd8e/tests/integration/test_hedged_requests/test.py#L86-L105 | 2021-03-12T11:52:07 | 2021-03-12T11:52:07 | {} | MEMBER |
ClickHouse | ClickHouse | 797,450,042 | 21,651 | UnamedRus | https://github.com/ClickHouse/ClickHouse/issues/18137
try `set read_backoff_min_latency_ms=0` | 2021-03-12T12:07:27 | 2021-03-12T12:07:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,483,742 | 21,657 | den-crane | What Clickhouse version do you use?
How we can reproduce this? Do you have SQL to run? Create table? Data sample?
Can you share `cat /etc/security/limits.d/clickhouse.conf` | 2021-03-12T13:16:53 | 2021-03-12T13:16:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,557,785 | 21,468 | ghost | @alesapin
> table schema
Do you need partition and ordering information?
```SQL
CREATE TABLE test (
code String,
id String,
date int, -- 20210308
...,
) ReplicatedMergeTree(...) PARTITION BY date ORDER BY code, id
```
> parts count in each partition
7~15 parts in a partition
> drop
I ... | 2021-03-11T08:21:29 | 2021-03-11T08:24:13 | {} | NONE |
ClickHouse | ClickHouse | 796,600,285 | 21,407 | alesapin | No related failures. | 2021-03-11T09:32:04 | 2021-03-11T09:32:04 | {} | MEMBER |
ClickHouse | ClickHouse | 797,254,026 | 21,599 | filimonov | Please also ensure that https://www.shellcheck.net/ is happy with your changes.
And it seem we need to start creating test suit for dockerfile / entrypoint itself...
| 2021-03-12T05:54:19 | 2021-03-12T05:54:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,478,579 | 9,259 | tavplubix | Fixed in #13450 | 2021-03-12T13:07:13 | 2021-03-12T13:07:13 | {} | MEMBER |
ClickHouse | ClickHouse | 797,539,845 | 12,561 | tavplubix | Fixed in 91cbd49e85ee3de6c93719e17bbf7e7e422d889e
Test will be added in #21661 | 2021-03-12T14:54:21 | 2021-03-12T14:54:21 | {} | MEMBER |
ClickHouse | ClickHouse | 797,669,495 | 21,624 | azat | BTW There is another PR (#21325) that is trying to address this | 2021-03-12T18:17:58 | 2021-03-12T18:17:58 | {
"rocket": 1
} | MEMBER |
ClickHouse | ClickHouse | 796,616,332 | 10,401 | UnamedRus | ```
SELECT groupArray(
(
SELECT dummy
FROM system.one
) AS k) AS n
FROM clusterAllReplicas('test_cluster_two_shards', system, parts)
Query id: 987a9da9-d4f1-43e3-a632-7a4ffafa7900
↗ Progress: 470.00 rows, 263.43 KB (3.95 thousand rows/s., 2.21 MB/s.)
Received exception from server (... | 2021-03-11T09:58:41 | 2021-03-11T09:58:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,482,694 | 21,350 | joein | @den-crane Sorry for my silence.
I attached sample.log, it contains last 1-2 minute before crash and some time after restart.
Maybe you can find the reason or give us an advice about the most important messages there.
Thank you!
[sample.log](https://github.com/ClickHouse/ClickHouse/files/6130179/sample.log)
| 2021-03-12T13:14:58 | 2021-03-12T13:14:58 | {} | NONE |
ClickHouse | ClickHouse | 797,536,193 | 21,470 | akuzm | А, ещё чуть не пропустил -- здесь `contrib/librdkafka` поменялась по ошибке, это я наверное сам поправлю перед мержем.
Если кликхаус из исходников не собираете, то имеет смысл вообще не делать `git submodule init`, чтобы эти файлы игнорировались. На существующем репозитории можно сделать `git submodule foreach git s... | 2021-03-12T14:48:27 | 2021-03-12T14:48:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,664,417 | 21,643 | azat | >Try to run CI/CD again.
I guess there is no need in this, CI will catch it later (if any).
> The issue with TestFlows check will be fixed with #21673.
Great, thanks! | 2021-03-12T18:09:11 | 2021-03-12T18:09:11 | {} | MEMBER |
ClickHouse | ClickHouse | 797,670,430 | 21,642 | alexey-milovidov | > P.S. looks like fsync related settings are not production ready, maybe worth adding some note about this?
Currently we don't use fsync related settings on any of production services in Yandex, because they don't make any improvement in case when geographically distributed replication is used with quorum inserts (t... | 2021-03-12T18:21:01 | 2021-03-12T18:21:01 | {} | MEMBER |
ClickHouse | ClickHouse | 796,538,798 | 12,701 | i4late | you have mysql engine tables ? | 2021-03-11T07:48:32 | 2021-03-11T07:48:32 | {} | NONE |
ClickHouse | ClickHouse | 796,543,876 | 21,594 | filimonov | It maybe related somehow. See
#16918
#2261
#18448
#17556 | 2021-03-11T07:57:05 | 2021-03-11T07:57:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,571,687 | 17,649 | vdimir | After this change https://github.com/ClickHouse/ClickHouse/pull/19885 this `CREATE` query doesn't produce exception. Type of column in the result table would be supertype for left and right tables. If there's no supertype, `CREATE` throws exception, but table is not created. | 2021-03-11T08:45:52 | 2021-03-11T08:45:52 | {} | MEMBER |
ClickHouse | ClickHouse | 796,585,874 | 21,599 | garutilorenzo | Hi @filimonov
> * Motivation is not 100% clear. The only clear improvement i see is ability to source the entrypoint w/o executing it. So you can for example create some adjusted entrypoint which will source the main one. It's nice and common practice.
my intention was to skip some step if the container already ... | 2021-03-11T09:08:52 | 2021-03-11T09:08:52 | {} | NONE |
ClickHouse | ClickHouse | 797,324,294 | 21,457 | alexey-milovidov | This is "improvement", not a bugfix. | 2021-03-12T08:29:13 | 2021-03-12T08:29:13 | {} | MEMBER |
ClickHouse | ClickHouse | 796,606,012 | 21,364 | amosbird | > How local tables will be mapped to particular shard / replica?
It'll be a deterministic algorithm that deduces the local table name out from `shard_id + distributed table name`. The mapping is thus defined implicitly.
> How they will be connected to the ImplicitDistributed?
ImplicitDistributed shares the sa... | 2021-03-11T09:41:31 | 2021-03-11T09:41:31 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 796,641,686 | 10,278 | filimonov | The reason:
```
SELECT
toUInt16(number) + 49700 AS k,
toDate(k) AS d
FROM numbers(20)
┌─────k─┬──────────d─┐
│ 49700 │ 2106-01-28 │
│ 49701 │ 2106-01-29 │
│ 49702 │ 2106-01-30 │
│ 49703 │ 2106-01-31 │
│ 49704 │ 2106-02-01 │
│ 49705 │ 2106-02-02 │
│ 49706 │ 2106-02-03 │
│ 49707 │ 2106-02-04 │
... | 2021-03-11T10:40:03 | 2021-03-11T10:41:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 796,994,577 | 21,631 | UnamedRus | ```
WITH
groupArray((date_start, date_end, value)) AS arr,
arrayJoin(arraySplit((x, y) -> (y > 900), arr, arrayDifference(CAST(arr.1, 'Array(UInt32)')))) AS splitted
SELECT
arrayReduce('min', splitted.1) AS start,
arrayReduce('max', splitted.2) AS end,
arraySum(splitted.3) AS sum
FROM
(
... | 2021-03-11T19:35:27 | 2021-03-11T19:35:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,364,815 | 21,613 | hexiaoting | assign to me, should I use `pow(2, arraySum(i -> log2(i), x))` the algorithm to implement arrayProduct? | 2021-03-12T09:34:20 | 2021-03-12T09:48:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 797,526,086 | 21,648 | filimonov | > there are only 2 ways to improve consuming performance, adding consumer('s threads) or increasing the value of
There are more:
1) update clickhouse version
2) adjust [librdkafka settings](https://github.com/edenhill/librdkafka/blob/master/CONFIGURATION.md) namely queued.min.messages and queued.max.messages.kbyt... | 2021-03-12T14:31:37 | 2021-03-12T14:31:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,562,490 | 15,641 | alexey-milovidov | Ok. | 2020-11-05T18:35:39 | 2020-11-05T18:35:39 | {} | MEMBER |
ClickHouse | ClickHouse | 722,604,193 | 16,231 | den-crane | hmm?
test : ([1],[5]) 4 4
expected: ([1],[4]) │ 4 │ 4
not 1,5 | 2020-11-05T19:46:52 | 2020-11-05T19:47:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 722,622,789 | 16,636 | alexey-milovidov | @hombit The test does not work. | 2020-11-05T20:23:08 | 2020-11-05T20:23:08 | {} | MEMBER |
ClickHouse | ClickHouse | 723,080,475 | 16,132 | zhang2014 | > Could you please also add an integration test for this?
Ok, I will add it . | 2020-11-06T13:27:59 | 2020-11-06T13:27:59 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.