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 | 616,266,671 | 6,732 | alexey-milovidov | Close as cannot reproduce. | 2020-04-20T01:49:27 | 2020-04-20T01:49:27 | {} | MEMBER |
ClickHouse | ClickHouse | 616,269,747 | 4,086 | alexey-milovidov | It is possible that memory cannot be allocated when there are too many virtual mappings:
Example:
```
$ wc -l /proc/$(pidof clickhouse-server)/maps
65518 /proc/22342/maps
```
The default limit is 65536. | 2020-04-20T02:03:51 | 2020-04-20T02:03:51 | {} | MEMBER |
ClickHouse | ClickHouse | 616,284,409 | 10,326 | den-crane | ReplicatedAggregatingMergeTree -- needs AggregateFunctions. It does not calculate sums automatically as summingMT does.
1. The easiest and the most efficient way in case is to use SummingMergeTree
```
CREATE MATERIALIZED VIEW materialized.t_minute_scene_monitor_local_replica on cluster ch_cluster (
hour String... | 2020-04-20T03:11:05 | 2020-04-20T03:15:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,387,279 | 2,219 | alexey-milovidov | I've added your test case to the test suite. | 2020-04-20T08:16:10 | 2020-04-20T08:16:10 | {} | MEMBER |
ClickHouse | ClickHouse | 616,389,852 | 2,380 | alexey-milovidov | Does not reproduce on newer versions. | 2020-04-20T08:21:05 | 2020-04-20T08:21:05 | {} | MEMBER |
ClickHouse | ClickHouse | 616,393,561 | 2,718 | alexey-milovidov | Aerospike: #5629
Cassandra: #4978 | 2020-04-20T08:27:53 | 2020-04-20T08:27:53 | {} | MEMBER |
ClickHouse | ClickHouse | 613,031,919 | 10,176 | alexey-milovidov | Please note that this will actually introduce security vulnerability instead of fixing anything. | 2020-04-13T18:34:54 | 2020-04-20T08:32:02 | {} | MEMBER |
ClickHouse | ClickHouse | 615,890,434 | 9,893 | alexey-milovidov | Let's postpone it. | 2020-04-18T15:32:10 | 2020-04-18T15:32:10 | {} | MEMBER |
ClickHouse | ClickHouse | 615,984,178 | 10,124 | alexey-milovidov | The time is divided by N-hour intervals from unspecified hour. | 2020-04-18T23:57:04 | 2020-04-18T23:57:04 | {} | MEMBER |
ClickHouse | ClickHouse | 616,108,539 | 10,243 | olevole | @drook replxx exists in the ports tree, we need to add CMAKE_REQUIRED_INCLUDES:
diff -ruN /usr/ports/databases/clickhouse/Makefile /root/clickhouse/Makefile
```
--- /usr/ports/databases/clickhouse/Makefile 2020-04-04 23:29:22.163375000 +0300
+++ /root/clickhouse/Makefile 2020-04-19 14:07:49.897556000 +0300
... | 2020-04-19T11:23:23 | 2020-04-19T11:23:23 | {} | NONE |
ClickHouse | ClickHouse | 616,213,141 | 9,642 | alexey-milovidov | Threre is a very small and simple patch from Alexander Burmak, I will apply it in nearest... hours. | 2020-04-19T19:39:03 | 2020-04-19T19:39:03 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 616,383,058 | 2,069 | alexey-milovidov | Так как добавлена возможность создавать словари запросами CREATE DICTIONARY, теперь не требуется отдельно создавать таблицы с ENGINE = Dictionary, и актуальность задачи отсутствует. | 2020-04-20T08:07:56 | 2020-04-20T08:08:27 | {} | MEMBER |
ClickHouse | ClickHouse | 616,388,837 | 2,341 | alexey-milovidov | @KochetovNicolai ^ | 2020-04-20T08:19:14 | 2020-04-20T08:19:14 | {} | MEMBER |
ClickHouse | ClickHouse | 616,129,698 | 5,618 | blinkov | @jangorecki only partially, I think, there's always a lot of room for improvement | 2020-04-19T13:04:38 | 2020-04-19T13:04:38 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,235,066 | 10,341 | azat | >Maybe we can solve the task by introducing another QueryProcessingStage: WithMergeableStateAfterAggregation (it means that aggregate functions were calculated and finalized) and process it inside InterpreterSelectQuery?
Decided to address this separately (to avoid polluting this patch set, since I guess there will ... | 2020-04-19T22:23:34 | 2020-04-19T22:23:34 | {} | MEMBER |
ClickHouse | ClickHouse | 616,261,511 | 8,995 | alexey-milovidov | @filimonov Yes, it will show memory leak (assuming the process is correctly terminated and assuming that increasing memory usage is attributed to memory leak). Unbounded grow of memory usage can also be attributed to the grow of size of some data structures that are not actually leaked. | 2020-04-20T01:24:23 | 2020-04-20T01:28:20 | {} | MEMBER |
ClickHouse | ClickHouse | 616,274,695 | 10,326 | BernieJiangDL | > Materialized view is an insert trigger it catches data from inserts.
> Did you create your MV at all CH nodes?
>
> It would be nice to see DDL of MV. Very often users create MV table with wrong `partition by` / `order by` -- inconsistent with MV select.
Thanks for your reply.
1. Yes, I created MV at all my C... | 2020-04-20T02:26:44 | 2020-04-20T02:26:44 | {} | NONE |
ClickHouse | ClickHouse | 616,279,997 | 5,341 | alexey-milovidov | This part of code (loading of external models) was rewritten since then. | 2020-04-20T02:49:54 | 2020-04-20T02:49:54 | {} | MEMBER |
ClickHouse | ClickHouse | 616,280,252 | 2,642 | alexey-milovidov | > Do the following MergeTree settings work for you?
These settings should work for this task. | 2020-04-20T02:50:59 | 2020-04-20T02:51:06 | {} | MEMBER |
ClickHouse | ClickHouse | 616,293,164 | 10,326 | BernieJiangDL | > ReplicatedAggregatingMergeTree -- needs AggregateFunctions. It does not calculate sums automatically as summingMT does.
>
> 1. The easiest and the most efficient way in case is to use SummingMergeTree
>
> ```
> CREATE MATERIALIZED VIEW materialized.t_minute_scene_monitor_local_replica on cluster ch_cluster (
... | 2020-04-20T03:50:39 | 2020-04-20T03:50:39 | {} | NONE |
ClickHouse | ClickHouse | 616,307,397 | 10,368 | alexey-milovidov | Maybe simply check all parts participating in merge? | 2020-04-20T04:51:41 | 2020-04-20T04:51:41 | {} | MEMBER |
ClickHouse | ClickHouse | 616,363,362 | 10,327 | crakjie | @filimonov maybe it's a bad understanding of the documentation but [here](https://clickhouse.tech/docs/en/sql_reference/functions/conditional_functions/#if) it's wrote that :
> The function executes then and else expressions and returns its result, depending on whether the condition cond ended up being zero or not. ... | 2020-04-20T07:28:08 | 2020-04-20T07:28:08 | {} | NONE |
ClickHouse | ClickHouse | 616,382,038 | 1,966 | alexey-milovidov | > Да, совсем несложно добавить возможность учёта столбцов в заголовке.
Добавлено, настройка `input_format_with_names_use_header`. | 2020-04-20T08:06:01 | 2020-04-20T08:06:01 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 616,391,190 | 2,455 | alexey-milovidov | @filimonov This feature was implemented by @nikitamikhaylov in #8303 | 2020-04-20T08:23:35 | 2020-04-20T08:23:35 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 615,891,143 | 10,348 | alexey-milovidov | It will fix automatically on next commit where README.md will be unchanged. | 2020-04-18T15:36:56 | 2020-04-18T15:36:56 | {} | MEMBER |
ClickHouse | ClickHouse | 615,983,710 | 10,124 | alexey-milovidov | It looks like just unspecified behaviour. | 2020-04-18T23:54:36 | 2020-04-18T23:54:36 | {} | MEMBER |
ClickHouse | ClickHouse | 616,131,442 | 10,355 | st3rv04ka | Does this help me add a new column without changing the data file? I think this solution will only work if I have more columns in the file than in the table. Can you give an example, please? | 2020-04-19T13:13:07 | 2020-04-19T13:13:07 | {} | NONE |
ClickHouse | ClickHouse | 616,213,461 | 8,962 | alexey-milovidov | @killwort We have intention to make most of the typical use cases compatible.
We'll make CREATE syntax compatible: by adding support for complex type expressions like CHAR VARYING NOT NULL... | 2020-04-19T19:41:19 | 2020-04-19T19:41:19 | {} | MEMBER |
ClickHouse | ClickHouse | 616,283,621 | 7,885 | alexey-milovidov | It's most likely fixed in #7762 | 2020-04-20T03:07:16 | 2020-04-20T03:07:16 | {} | MEMBER |
ClickHouse | ClickHouse | 616,293,567 | 10,326 | BernieJiangDL | > ReplicatedAggregatingMergeTree -- needs AggregateFunctions. It does not calculate sums automatically as summingMT does.
>
> 1. The easiest and the most efficient way in case is to use SummingMergeTree
>
> ```
> CREATE MATERIALIZED VIEW materialized.t_minute_scene_monitor_local_replica on cluster ch_cluster (
... | 2020-04-20T03:52:50 | 2020-04-20T03:52:50 | {} | NONE |
ClickHouse | ClickHouse | 615,970,008 | 10,355 | den-crane | https://clickhouse.tech/docs/en/sql_reference/table_functions/input/ | 2020-04-18T22:52:42 | 2020-04-18T22:52:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,169,885 | 10,355 | den-crane | ```
CREATE TABLE db.database (
product String,
product_num String,
date Date
) ENGINE = ReplacingMergeTree() PARTITION BY toMonday(date)
ORDER BY (product, product_num) SETTINGS index_granularity = 8192;
echo aaaa,bbbb|clickhouse-client --query="INSERT INTO db.database SELECT a, b, '2020-01-01' FROM inp... | 2020-04-19T16:00:42 | 2020-04-19T16:06:50 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,201,409 | 9,642 | azat | >Will be reimplemented in a different way.
By the way, will it be done in the nearest feature?
I'm kind of applied it to my tree, and that said that I'm using it (with some fixes). | 2020-04-19T18:25:31 | 2020-04-19T18:25:31 | {} | MEMBER |
ClickHouse | ClickHouse | 616,261,559 | 8,995 | alexey-milovidov | @akuzm just forgot about this task. | 2020-04-20T01:24:42 | 2020-04-20T01:24:42 | {} | MEMBER |
ClickHouse | ClickHouse | 616,450,003 | 9,707 | maqroll | @oandrew
Fails with
```
2020.04.20 10:02:16.486936 [ 53 ] {} <Error> void DB::StorageKafka::threadFunc(): Code: 8, e.displayText() = DB::Exception: Field source.name not found in Avro schema, Stack trace (when copying this message, always include the lines below):
0. Poco::Exception::Exception(std::__1::basic_... | 2020-04-20T10:12:52 | 2020-04-20T10:27:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,514,268 | 346 | 4ertus2 | > @4ertus2 I tried and it seems `emptyArrayToSingle` is not impacted by `join_use_nulls`. What is the expected behavior ?
Let's make an issue. It could be something unknown or related to #9253 | 2020-04-20T12:15:27 | 2020-04-20T12:15:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,261,904 | 10,172 | alexey-milovidov | Not fixed: https://clickhouse-test-reports.s3.yandex.net/10359/237db2668874014e960bdacc4b7917cdf93869e5/functional_stateless_tests_(ubsan).html | 2020-04-20T01:26:10 | 2020-04-20T01:26:10 | {} | MEMBER |
ClickHouse | ClickHouse | 616,337,962 | 10,360 | alexey-milovidov | Documentation is wrong.
CC @BayoNet | 2020-04-20T06:26:52 | 2020-04-20T06:26:52 | {} | MEMBER |
ClickHouse | ClickHouse | 616,384,750 | 2,219 | alexey-milovidov | Does not reproduce in recent versions. | 2020-04-20T08:11:13 | 2020-04-20T08:11:13 | {} | MEMBER |
ClickHouse | ClickHouse | 616,392,379 | 346 | crakjie | @4ertus2 I tried and it seems `emptyArrayToSingle` is not impacted by `join_use_nulls`. What is the expected behavior ? | 2020-04-20T08:25:46 | 2020-04-20T08:25:46 | {} | NONE |
ClickHouse | ClickHouse | 616,507,420 | 2,341 | KochetovNicolai | Воспроизвел на 1.1.54380 во так:
```
CREATE TABLE tr_checks (date_check Date) ENGINE = MergeTree(date_check, (date_check), 8192);
insert into tr_checks values (addMonths(toStartOfMonth(today()), -3) - 1);
insert into tr_checks values (addMonths(toStartOfMonth(today()), -3) + 1);
ALTER TABLE tr_checks add colum... | 2020-04-20T12:01:25 | 2020-04-20T12:01:25 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 616,589,070 | 8,995 | akuzm | > @akuzm just forgot about this task.
Thankfully we have github so we don't have to remember. The reason I'm not doing anything about this task is that it doesn't seem a particularly high priority to me, and there is always a lot of other things to do. | 2020-04-20T14:26:24 | 2020-04-20T14:26:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 615,896,285 | 9,642 | alexey-milovidov | Will be reimplemented in a different way. | 2020-04-18T16:10:30 | 2020-04-18T16:10:30 | {} | MEMBER |
ClickHouse | ClickHouse | 615,902,582 | 5,346 | AlessandroSechi | Same here for 19.17.6. Upgrading to 20.3.7 revision 54433 solved the issue. | 2020-04-18T16:51:57 | 2020-04-18T16:51:57 | {} | NONE |
ClickHouse | ClickHouse | 616,353,461 | 10,320 | maison2710 | Thank you @alexey-milovidov
I was able to release deb packages by using docker image `yandex/clickhouse-deb-builder` :
1. `docker run --network=host --rm -it --volume=/data/test/output:/output --volume=/data/test/build:/build --volume=/data/test/local_ccache:/ccache -e DEB_CC=gcc-9 -e DEB_CXX=g++-9 -e CCACHE_DIR... | 2020-04-20T07:05:39 | 2020-04-20T07:06:26 | {} | NONE |
ClickHouse | ClickHouse | 616,379,844 | 9,241 | azat | Yeah, the problem is that after #8750:
- the aggregator starts checking available space on disk
- but it does not takes compression into account
- and if there will be enough threads it will fail, since the allocation will be done in parallel and all the reservation will be active
So what we can do is just remove... | 2020-04-20T08:01:48 | 2020-04-20T08:01:48 | {} | MEMBER |
ClickHouse | ClickHouse | 616,448,910 | 10,243 | drook | @olevole Thanks a lot for your patch, I took the part about replxx from it, but I still cannot build 20.3.7.48 neitehr from the light GH0 version, nor for from the full-recursive (can you ?).
This thing still stumbles for me at the cityhash (though replxx is working now):
````
# make
===> License APACHE20 acce... | 2020-04-20T10:10:54 | 2020-04-20T10:10:54 | {} | NONE |
ClickHouse | ClickHouse | 616,546,879 | 2,718 | filimonov | Redis is delivered, but "Introduce direct dictionary layout" - not yet (w/o that ClickHouse still copy everything to own memory). | 2020-04-20T13:17:59 | 2020-04-20T13:17:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,571,255 | 10,368 | den-crane | >will try to fetch it instead -> No active replica has part
a) Introduce some counter `fetch_try_no_active` and give up after 10 (100 ( configurable)) tries?
b) Without automation: `alter table drop part xxx force` / `optimize table partition yyyy force` | 2020-04-20T13:57:15 | 2020-04-20T13:59:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,003,153 | 10,352 | zplzpl | all options , not working
1. SETTINGS optimize_read_in_order = 0
```
SELECT
*,
neighbor(sold_cnt, -1) AS prev_sold_cnt
FROM
(
SELECT
snapshot_time,
product_id,
sold_cnt
FROM fact_product_sn
WHERE product_id = 2102982256
ORDER BY snapshot_time ASC
... | 2020-04-19T01:37:56 | 2020-04-19T01:45:26 | {} | NONE |
ClickHouse | ClickHouse | 616,206,893 | 9,559 | alexey-milovidov | It was broken with Processors:
```
Alexey Milovidov, [19.04.20 21:44]
@kochetovnicolai У нас регрессия с процессорами. Пока не покрыта тестами, я собираюсь исправить:
milovidov@milovidov-desktop:~/work/ClickHouse$ curl http://localhost:8123/?experimental_use_processors=0 --data-binary "SELECT DISTINCT number % 5 ... | 2020-04-19T18:54:37 | 2020-04-19T18:54:37 | {} | MEMBER |
ClickHouse | ClickHouse | 616,353,262 | 7,202 | alexey-milovidov | It happens for time zones that have offset changed between positive and negative. | 2020-04-20T07:05:08 | 2020-04-20T07:05:08 | {} | MEMBER |
ClickHouse | ClickHouse | 616,192,172 | 10,263 | azat | Actually not sure that "pr-improvement" will be enough, since it can be pretty tricky to debug the problems with distributed sends i.e. why it became slower (for "regular" user), maybe backward incompatible is better? | 2020-04-19T17:39:04 | 2020-04-19T17:39:04 | {} | MEMBER |
ClickHouse | ClickHouse | 616,249,861 | 10,286 | alexey-milovidov | > Sync of merged data part on removal of any of source data parts.
This is in question because fsync of a file that was written and closed is not guaranteed to work properly on Linux: https://stackoverflow.com/questions/37288453/calling-fsync2-after-close2 | 2020-04-20T00:16:16 | 2020-04-20T00:16:16 | {} | MEMBER |
ClickHouse | ClickHouse | 616,269,327 | 4,662 | alexey-milovidov | Does it reproduce on recent versions? | 2020-04-20T02:01:51 | 2020-04-20T02:01:51 | {} | MEMBER |
ClickHouse | ClickHouse | 616,392,561 | 2,562 | alexey-milovidov | This is normal and does not indicate any errors itself. | 2020-04-20T08:26:07 | 2020-04-20T08:26:07 | {} | MEMBER |
ClickHouse | ClickHouse | 616,392,784 | 2,564 | alexey-milovidov | @azat FYI | 2020-04-20T08:26:31 | 2020-04-20T08:26:31 | {} | MEMBER |
ClickHouse | ClickHouse | 616,423,159 | 5,591 | BayoNet | Ok. I'll do it ASAP. | 2020-04-20T09:22:06 | 2020-04-20T09:22:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,636,213 | 10,000 | akuzm | One thing we can improve is to dispatch on whether the argument is tuple statically, not dynamically -- e.g. add two implementaitons of `sumMap`, one takes normal columns and the other takes Tuple, and then create the right one at initialization stage, the same way we create different specializations for different type... | 2020-04-20T15:42:25 | 2020-04-20T15:42:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 615,889,182 | 10,343 | alexey-milovidov | It's very beautiful. Like a set of kids' pencils... | 2020-04-18T15:23:58 | 2020-04-18T15:23:58 | {
"laugh": 1
} | MEMBER |
ClickHouse | ClickHouse | 615,904,630 | 4,003 | inakivb | Hi,
I've just have a problem that may be related (or be the same)
Querying HTTP API whit session enabled, using the same session_id, from time to time it start to make all queries fail with the error below.
It seems that, removing the session makes everything work fine.
```
(version 20.3.5.21 (official build))... | 2020-04-18T17:05:01 | 2020-04-18T17:05:01 | {} | NONE |
ClickHouse | ClickHouse | 616,167,868 | 10,358 | filimonov | Add smth like that to your config.xml:
```xml
<kafka>
<security_protocol>sasl_plaintext</security_protocol>
<sasl_mechanism>PLAIN</sasl_mechanism>
<sasl_username>test</sasl_username>
<sasl_password>test</sasl_password>
<!--
<debug>all</debug>
<auto_offset_reset>latest</auto_offset_reset>
<compression_type>sna... | 2020-04-19T15:52:19 | 2020-04-19T16:07:15 | {
"hooray": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,224,904 | 9,707 | oandrew | @maqroll
Could you test if my PR covers your use case?
You can download a build from https://clickhouse-builds.s3.yandex.net/10354/68f2e934439766f235361abd4339164c8b8dd7d0/report.html | 2020-04-19T21:04:16 | 2020-04-19T21:04:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,268,387 | 7,970 | alexey-milovidov | We don't support gcc-8 anymore. Only gcc-9. | 2020-04-20T01:57:18 | 2020-04-20T01:57:18 | {} | MEMBER |
ClickHouse | ClickHouse | 616,274,339 | 515 | alexey-milovidov | For `dictGetHierarchy` it was fixed in #1064 | 2020-04-20T02:25:07 | 2020-04-20T02:25:07 | {} | MEMBER |
ClickHouse | ClickHouse | 616,281,507 | 5,141 | alexey-milovidov | Fixed. | 2020-04-20T02:56:50 | 2020-04-20T02:56:50 | {} | MEMBER |
ClickHouse | ClickHouse | 616,256,692 | 9,241 | alexey-milovidov | @azat could you please look at this issue? | 2020-04-20T00:59:24 | 2020-04-20T00:59:24 | {} | MEMBER |
ClickHouse | ClickHouse | 616,453,960 | 10,243 | drook | [CMakeOutput.log](https://github.com/ClickHouse/ClickHouse/files/4502745/CMakeOutput.log)
| 2020-04-20T10:20:35 | 2020-04-20T10:20:35 | {} | NONE |
ClickHouse | ClickHouse | 615,888,511 | 10,301 | alexey-milovidov | TSV without header does not support BOM because it is not possible to distinguish it with binary data in the value in first row first column. | 2020-04-18T15:19:44 | 2020-04-18T15:19:44 | {} | MEMBER |
ClickHouse | ClickHouse | 616,391,998 | 2,373 | alexey-milovidov | Fixed some time ago. | 2020-04-20T08:25:07 | 2020-04-20T08:25:07 | {} | MEMBER |
ClickHouse | ClickHouse | 615,890,810 | 10,348 | alexey-milovidov | Performance test is broken because it tries to analyze README.md file. | 2020-04-18T15:34:45 | 2020-04-18T15:34:45 | {} | MEMBER |
ClickHouse | ClickHouse | 615,908,071 | 10,352 | nauu | Maybe it was caused by #8545. | 2020-04-18T17:27:44 | 2020-04-18T17:27:44 | {} | MEMBER |
ClickHouse | ClickHouse | 616,207,903 | 9,559 | alexey-milovidov | Help was provided! | 2020-04-19T19:01:57 | 2020-04-19T19:01:57 | {} | MEMBER |
ClickHouse | ClickHouse | 616,269,775 | 4,086 | alexey-milovidov | I'll close this as obsolete. | 2020-04-20T02:04:00 | 2020-04-20T02:04:00 | {} | MEMBER |
ClickHouse | ClickHouse | 616,281,040 | 8,655 | alexey-milovidov | That's reasonable, let's do that. @4ertus2 ? | 2020-04-20T02:54:46 | 2020-04-20T02:54:56 | {} | MEMBER |
ClickHouse | ClickHouse | 616,661,102 | 10,362 | alexey-milovidov | My collegue @velom from Yandex.Metrica requested (he is too shy to post a comment on GitHub) that `max_server_memory_usage` should be server setting in config.xml instead of user-level setting. Will do. | 2020-04-20T16:21:51 | 2020-04-20T16:21:51 | {} | MEMBER |
ClickHouse | ClickHouse | 615,890,368 | 9,893 | alexey-milovidov | @qoega It will change the behaviour of `empty` function:
```
SELECT empty(toFixedString('', 1 + randConstant() % 100));
``` | 2020-04-18T15:31:43 | 2020-04-18T15:31:43 | {} | MEMBER |
ClickHouse | ClickHouse | 615,984,265 | 10,124 | alexey-milovidov | If you are comparing with some other implementation - please add the reference. | 2020-04-18T23:57:34 | 2020-04-18T23:57:34 | {} | MEMBER |
ClickHouse | ClickHouse | 616,208,408 | 8,962 | killwort | Why bother with MySQL/PSQL type-compatibility if other SQL parts are not compatible with anything else? E.g. CREATE syntax, bulk INSERT syntax, SELECT/JOIN syntax. | 2020-04-19T19:05:36 | 2020-04-19T19:05:36 | {} | NONE |
ClickHouse | ClickHouse | 616,235,704 | 10,341 | alexey-milovidov | Ok. | 2020-04-19T22:27:46 | 2020-04-19T22:27:46 | {} | MEMBER |
ClickHouse | ClickHouse | 616,238,249 | 10,243 | olevole | @drook with GH_TUPLES we can use the CH installation via git and not violate the FreeBSD ports concept, e.g:
https://github.com/olevole/fbsd-ports/tree/master/databases/clickhouse-devel
Tested with FreeBSD 13-CURRENT ( https://pastebin.com/7G10sqgQ )
I am also interested in CH on FreeBSD platform ( besides Li... | 2020-04-19T22:47:03 | 2020-04-19T22:47:21 | {} | NONE |
ClickHouse | ClickHouse | 616,265,703 | 6,269 | alexey-milovidov | All the unneeded logic from `checkDataPart` was removed. | 2020-04-20T01:44:43 | 2020-04-20T01:44:43 | {
"hooray": 3
} | MEMBER |
ClickHouse | ClickHouse | 616,279,114 | 9,854 | alexey-milovidov | @filimonov MOVEs are not production ready until the tests will be enabled: #8840 | 2020-04-20T02:45:58 | 2020-04-20T02:45:58 | {} | MEMBER |
ClickHouse | ClickHouse | 616,301,667 | 5,257 | alexey-milovidov | @alesapin How did you get this trace?
Do you think that the issue is still relevant? | 2020-04-20T04:30:59 | 2020-04-20T04:30:59 | {} | MEMBER |
ClickHouse | ClickHouse | 616,310,015 | 10,368 | alexey-milovidov | How to fix it manually:
1. `SELECT * FROM system.replication_queue WHERE create_time < now() - INTERVAL 1 DAY AND type = 'MERGE_PARTS' AND last_exception LIKE '%No active replica has part%' \G`
2. `SELECT replica_path || '/queue/' || node_name FROM system.replication_queue JOIN system.replicas USING (database, tabl... | 2020-04-20T05:01:03 | 2020-04-20T05:24:59 | {
"+1": 28
} | MEMBER |
ClickHouse | ClickHouse | 616,643,572 | 10,000 | akuzm | By the way, at the first glance the sumMap implementation we have now looks kinda suboptimal.
@mkabilov sent me a flamegraph from a server that runs sumMap a lot, made with our `system.trace_log`. This wonderful site won't let me attach an svg, so here it is: https://sendeyo.com/up/d/e285888839
You can see lots of ... | 2020-04-20T15:53:54 | 2020-04-20T16:00:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 615,914,256 | 10,341 | alexey-milovidov | Ok. The code becomes complicated and bug-prone but still acceptable.
Maybe we can solve the task by introducing another QueryProcessingStage: WithMergeableStateAfterAggregation (it means that aggregate functions were calculated and finalized) and process it inside InterpreterSelectQuery?
(we can experiment in a s... | 2020-04-18T18:01:11 | 2020-04-18T18:01:43 | {} | MEMBER |
ClickHouse | ClickHouse | 615,915,920 | 10,341 | azat | > Maybe we can solve the task by introducing another QueryProcessingStage: WithMergeableStateAfterAggregation (it means that aggregate functions were calculated and finalized) and process it inside InterpreterSelectQuery?
Yeah, I was just "experimenting"
Another stage is a great idea, will take a look (even distinc... | 2020-04-18T18:06:57 | 2020-04-18T18:06:57 | {} | MEMBER |
ClickHouse | ClickHouse | 615,931,229 | 10,352 | den-crane | Because the subquery returned 2 blocks
try
SELECT
snapshot_time,
product_id,
sold_cnt
FROM fact_product
WHERE product_id = 2102982256
ORDER BY snapshot_time ASC **setttings optimize_read_in_order=0**
or
**ORDER BY snapshot_time ASC, product_id+0**
or
**O... | 2020-04-18T19:22:00 | 2020-04-18T19:57:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,189,020 | 10,357 | alexey-milovidov | Maybe @YiuRULE is interested in this task? | 2020-04-19T17:25:06 | 2020-04-19T17:25:06 | {} | MEMBER |
ClickHouse | ClickHouse | 616,192,808 | 10,263 | alexey-milovidov | It's only relevant when using a huge number of Distributed tables (rare case). And it should not become slower as we have 16 background threads. Data is sent almost as is without any processing on our side, so it will either saturate the network or there are slow peers (that will require additional debugging). | 2020-04-19T17:41:55 | 2020-04-19T17:41:55 | {} | MEMBER |
ClickHouse | ClickHouse | 616,244,606 | 10,352 | den-crane | WA:
```
insert into fact_product values('2020-02-11',2102982256,179819), ('2020-02-12',2102982256,180663 )
SELECT
*,
neighbor(sold_cnt, -1) AS prev_sold_cnt
FROM
(
SELECT *
FROM
(
SELECT
snapshot_time,
product_id,
sold_cnt
FROM fac... | 2020-04-19T23:37:53 | 2020-04-19T23:37:53 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 616,266,209 | 10,364 | alexey-milovidov | This is minor bugfix, thus no-backport.
There are no tests for this bugfix, because it's technically difficult to provide. | 2020-04-20T01:47:03 | 2020-04-20T01:47:03 | {} | MEMBER |
ClickHouse | ClickHouse | 616,236,541 | 10,362 | azat | (And just a note it includes #10361) | 2020-04-19T22:33:50 | 2020-04-19T22:33:50 | {} | MEMBER |
ClickHouse | ClickHouse | 616,266,621 | 6,732 | alexey-milovidov | It is possible that memory cannot be allocated when there are too many virtual mappings:
Example:
```
$ wc -l /proc/$(pidof clickhouse-server)/maps
65518 /proc/22342/maps
```
The default limit is 65536. | 2020-04-20T01:49:10 | 2020-04-20T01:49:10 | {} | MEMBER |
ClickHouse | ClickHouse | 616,280,670 | 8,973 | alexey-milovidov | Similar to https://github.com/ClickHouse/ClickHouse/issues/7540 | 2020-04-20T02:53:02 | 2020-04-20T02:53:02 | {} | MEMBER |
ClickHouse | ClickHouse | 616,281,312 | 6,833 | alexey-milovidov | Fixed in master. | 2020-04-20T02:55:53 | 2020-04-20T02:55:53 | {} | MEMBER |
ClickHouse | ClickHouse | 616,282,200 | 5,025 | alexey-milovidov | We don't have `HaMergeTree` in our codebase. | 2020-04-20T03:00:09 | 2020-04-20T03:00:09 | {} | MEMBER |
ClickHouse | ClickHouse | 616,795,976 | 10,379 | alexey-milovidov | Test fail added to CI fest. | 2020-04-20T20:39:06 | 2020-04-20T20:39:06 | {} | MEMBER |
ClickHouse | ClickHouse | 616,859,616 | 10,336 | den-crane | I configured my system CH+nginx exactly as your settings
```
cat /etc/clickhouse-server/conf.d/user_default.xml
<?xml version="1.0" ?>
<yandex>
<users>
<default>
<networks replace="replace">
<ip>::1</ip>
<ip>127.0.0.1</ip>
</networks>
... | 2020-04-20T23:16:26 | 2020-04-20T23:18:16 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.