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 | 525,378,889 | 6,690 | PlatonB | Сделал одинарными - тот же результат. | 2019-08-27T16:25:53 | 2019-08-27T16:25:53 | {} | NONE |
ClickHouse | ClickHouse | 525,804,381 | 6,705 | 4ertus2 | fixes #6284 #4877 | 2019-08-28T15:45:33 | 2019-08-28T15:45:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,836,836 | 6,638 | SaltTan | A few examples.
A 'group by' query: https://gist.github.com/SaltTan/cc5e5e440432d9ed630a772528f94908
```
SELECT f_string
FROM distributed_table
WHERE partition_key = '2019-08-23' AND pk_f1 = 118129 AND pk_f2= -2
GROUP BY f_string LIMIT 10
```
A 'group by' query with two-level aggregation: https://gist.github.... | 2019-08-28T17:10:56 | 2019-08-28T17:15:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,945,980 | 6,716 | alexey-milovidov | `clickhouse-benchmark` run queries continuously using specified number of connections (one connection by default, controlled by `-c` argument). If you run it with single connection (by default), it will perform single query at any moment of time, and measure latency, not throughput. Latency will include network delays.... | 2019-08-28T22:25:44 | 2019-08-28T22:25:44 | {} | MEMBER |
ClickHouse | ClickHouse | 525,299,577 | 6,688 | alesapin | Сообщение, которое выкидывается при неправильном указании параметров движка, очень страшное. Особенно в данном примере, когда не указаны ни параметры, ни `ORDER BY...` показывается именно старый вариант, хотя разумнее бы было новый:
```
Examples:
MergeTree(EventDate, (CounterID, EventDate), 8192)
MergeTree(Even... | 2019-08-27T13:24:47 | 2019-08-27T13:26:03 | {} | MEMBER |
ClickHouse | ClickHouse | 525,376,339 | 6,690 | den-crane | "uiop"
Двойные кавычки это обозначение литерала -- объекта базы данных (имя поля, имя таблицы)
select "поле" from "таблица"
для значений строк используются только одинарные кавычки
VALUES ('qwerty', 1000) | 2019-08-27T16:19:20 | 2019-08-27T16:19:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,438,902 | 6,678 | alexey-milovidov | Overestimation of free space is fine. | 2019-08-27T19:00:31 | 2019-08-27T19:00:31 | {} | MEMBER |
ClickHouse | ClickHouse | 525,440,148 | 6,691 | alexey-milovidov | TODO the setting was added in a group of obsolete settings. | 2019-08-27T19:03:51 | 2019-08-27T19:03:51 | {} | MEMBER |
ClickHouse | ClickHouse | 525,623,589 | 1,661 | blinkov | @MeteHanC mutations are done asynchronously, it just got scheduled that fast, not executed. To check how they are going you need to look into `system.mutations` table. | 2019-08-28T07:40:24 | 2019-08-28T07:40:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,778,539 | 4,877 | 4ertus2 | #6705 | 2019-08-28T14:45:33 | 2019-08-28T15:46:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,876,890 | 6,714 | MeteHanC | I think this could be related to : #6496 | 2019-08-28T18:55:34 | 2019-08-28T18:55:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,596,093 | 6,594 | WestDragon | Есть два уточнения которые возможно помогут, в моём случае база запущена в докере и таблица заполнена через запрос:
`INSERT INTO invoice.esf_item_test
SELECT
id,
general_number,
general_issuance_date,
name,
code,
oced,
description,
units,
quantity,
price,
cost,
summa_excise,
summa_vat,
... | 2019-08-28T05:57:45 | 2019-08-28T05:57:45 | {} | NONE |
ClickHouse | ClickHouse | 525,598,158 | 3,490 | lapkofear | Ch version: 19.13.1.11. For example above difference is ~1s on AMD vs 60ms on Intel. Symptoms are the same, but test from your commit does not show any degradation. | 2019-08-28T06:07:40 | 2019-08-28T06:07:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,866,101 | 6,205 | alexey-milovidov | I understand the idea and the code looks Ok. Could you please merge with master? | 2019-08-28T18:26:07 | 2019-08-28T18:26:07 | {} | MEMBER |
ClickHouse | ClickHouse | 525,945,838 | 6,716 | den-crane | >Local Table
>**SETTINGS index_granularity = 128** !!!!
| 2019-08-28T22:25:14 | 2019-08-28T22:25:14 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,187,588 | 6,661 | blitzarx1 | @filimonov After I changed the version to 19.13.3.26 everything works as expected. | 2019-08-27T07:59:55 | 2019-08-27T07:59:55 | {} | NONE |
ClickHouse | ClickHouse | 525,290,105 | 6,688 | alexey-milovidov | Параметры в MergeTree не требуются, но при этом обязательным является указание ORDER BY (ключа сортировки). | 2019-08-27T13:01:00 | 2019-08-27T13:01:00 | {} | MEMBER |
ClickHouse | ClickHouse | 525,346,531 | 1,522 | seanlook | Got the same error in distributed table:
```
0. clickhouse-server(StackTrace::StackTrace()+0x16) [0x65cbc16]
1. clickhouse-server(DB::RemoteBlockOutputStream::RemoteBlockOutputStream(DB::Connection&, std::string const&, DB::Settings const*)+0xa13) [0x60a50c3]
2. clickhouse-server(DB::StorageDistributedDirectoryMoni... | 2019-08-27T15:10:56 | 2019-08-27T15:15:37 | {} | NONE |
ClickHouse | ClickHouse | 525,439,823 | 6,691 | alexey-milovidov | TODO change setting name to avoid ambiguouty. | 2019-08-27T19:02:56 | 2019-08-27T19:02:56 | {} | MEMBER |
ClickHouse | ClickHouse | 525,714,206 | 6,667 | alexey-milovidov | (at least it should fix performance regression for AMD) | 2019-08-28T12:04:03 | 2019-08-28T12:04:03 | {} | MEMBER |
ClickHouse | ClickHouse | 525,577,739 | 6,562 | MetikovVadim | Insertion took the same time:
0 rows in set. Elapsed: 48626.413 sec. Processed 50.60 billion rows, 6.06 TB (1.04 million rows/s., 124.68 MB/s.)
and
0 rows in set. Elapsed: 48974.664 sec. Processed 50.60 billion rows, 6.06 TB (1.03 million rows/s., 123.79 MB/s.)
After INSERT I run OPTIMIZE TABLE ... PARTITION ... ... | 2019-08-28T04:20:01 | 2019-08-28T04:20:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,740,003 | 26 | filimonov | https://github.com/yandex/ClickHouse/issues/6697 | 2019-08-28T13:15:14 | 2019-08-28T13:15:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,224,646 | 6,168 | alexey-milovidov | The concept is Ok. Yes, it will be better to add support for Replicated tables here. | 2019-08-27T09:39:46 | 2019-08-27T09:39:46 | {} | MEMBER |
ClickHouse | ClickHouse | 525,225,058 | 6,562 | MetikovVadim | I have recreated tables with all columns. During INSERT(5% of both tables) i get very short merge:
┌─table──────────┬─elapsed1┬progress1┬estimated─┬─result_part_name─┐
│ graphite_day_test2 │ 0.13 │ 0.07 │ 1.64 │ 20190822_2668_2803_27 │
│ graphite_reverse_test3 │ 0.04 │ 0.01 │ 2.22 │ 2... | 2019-08-27T09:40:52 | 2019-08-27T09:40:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,290,800 | 6,688 | PlatonB | Но он же в документации в квадратных скобках:
> CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster]
> (
> name1 [type1] [DEFAULT|MATERIALIZED|ALIAS expr1] [TTL expr1],
> name2 [type2] [DEFAULT|MATERIALIZED|ALIAS expr2] [TTL expr2],
> ...
> INDEX index_name1 expr1 TYPE type1(...) GR... | 2019-08-27T13:02:52 | 2019-08-27T13:02:52 | {} | NONE |
ClickHouse | ClickHouse | 525,300,082 | 6,594 | vmarkovtsev | I can give you a 300GB dump of my table (it is not secret). I am tar-ing it. | 2019-08-27T13:26:00 | 2019-08-27T13:26:00 | {} | NONE |
ClickHouse | ClickHouse | 525,306,633 | 6,657 | samz406 | I updated ck, it's only execute select sql is ok ,but create view it's not work.
| 2019-08-27T13:41:09 | 2019-08-27T13:41:09 | {} | NONE |
ClickHouse | ClickHouse | 525,423,858 | 5,014 | alexey-milovidov | @243f6a8885a308d313198a2e037 Sorry, I missed your answer.
This issue will be solved (at least for some of these cases) in https://github.com/yandex/ClickHouse/issues/6045 | 2019-08-27T18:21:51 | 2019-08-27T18:21:51 | {
"heart": 1,
"hooray": 1,
"laugh": 1
} | MEMBER |
ClickHouse | ClickHouse | 525,439,443 | 6,691 | alexey-milovidov | My changes:
- fixed overflow in calculation of space;
- changed variable name to conform code style;
- better exception message;
- don't do this check if ClickHouse happens to build with old Poco library. | 2019-08-27T19:01:59 | 2019-08-27T19:01:59 | {} | MEMBER |
ClickHouse | ClickHouse | 525,524,028 | 6,343 | alexey-milovidov | TODO: rename T_test to StudentTTest, move code to .cpp. | 2019-08-27T23:38:38 | 2019-08-27T23:38:38 | {} | MEMBER |
ClickHouse | ClickHouse | 525,612,093 | 6,671 | MeteHanC | > Yes, FINAL affects performance dramatically.
Thank you! And this issue could be helpful as well : #1661 | 2019-08-28T07:02:49 | 2019-08-28T07:02:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,712,533 | 6,667 | akuzm | The regressions are bigger than improvements. The biggest are in queries that materialize constant columns. Since we can overallocate memory for columns by a large margin, populating it does more harm than good. Maybe we should populate only for arenas, which use smaller chunks. | 2019-08-28T11:59:07 | 2019-08-28T11:59:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,834,370 | 6,708 | 4ertus2 | Closes #4858 | 2019-08-28T17:03:52 | 2019-08-28T17:04:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,849,490 | 6,707 | jcdang | PR: https://github.com/yandex/ClickHouse/pull/6710 | 2019-08-28T17:44:18 | 2019-08-28T17:44:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,217,356 | 2,507 | filimonov | Problem is still here.
Simple test case:
```
SELECT
'settings' AS source,
*
FROM system.settings
UNION ALL
SELECT
'settings' AS source,
*
FROM remote('127.0.0.1', 'system', 'settings')
Received exception from server (version 19.13.3):
Code: 171. DB::Exception: Received from localho... | 2019-08-27T09:19:55 | 2019-08-27T09:21:48 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,274,005 | 6,661 | filimonov | > @filimonov After I changed the version to 19.13.3.26 everything works as expected.
@bl17zar Thank you for the feedback.
@bespilotnayakartoshka unfortunatelly that information is not enough to reproduce the problem.
Please describe your setup. How many CH nodes, how many Kafka brokers, Kafka versions, how man... | 2019-08-27T12:16:10 | 2019-08-27T12:17:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,374,791 | 6,678 | weiqxu | Thanks for comments.
Does the team has the plan to implement it, or I can implement it now, since we want to use it in the production environment.
As my understand from your comment, we can implement LimitWriteBuffer, the origin code
`WriteBufferFromFile file_buf(path);`
`CompressedWriteBuffer compressed_buf(... | 2019-08-27T16:15:24 | 2019-08-27T16:27:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,393,828 | 6,594 | vmarkovtsev | Here is the table: https://drive.google.com/file/d/1oyeD-oNuDkyKyhYXvdaXf_SeoUS4w0dK (230GB)
How to reproduce:
```
ALTER TABLE uasts ADD INDEX value value bloom_filter(0.01) granularity 4;
OPTIMIZE TABLE uasts FINAL;
// 3 hours later...
SELECT count(*) FROM uasts WHERE value='numpy';
``` | 2019-08-27T17:04:53 | 2019-08-27T17:05:05 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 525,598,714 | 6,549 | golubev-ml | I have same issue - sometimes my mutations get stuck and I get
`Code: 76, e.displayText() = DB::ErrnoException: Cannot open file /var/lib/clickhouse/data/database/table/tmp_mut_20190817_20190819_11342_12974_9_13562/column.mrk, errno: 13, strerror: Permission denied (version 19.11.3.11 (official build))`
I run the se... | 2019-08-28T06:10:12 | 2019-08-28T06:10:12 | {} | NONE |
ClickHouse | ClickHouse | 525,627,503 | 1,661 | MeteHanC | > mutations are done asynchronously, it just got scheduled that fast, not executed. To check how they are going you need to look into `system.mutations` table.
I understand. I checked it and it is being done in no time. 'done' column was set to '1' immediately and it was also replicated to other CH nodes instantly a... | 2019-08-28T07:52:35 | 2019-08-28T07:52:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,679,627 | 3,490 | alexey-milovidov | Does this fix help https://github.com/yandex/ClickHouse/pull/6667 ?
(Go to "ClickHouse build check" and download a package for test). | 2019-08-28T10:16:39 | 2019-08-28T10:16:39 | {} | MEMBER |
ClickHouse | ClickHouse | 525,739,426 | 6,697 | filimonov | Correlated subqueries are not supported, and currently no schedule for implementing that.
You can rewrite your query using JOIN.
P.S. We still have no issue to track requests for that feature. Let's use this one. | 2019-08-28T13:13:49 | 2019-08-28T14:39:33 | {
"+1": 1,
"confused": 1,
"eyes": 4
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,828,280 | 4,858 | 4ertus2 | ```
SELECT toUInt64(1) x FROM (select 1) group by 1 HAVING x in (SELECT countIf(y, z == 1) FROM (SELECT 1 y, 1 z))
┌─x─┐
│ 1 │
└───┘
```
It looks like it's already fixed somewhere. I'll add a test. | 2019-08-28T16:46:55 | 2019-08-28T16:46:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,387,666 | 6,690 | den-crane | ну и table и function зарезервированные слова.
не надо так называть таблицы
поэтому
```
либо
INSERT INTO table table(col1, col2)
VALUES ('qwerty', 1000), ('uiop', 100), ('asdf', 400), ('ghjkl', 700), ('zxcv', 600)
либо
INSERT INTO "table"(col1, col2)
VALUES ('qwerty', 100... | 2019-08-27T16:48:56 | 2019-08-27T16:51:31 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,720,422 | 3,430 | alesapin | > why do we need global index for subquery? this will cause random _subqueryXX vary if it appear in select list run to run.
The root cause is expressions compilation cache where we store pairs `expression_hash -> bytecode`. As I remember subqueries execute before compilation. The result came to the compiler as actio... | 2019-08-28T12:21:39 | 2019-08-28T12:21:39 | {} | MEMBER |
ClickHouse | ClickHouse | 525,859,504 | 6,185 | alexey-milovidov | @KochetovNicolai You had to look at this PR one month ago. | 2019-08-28T18:09:05 | 2019-08-28T18:09:05 | {} | MEMBER |
ClickHouse | ClickHouse | 525,880,232 | 5,960 | alexey-milovidov | Is it obsolete? | 2019-08-28T19:05:02 | 2019-08-28T19:05:02 | {} | MEMBER |
ClickHouse | ClickHouse | 525,938,023 | 6,710 | alexey-milovidov | Three tests have failed. | 2019-08-28T21:55:06 | 2019-08-28T21:55:06 | {} | MEMBER |
ClickHouse | ClickHouse | 525,217,190 | 6,168 | YiuRULE | Ok no problem :)
Maybe could I have a confirmation about the workflow ? if it's ok I will probably work already on a `MOVE` command concerning the `Replicated` family table, this way the reviewer will have only one PR to review instead of two | 2019-08-27T09:19:27 | 2019-08-27T09:19:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,291,847 | 6,685 | alesapin | Failed test in debug-mode because debug build is too slow. Will disable it in CI for debug build (it doesn't affect coverage). | 2019-08-27T13:05:31 | 2019-08-27T13:05:31 | {} | MEMBER |
ClickHouse | ClickHouse | 525,296,323 | 6,594 | alexey-milovidov | > I wonder if this has something to do with my FS mount options
It should not relate. | 2019-08-27T13:17:06 | 2019-08-27T13:17:06 | {} | MEMBER |
ClickHouse | ClickHouse | 525,435,116 | 6,678 | alexey-milovidov | > Does the team has the plan to implement it, or I can implement it now, since we want to use it in the production environment.
We don't have a plan to implement it in our team.
> As my understand from your comment, we can implement LimitWriteBuffer, the origin code
WriteBufferFromFile file_buf(path);
Compresse... | 2019-08-27T18:50:52 | 2019-08-27T19:07:12 | {} | MEMBER |
ClickHouse | ClickHouse | 525,445,247 | 6,678 | alexey-milovidov | I think, this implementation is Ok, no need to implement LimitWriteBuffer. | 2019-08-27T19:18:39 | 2019-08-27T19:18:39 | {} | MEMBER |
ClickHouse | ClickHouse | 525,649,595 | 6,661 | mikekytyzov | @filimonov
So. I have 2 shards with 2 replicas each with current table setup:
```
CREATE TABLE my_super_table ON CLUSTER my_cluster
(some_id UInt64,
....
created_at DateTime)
ENGINE = ReplicatedReplacingMergeTree('/clickhouse/tables/{shard}/my_super_table', '{re... | 2019-08-28T08:54:50 | 2019-08-28T11:46:10 | {} | NONE |
ClickHouse | ClickHouse | 525,798,501 | 6,689 | alexey-milovidov | What did @vitlibar say? | 2019-08-28T15:31:25 | 2019-08-28T15:31:25 | {} | MEMBER |
ClickHouse | ClickHouse | 525,202,801 | 6,562 | MetikovVadim | On 5% I catched one merge on non-rollup table:
┌─table───────┬elapsed1┬progress1┬estimated┬result_part_name──────┐
│ graphite_reverse │ 3.7 │ 0.85 │ 0.64 │ 20190827_4177_4193_2 │
└───────────┴──────┴──────┴───────┴────────────────┘
| 2019-08-27T08:41:45 | 2019-08-27T08:41:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,229,104 | 4,689 | kirill001 | The same issue with DataGrip and DBeaver on Ubuntu 19.04.
Here is an error from /var/log/clickhouse-server/clickhouse-server.err.log
> 2019.08.27 11:42:05.663008 [ 20 ] {} <Error> virtual DB::WriteBufferFromHTTPServerResponse::~WriteBufferFromHTTPServerResponse(): Poco::Exception. Code: 1000, e.code() = 0, e.displa... | 2019-08-27T09:51:58 | 2019-08-27T09:52:25 | {} | NONE |
ClickHouse | ClickHouse | 525,300,209 | 6,688 | alexey-milovidov | Давай переделаем в этом сообщении на актуальный синтаксис. Наверное @BayoNet сможет закоммитить? | 2019-08-27T13:26:16 | 2019-08-27T13:26:16 | {} | MEMBER |
ClickHouse | ClickHouse | 525,573,366 | 6,562 | MetikovVadim | After INSERT I run OPTIMIZE TABLE ... PARTITION ...
on both tables. I get :
┌─table─────────┬─elapsed1┬progress1┬estimated┬result_part_name─┐
│ graphite_reverse_test3 │ 6.88 │ 0.6 │ 4.49 │ 20190822_1_48245_104 │
│ graphite_day_test2 │ 4.54 │ 0.02 │ 211.24 │ 20190822_1_48243_82 │
└─... | 2019-08-28T03:53:33 | 2019-08-28T03:53:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,619,424 | 6,461 | heneyin | Good, thanks! | 2019-08-28T07:27:18 | 2019-08-28T07:27:18 | {} | NONE |
ClickHouse | ClickHouse | 525,856,960 | 6,710 | alexey-milovidov | Could you please also add a test?
You can use `.sh` test. Similar tests are located in `dbms/tests/queries/0_stateless`. | 2019-08-28T18:02:23 | 2019-08-28T18:02:23 | {} | MEMBER |
ClickHouse | ClickHouse | 525,876,449 | 6,638 | SaltTan | It turned out the data type was different for the 'group by' column: String in the local table, LowCardinality(String) in the distributed table.
Here is a test case to reproduce (I couldn't reproduce without a cluster of two servers):
```
<remote_servers>
<test>
<shard>
<re... | 2019-08-28T18:54:20 | 2019-08-28T18:54:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,197,684 | 6,683 | woozhijun |
> It looks like it's connected to some query from system.parts. Can you try to find that query in /var/log/clickhouse-server/clickhouse-server.log ? It should be somewhere before the segfault.
I check the log and found,
`The part /clickhouse/tables/shard/nx_access/replicas/clickhouse6-master/parts/20190726_100844... | 2019-08-27T08:27:41 | 2019-08-27T08:27:41 | {} | NONE |
ClickHouse | ClickHouse | 525,527,889 | 6,692 | alexey-milovidov | The setting `queue_max_wait_ms` is overloaded.
Need to add three more settings:
- replace_running_query_max_wait_ms;
- kafka_max_wait_ms;
- connection_max_wait_ms.
(TODO: figure out better names) | 2019-08-27T23:57:11 | 2019-08-27T23:57:24 | {} | MEMBER |
ClickHouse | ClickHouse | 525,708,207 | 322 | akonyaev90 | Released in
https://github.com/yandex/ClickHouse/pull/3172
https://github.com/yandex/ClickHouse/pull/2948
I think this issue can be closed | 2019-08-28T11:46:14 | 2019-08-28T11:46:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,714,111 | 6,667 | alexey-milovidov | You can populate only for hash tables (adding a flag similar to `clear_memory_`). | 2019-08-28T12:03:43 | 2019-08-28T12:03:43 | {} | MEMBER |
ClickHouse | ClickHouse | 525,224,949 | 6,659 | yingfeng | It's the same issue with [6575](https://github.com/yandex/ClickHouse/issues/6575), we have a solution right now, to be discussed there | 2019-08-27T09:40:37 | 2019-08-27T09:40:37 | {} | NONE |
ClickHouse | ClickHouse | 525,292,871 | 6,594 | alesapin | Can't reproduce it locally with simple fake data :(. I will be glad If anybody can write a script to reproduce the issue. Still investigating. | 2019-08-27T13:08:12 | 2019-08-27T13:08:12 | {} | MEMBER |
ClickHouse | ClickHouse | 525,313,283 | 6,667 | alexey-milovidov | Please merge with master for performance test to run. | 2019-08-27T13:56:50 | 2019-08-27T13:56:50 | {} | MEMBER |
ClickHouse | ClickHouse | 525,391,372 | 6,594 | alesapin | Reproduced possible another, but related issue:
```
:) SET allow_experimental_data_skipping_indices=1;
:) CREATE TABLE some_table
(
`Key` UInt64,
`Val1` UInt64,
`Val2` UInt64,
INDEX i1 Val1 TYPE minmax GRANULARITY 1
)
ENGINE = MergeTree()
ORDER BY Key
SETTINGS vertical_merge_algorithm_min_... | 2019-08-27T16:58:33 | 2019-08-27T16:58:33 | {} | MEMBER |
ClickHouse | ClickHouse | 525,837,978 | 6,638 | SaltTan | There is no crash with `SET distributed_aggregation_memory_efficient = 0` | 2019-08-28T17:13:57 | 2019-08-28T17:13:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,862,077 | 6,205 | alexey-milovidov | Sorry, I forgot about this PR. Will take a look. | 2019-08-28T18:15:53 | 2019-08-28T18:15:53 | {} | MEMBER |
ClickHouse | ClickHouse | 525,605,434 | 6,549 | thyn | Did chmod help you? Did you try? | 2019-08-28T06:38:38 | 2019-08-28T06:38:38 | {} | NONE |
ClickHouse | ClickHouse | 525,620,602 | 1,661 | MeteHanC | Why 'Alter table update' was not suggested in this case? Correct me if I am wrong but it says **_'... mutations are intended for heavy operations that change a lot of rows in a table'_** on the offcial [documentation](https://clickhouse.yandex/docs/en/query_language/alter/#alter-mutations).
It is not that much but I... | 2019-08-28T07:30:57 | 2019-08-28T07:31:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,801,943 | 6,699 | filimonov | It's good now (also for `merge_tree_settings`).
| 2019-08-28T15:39:32 | 2019-08-28T15:39:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,238,673 | 6,562 | MetikovVadim | On 10% we could see the different merge duration:
┌─table──────────┬elapsed1┬progress1┬estimated┬─result_part_name─┐
│ graphite_reverse_test3 │ 0.41 │ 0.24 │ 1.27 │ 20190822_4971_5132_30 │
│ graphite_day_test2 │ 3.15 │ 0.63 │ 1.79 │ 20190822_4968_5120_30 │
| 2019-08-27T10:19:22 | 2019-08-27T10:19:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,267,055 | 6,500 | abyss7 | > FYI. When I tried to use `link_libraries`, CMake placed these libraries in two places: at the front of all other libraries and at the end. Example:
>
> `glibc_compatibility.a libc.so other libraries glibc_compatibility.a libc.so`
>
> And it doesn't work properly.
I hope I overcame it using INTERFACE librarie... | 2019-08-27T11:53:54 | 2019-08-27T11:53:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,297,045 | 6,689 | alesapin | In this pull request, I'm trying to improve `SettingsCollection` interface with the help of @vitlibar. Do not merge until he will take a look. | 2019-08-27T13:18:50 | 2019-08-27T13:18:50 | {} | MEMBER |
ClickHouse | ClickHouse | 525,323,684 | 6,602 | abyss7 | Which exact feature required cmake ≥ 3.12 ? | 2019-08-27T14:21:05 | 2019-08-27T14:21:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,546,274 | 4,672 | sundy-li | Sorry, I will make another PR later. | 2019-08-28T01:31:19 | 2019-08-28T01:31:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,678,815 | 6,699 | alexey-milovidov | CC @vitlibar @alesapin | 2019-08-28T10:14:31 | 2019-08-28T10:14:31 | {} | MEMBER |
ClickHouse | ClickHouse | 525,807,549 | 3,239 | 4ertus2 | #6697 | 2019-08-28T15:52:56 | 2019-08-28T15:52:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,210,063 | 6,562 | MetikovVadim | That is the root of error:
OPTIMIZE TABLE graphite_day_test2 PARTITION '2019-08-22'
Received exception from server (version 19.5.4):
Code: 10. DB::Exception: Received from localhost:9000, 127.0.0.1. DB::Exception: Not found column Time in block. There are only columns: Date, Path.
| 2019-08-27T09:01:02 | 2019-08-27T09:01:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,254,104 | 6,628 | alesapin | Our CI uses exactly ubuntu 18.04 and checks every commit with two different compilers (`gcc-9` and `clang-8`) and multiple cmake options. You can see build reports in status reference for each commit in master and PR. Example https://clickhouse-builds.s3.yandex.net/0/7985270624e877de15f958f0e3f19a23365f34de/report.html... | 2019-08-27T11:09:30 | 2019-08-27T11:09:48 | {} | MEMBER |
ClickHouse | ClickHouse | 525,291,729 | 6,686 | alexey-milovidov | Version 19.14 master is also affected. | 2019-08-27T13:05:11 | 2019-08-27T13:05:11 | {} | MEMBER |
ClickHouse | ClickHouse | 525,294,612 | 6,598 | BayoNet | @zhang2014 @sundy-li
Hello. I've inserted the English block into docs/zh/query_language/create.md because the docs build fails without of it. Add it to your translation plan please :-). | 2019-08-27T13:12:39 | 2019-08-27T13:12:39 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,297,878 | 6,688 | alexey-milovidov | @BayoNet Требуется пояснение в документации.
Если используется актуальный способ задания свойств MergeTree, то ORDER BY является обязательным. | 2019-08-27T13:20:48 | 2019-08-27T13:21:01 | {
"+1": 3
} | MEMBER |
ClickHouse | ClickHouse | 525,312,866 | 6,671 | alexey-milovidov | Yes, FINAL affects performance dramatically. | 2019-08-27T13:55:53 | 2019-08-27T13:55:53 | {} | MEMBER |
ClickHouse | ClickHouse | 525,330,229 | 6,678 | alexey-milovidov | I thought it should be checked while writing. We may implement `LimitWriteBuffer` like `LimitReadBuffer` and wrap the most deep WriteBuffer (that actually writes to file) in it. | 2019-08-27T14:35:16 | 2019-08-27T14:35:16 | {} | MEMBER |
ClickHouse | ClickHouse | 525,728,503 | 3,490 | alexey-milovidov | Also we need to test for hypothesis that I had one year ago:
- page faults happen more frequently than expected: after a page was read with page fault, subsequent read from the same page may introduce another page fault (that should not happen but happens on AMD processor under unknown circumstances). | 2019-08-28T12:44:56 | 2019-08-28T12:44:56 | {} | MEMBER |
ClickHouse | ClickHouse | 525,948,171 | 6,716 | ghost | ok thanks.
Also related to this:
>Local Table
SETTINGS index_granularity = 128 !!!!
I was testing limiting the index granularity for smaller return sets/single record lookups based on these posts:
https://www.altinity.com/blog/clickhouse-in-the-storm-part-1
https://www.altinity.com/blog/clickhouse-in-the-... | 2019-08-28T22:35:09 | 2019-08-28T22:35:09 | {} | NONE |
ClickHouse | ClickHouse | 525,188,448 | 6,562 | MetikovVadim | I run INSERT ... SELECT ...
Both of requests are completed less than 1% and i get 1 running merge :
SELECT
table,
floor(elapsed, 2) AS elapsed1,
floor(progress, 2) AS progress1,
floor((elapsed / progress) * (1 - progress), 2) AS estimated,
result_part_name
FROM system.merges
ORDER BY ... | 2019-08-27T08:02:21 | 2019-08-27T08:02:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,206,493 | 6,562 | MetikovVadim | The experiment was interrupted on 6% with message:
Code: 252. DB::Exception: Received from localhost:9000, 127.0.0.1. DB::Exception: Too many parts (3000). Merges are processing significantly slower than inserts..
in both cases. Why CH does not merge inserted data ? | 2019-08-27T08:51:52 | 2019-08-27T08:52:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,211,123 | 6,683 | alexey-milovidov | This bug is fixed in https://github.com/yandex/ClickHouse/pull/6514 and backported to 19.13, 19.11. | 2019-08-27T09:03:50 | 2019-08-27T09:03:50 | {} | MEMBER |
ClickHouse | ClickHouse | 525,291,895 | 6,686 | alexey-milovidov | ```
/home/milovidov/work/ClickHouse/contrib/libcxx/include/vector:1542: _LIBCPP_ASSERT '__n < size()' failed. vector[] index out of bounds
2019.08.27 16:04:52.013159 [ 56 ] {} <Fatal> BaseDaemon: ########################################
2019.08.27 16:04:52.013353 [ 56 ] {} <Fatal> BaseDaemon: (version 19.14.1.1) (fr... | 2019-08-27T13:05:38 | 2019-08-27T13:05:38 | {} | MEMBER |
ClickHouse | ClickHouse | 525,303,132 | 6,594 | alesapin | I already have examples of broken columns (thanks to @EvgenyVinogradov), and the problem is clear -- in some cases, we don't write marks for index columns. But according to reports the amount of missed marks is always different and I cannot find any similarities between them...
| 2019-08-27T13:33:16 | 2019-08-27T13:33:16 | {} | MEMBER |
ClickHouse | ClickHouse | 525,697,650 | 3,430 | chenxing-xc | why do we need global index for subquery? this will cause random _subqueryXX vary if it appear in select list run to run. | 2019-08-28T11:11:37 | 2019-08-28T11:11:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 525,796,909 | 6,692 | alexey-milovidov | Added all the required settings. | 2019-08-28T15:27:52 | 2019-08-28T15:27:52 | {} | MEMBER |
ClickHouse | ClickHouse | 525,883,859 | 6,710 | alexey-milovidov | Ok. And we need `reference` file for the test. It is just stdout:
`./00936_multiquery_http.sh > 00936_multiquery_http.reference` | 2019-08-28T19:14:38 | 2019-08-28T19:14:50 | {} | MEMBER |
ClickHouse | ClickHouse | 524,598,875 | 6,575 | alexey-milovidov | Yes, it becomes almost obvious... | 2019-08-25T04:09:58 | 2019-08-25T04:09:58 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.