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 | 619,191,298 | 10,488 | surekhasaharan | tried without cat, still same error
```
clickhouse-client --query "INSERT INTO table1 FORMAT ORC" < sample_orc
Code: 173. DB::ErrnoException: Allocator: Cannot mmap 128.00 MiB., errno: 12, strerror: Cannot allocate memory
```
Am I missing some config ? | 2020-04-24T19:06:24 | 2020-04-24T19:06:24 | {} | NONE |
ClickHouse | ClickHouse | 619,406,758 | 10,493 | den-crane | https://clickhouse.tech/docs/en/operations/settings/settings/#settings-optimize_skip_unused_shards
**optimize_skip_unused_shards**
_Enables or disables skipping of unused shards for SELECT queries that have sharding key condition in PREWHERE/WHERE (assumes that the data is distributed by sharding key, otherwise d... | 2020-04-25T16:39:52 | 2020-04-25T16:40:02 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,452,335 | 10,493 | byx313 | > https://clickhouse.tech/docs/en/operations/settings/settings/#settings-optimize_skip_unused_shards
>
> **optimize_skip_unused_shards**
>
> _Enables or disables skipping of unused shards for SELECT queries that have sharding key condition in PREWHERE/WHERE (assumes that the data is distributed by sharding key, o... | 2020-04-25T23:05:54 | 2020-04-25T23:05:54 | {} | NONE |
ClickHouse | ClickHouse | 618,359,361 | 2,278 | alexey-milovidov | @winter7 Default value (autodetect) should be ok for all use cases. | 2020-04-23T11:56:15 | 2020-04-23T11:56:15 | {} | MEMBER |
ClickHouse | ClickHouse | 618,591,018 | 10,453 | alexey-milovidov | If odbc-bridge transfers data in Native format and does not parse date and time, we don't care about timezones. | 2020-04-23T18:54:31 | 2020-04-23T18:54:31 | {} | MEMBER |
ClickHouse | ClickHouse | 618,769,277 | 2,278 | winter7 | @alexey-milovidov I don't think I can find default value in `config.xml` `<merge_tree>` nor `system.settings`. Is it related with version? | 2020-04-24T02:37:13 | 2020-04-24T02:37:13 | {} | NONE |
ClickHouse | ClickHouse | 618,774,415 | 10,436 | gglinux | DB::ExpressionBlockInputStream::readImpl, Get this, from core dump. @filimonov
| 2020-04-24T02:59:13 | 2020-04-24T06:04:15 | {} | NONE |
ClickHouse | ClickHouse | 618,888,857 | 10,055 | filimonov | It seems like installing ca-certificates, and ajusting path don't help.
If the solution from the ticket above don't help - it may be some bug related to openssl version used by clickhouse - it's uses unreleased prealpha openssl 3 since 20.1:
* https://github.com/ClickHouse/ClickHouse/pull/8956#issuecomment-5812551... | 2020-04-24T08:50:42 | 2020-04-24T11:33:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,031,949 | 9,707 | maqroll | Looks great.
What if `"before.Value.field1.string" String DEFAULT "hello"`?
Would take into account the default value instead of the default default value by type?
| 2020-04-24T14:06:53 | 2020-04-24T14:06:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,915,240 | 5,491 | zlen131313 | > @CurtizJ has found a way to remove parts in case if whole partition is going to be removed.
What is this way? | 2020-04-24T09:48:38 | 2020-04-24T18:56:52 | {} | NONE |
ClickHouse | ClickHouse | 619,463,948 | 10,503 | alexey-milovidov | @oandrew Yes, creating temporary std::vector/std::string/whatever in a loop is a common performance antipattern. | 2020-04-26T01:23:01 | 2020-04-26T01:23:01 | {} | MEMBER |
ClickHouse | ClickHouse | 618,460,206 | 10,459 | alesapin | Actually it's not 100% equal behavior as before, but I think it will work well. | 2020-04-23T15:21:43 | 2020-04-23T15:21:43 | {} | MEMBER |
ClickHouse | ClickHouse | 618,585,268 | 10,450 | den-crane | >&query=INSERT into default.is_bu_log values (1111,'2020040812','2020-04-08
Seems that you send your data in URL. URL is obviously limited and you don't use POST.
Check https://clickhouse.tech/docs/en/interfaces/http/
Insert and values in POST body
```
echo 'INSERT INTO t VALUES (1),(2),(3)' | curl 'http://... | 2020-04-23T18:45:56 | 2020-04-23T18:45:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,981,762 | 10,481 | 4ertus2 | Workaround `set multiple_joins_rewriter_version = 2` | 2020-04-24T12:33:36 | 2020-04-24T12:33:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,018,340 | 10,482 | adrian17 | I was able to reproduce on two separate machines, but indeed looks like a temporary issue. | 2020-04-24T13:42:34 | 2020-04-24T13:42:34 | {} | NONE |
ClickHouse | ClickHouse | 619,064,774 | 10,479 | qoega | You have to update reference
```
2020-04-24 15:48:24 01231_markdown_format: [ FAIL ] 0.12 sec. - result differs with reference:
2020-04-24 15:48:24 --- /usr/share/clickhouse-test/queries/0_stateless/01231_markdown_format.reference 2020-04-24 14:30:51.000000000 +0300
... | 2020-04-24T15:01:15 | 2020-04-24T15:01:15 | {} | MEMBER |
ClickHouse | ClickHouse | 619,407,872 | 10,435 | amosbird | Ok, I'll also do some tests about the extra time ThinLTO entails. | 2020-04-25T16:45:43 | 2020-04-25T16:45:43 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 619,445,807 | 10,488 | den-crane | Try to reduce max_insert_block_size
clickhouse-client `--max_insert_block_size=100000` | 2020-04-25T21:54:07 | 2020-04-25T21:54:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,962,703 | 9,545 | And42 | Well, one performance test is failing anyway :(
For now the only solution I see if that's critical is to add an additional switch to preferences to enable/disable perf events | 2020-04-24T11:45:41 | 2020-04-24T11:45:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,968,979 | 1,981 | depyronick | I was deleted the error log file manually with user ubuntu and created it myself, so the permission got changed to ubuntu. `/var/log/clickhouse` and the files inside it should be chowned to `clickhouse:clickhouse`. That fixed the error. | 2020-04-24T12:02:32 | 2020-04-24T12:02:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,442,801 | 10,357 | fibersel | Can i take this task if @YiuRULE has not started it yet? | 2020-04-25T21:25:32 | 2020-04-25T21:25:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,311,511 | 2,278 | winter7 | do you have any recommanded number about `max_part_loading_threads` in `merge_tree`? @filimonov | 2020-04-23T10:08:52 | 2020-04-23T10:08:52 | {} | NONE |
ClickHouse | ClickHouse | 618,348,986 | 9,545 | And42 | @alexey-milovidov @akuzm do queries reuse threads from a pool or create new ones when executed? If they reuse it's possible to call `syscall` only when a thread is created and not each time a query starts | 2020-04-23T11:31:03 | 2020-04-23T11:31:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,508,757 | 10,435 | amosbird | btw, we should also disable LTO for unitests | 2020-04-23T16:43:32 | 2020-04-23T16:43:32 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 618,646,618 | 10,458 | qoega | ==273==ERROR: AddressSanitizer: allocator is out of memory trying to allocate 0x10000f bytes | 2020-04-23T20:23:47 | 2020-04-23T20:23:47 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 619,013,116 | 10,482 | adrian17 | Seems like it was fixed already? I'll keep the issue open for confirmation. | 2020-04-24T13:33:43 | 2020-04-24T13:33:43 | {} | NONE |
ClickHouse | ClickHouse | 619,016,583 | 10,482 | blinkov | @adrian17 I think we weren't doing anything to fix it, probably you have just had some temporary issue. Can't reproduce on fresh Ubunut 18 at the moment. | 2020-04-24T13:39:32 | 2020-04-24T13:39:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,164,736 | 8,915 | SaltTan | The behaviour is the same in 20.3
DB::Exception: Column value is not under aggregate function and not in GROUP BY. (version 19.16.17.1 (official build)) (from [::ffff]:49392) (in query: SELECT alias AS value FROM system.settings ALL INNER JOIN (SELECT value AS alias, name FROM system.settings **HAVING** alias = 'ena... | 2020-04-24T18:03:26 | 2020-04-24T18:03:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,270,931 | 10,484 | alexey-milovidov | > tests/config/readonly.xml → tests/ctest/config/readonly.xml
These configs are used in Docker images for tests. | 2020-04-24T22:52:33 | 2020-04-24T22:52:33 | {} | MEMBER |
ClickHouse | ClickHouse | 619,454,015 | 10,493 | byx313 | > https://clickhouse.tech/docs/en/operations/settings/settings/#settings-optimize_skip_unused_shards
>
> **optimize_skip_unused_shards**
>
> _Enables or disables skipping of unused shards for SELECT queries that have sharding key condition in PREWHERE/WHERE (assumes that the data is distributed by sharding key, o... | 2020-04-25T23:24:06 | 2020-04-25T23:24:06 | {} | NONE |
ClickHouse | ClickHouse | 619,454,542 | 8,936 | byx313 | > Hi @byx313, I can't reproduce the issue that you are seeing. Here is my test case.
>
> 1. I enable live views and create a live view table.
>
> ```
> ClickHouse client version 20.2.1.1.
> Connecting to localhost:9000 as user default.
> Connected to ClickHouse server version 20.2.1 revision 54432.
>
> :) S... | 2020-04-25T23:30:34 | 2020-04-25T23:30:34 | {} | NONE |
ClickHouse | ClickHouse | 618,399,193 | 5,985 | alesapin | Now it works when all values in column is not equal to `NULL`. | 2020-04-23T13:35:19 | 2020-04-23T13:35:19 | {
"+1": 3
} | MEMBER |
ClickHouse | ClickHouse | 618,531,458 | 10,462 | den-crane | No. You need to repeat your query and it will be executed on a live replica. | 2020-04-23T17:22:25 | 2020-04-23T17:22:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,538,041 | 10,461 | den-crane | SET max_threads=1, max_insert_block_size=65536, max_insert_block_size=65536;
How many columns in your table?
>Any way to force this flush to disk to free some memory between files?
You don't need it. CH (jemalloc) does not deallocate memory for better performance. This behavior should not cause any issues, if ... | 2020-04-23T17:34:01 | 2020-04-23T17:36:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,706,232 | 10,464 | den-crane | http handlers -- not released yet https://github.com/ClickHouse/ClickHouse/pull/7774
master documentation -- https://clickhouse.tech/docs/en/interfaces/http/#predefined_http_interface
20.3 documentation https://clickhouse.tech/docs/v20.3/en/interfaces/http/ | 2020-04-23T22:34:54 | 2020-04-23T22:34:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,898,643 | 10,055 | filimonov | ```
openssl version
OpenSSL 1.1.1 11 Sep 2018
openssl s_client -state -nbio -connect data.gharchive.org:443
CONNECTED(00000005)
Turned on non blocking io
SSL_connect:before SSL initialization
SSL_connect:SSLv3/TLS write client hello
SSL_connect:error in SSLv3/TLS write client hello
write R BLOCK
SSL_conn... | 2020-04-24T09:11:54 | 2020-04-24T09:18:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,930,959 | 10,477 | alesapin | I expect 7 to 10 failed tests. | 2020-04-24T10:24:56 | 2020-04-24T10:24:56 | {} | MEMBER |
ClickHouse | ClickHouse | 619,038,379 | 9,724 | vitlibar | > @vitlibar What is the status here?
Fixed. | 2020-04-24T14:18:06 | 2020-04-24T14:18:06 | {} | MEMBER |
ClickHouse | ClickHouse | 618,593,158 | 9,545 | And42 | In the last commit I added a logic to create descriptors only once per thread. This added an issue: there can be only one active instance of `PerfEventsCounters` per thread at a given moment. If there is more than a one instance the resulting values are incorrect. I will fix it by not stopping event measurements after ... | 2020-04-23T18:57:41 | 2020-04-23T18:57:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,922,110 | 8,639 | CurtizJ | New failures:
`test_insert_quorum_with_move_partition` - `ALTER TABLE ... MOVE PARTITION TO TABLE` is not implemented in 20.1
`01254_array_of_unnamed_tuples` - function `generateRandom` is not implemented in 20.1
More `test_ttl_move` tests failed in comparison to previous release, but all these tests disabled as f... | 2020-04-24T10:04:16 | 2020-04-24T10:11:37 | {} | MEMBER |
ClickHouse | ClickHouse | 619,448,549 | 8,915 | SaltTan | @zhang2014 This is what happens when I send my query to CH 19.16
```
$ clickhouse-client
ClickHouse client version 20.3.7.46 (official build).
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 20.3.7 revision 54433.
SELECT alias AS value
FROM remote('ch-19-16', system.setting... | 2020-04-25T22:24:10 | 2020-04-25T22:24:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,305,774 | 10,398 | pic-Nick | >
>
> #10200 ?
No. I do not face such an issue. Only illegal query crashes. And it remains on 20.3.7.46.
>
>
> It's strange that you have JSONRowOutputFormat in stack trace, cause your query has no output format specified. Does the crash appear with default output format?
I ran this query from TABI... | 2020-04-23T09:57:09 | 2020-04-23T09:57:09 | {} | NONE |
ClickHouse | ClickHouse | 618,316,523 | 8,814 | winter7 | @filimonov thanks! You guys release so fast | 2020-04-23T10:18:29 | 2020-04-23T10:18:29 | {} | NONE |
ClickHouse | ClickHouse | 618,439,206 | 5,666 | nikitamikhaylov | I have no idea how to do it properly, but let's try. #10458 | 2020-04-23T14:48:59 | 2020-04-23T14:48:59 | {} | MEMBER |
ClickHouse | ClickHouse | 618,465,004 | 10,453 | alesapin | Unfortunately `--push/pop-state` supports only:
```
The option which are covered are: -Bdynamic, -Bstatic, -dn, -dy, -call_shared,
-non_shared, -static, -N, -n, --whole-archive, --no-whole-archive, -r, -Ur,
--copy-dt-needed-entries, --no-copy-dt-needed-entries, --as-needed, --no-as-needed,
... | 2020-04-23T15:28:53 | 2020-04-23T15:29:28 | {} | MEMBER |
ClickHouse | ClickHouse | 618,587,701 | 10,443 | den-crane | I would use `sed`
```
cat test.txt | sed 's/:/|/' | clickhouse-client --format_csv_delimiter = '|'
``` | 2020-04-23T18:49:32 | 2020-04-23T18:49:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,548,420 | 10,455 | den-crane | Why do you use Buffer over Distributed ? But not over MergeTree?
Could not reproduce:
```
CREATE TABLE some_table_local on cluster my_cluster (uuid String)
ENGINE = MergeTree PRIMARY KEY uuid ORDER BY uuid SETTINGS index_granularity = 8192
CREATE TABLE some_table_all on cluster my_cluster
(uuid St... | 2020-04-23T17:51:54 | 2020-04-23T19:19:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,755,240 | 10,455 | hydra-yl | @den-crane Thanks for the reply.
Actually, the reproduce steps are not identical to our production usage: a java application consume message from kafka and do a bit of processing then insert into clickhouse. The main purpose to use buffer table is to accelerate insertion speed. The goal is to insert 5billion rows i... | 2020-04-24T01:42:11 | 2020-04-24T01:42:40 | {} | NONE |
ClickHouse | ClickHouse | 619,057,871 | 10,243 | drook | Could you please clarify this: you mean that I need to re-report build issue for a fresh release of 20.3.x, if any, or, since 20.4 is building fine, you mean exactly it - so I should stick to 20.4.x ?
In the latter case - Yandex Managed Cloud CH offers only LTS versions, and 20.4 is a stable - so, since my employer ... | 2020-04-24T14:49:55 | 2020-04-24T14:50:44 | {} | NONE |
ClickHouse | ClickHouse | 619,223,144 | 10,483 | ramazanpolat | @fhmcn A workaround would be creating a named pipe at `/var/log/clickhouse/clickhouse-server.log`.
Then you reject lines you don't want(in your case rejecting lines that contain `default.SampleDict` will work) and write the filtered data to a real file.
Something like this:
```bash
$ rm -rf /var/log/clickhouse/... | 2020-04-24T20:25:13 | 2020-04-24T20:50:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,462,453 | 9,707 | oandrew | @maqroll
yep, it will use `DEFAULT` if specified
| 2020-04-26T01:04:27 | 2020-04-26T01:04:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,965,706 | 10,405 | CurtizJ | I've downloaded your data parts and tried to run different versions. But your query worked as expected. Also I tried the same versions, but with address sanitizer and had the same result. Maybe it's some kind of hardware issue? Can you reproduce the bug on other machine?
```
SELECT version()
┌─version()─┐
│ 20.... | 2020-04-24T11:54:04 | 2020-04-24T11:54:04 | {} | MEMBER |
ClickHouse | ClickHouse | 619,443,094 | 10,498 | vzakaznikov | @KochetovNicolai, I've added a test `0_stateless/01247_insert_into_watch_live_view_url.py ` but it does not pass. There seems to be an issue with the `url` table function. The data is being buffered somewhere by the StorageURL's output stream and does not get written out until the stream is finalized. Any ideas how it ... | 2020-04-25T21:28:20 | 2020-04-25T21:28:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,456,719 | 10,493 | den-crane | > > https://clickhouse.tech/docs/en/operations/settings/settings/#settings-optimize_skip_unused_shards
> > **optimize_skip_unused_shards**
> > _Enables or disables skipping of unused shards for SELECT queries that have sharding key condition in PREWHERE/WHERE (assumes that the data is distributed by sharding key, oth... | 2020-04-25T23:57:17 | 2020-04-25T23:57:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,489,313 | 10,488 | surekhasaharan | I had tried below
```
clickhouse-client --max_insert_block_size=1024 --query "INSERT INTO table1 FORMAT ORC" < sample_orc
Received exception from server (version 20.3.8):
Code: 241. DB::Exception: Received from localhost:9000. DB::Exception: Memory limit (for query) exceeded: would use 9.33 GiB (attempt to alloc... | 2020-04-26T06:06:44 | 2020-04-26T06:06:44 | {} | NONE |
ClickHouse | ClickHouse | 618,474,463 | 10,460 | vzakaznikov | @KochetovNicolai, any ideas what changed between 19.* and 20.* to cause this? | 2020-04-23T15:43:54 | 2020-04-23T15:43:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,635,896 | 9,849 | mvcalder-xbk | Fixed with the upgrade, thank you. | 2020-04-23T20:05:57 | 2020-04-23T20:05:57 | {} | NONE |
ClickHouse | ClickHouse | 619,028,058 | 9,724 | filimonov | @vitlibar What is the status here? | 2020-04-24T14:00:06 | 2020-04-24T14:00:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,087,042 | 10,455 | hydra-yl | @den-crane After change to distriubted -> Buffer -> MergeTree structure, problem is solved: No missing rows or/and duplicate rows. Maybe Clickhouse should not allow buffer over distributed table | 2020-04-24T15:39:03 | 2020-04-24T15:39:03 | {} | NONE |
ClickHouse | ClickHouse | 619,269,591 | 10,243 | alexey-milovidov | Successful build is only guaranteed for build variants that are tested in CI.
We accept patches for alternative builds but we apply them directly on master, without backporting.
If you build release version manually, you can also re-apply these patches manually. | 2020-04-24T22:46:54 | 2020-04-24T22:46:54 | {} | MEMBER |
ClickHouse | ClickHouse | 619,340,389 | 10,493 | weiqxu | it will send the query to all shard.
if the `channel` is the PARTATION KEY or PK column. other shards that didn't contain the value will return empty result very soon. | 2020-04-25T08:12:56 | 2020-04-25T08:12:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,463,070 | 10,503 | oandrew | @alexey-milovidov is it even possible it affects performance that much? | 2020-04-26T01:12:04 | 2020-04-26T01:12:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,499,753 | 10,475 | AlexanderJLiu | I think you should better use [clickhouse-jdbc](https://github.com/ClickHouse/clickhouse-jdbc).
These test [examples](https://github.com/ClickHouse/clickhouse-jdbc/tree/master/src/test/java/ru/yandex/clickhouse) may be help. | 2020-04-26T07:16:45 | 2020-04-26T07:16:45 | {} | NONE |
ClickHouse | ClickHouse | 618,438,781 | 10,454 | rsurgiewicz | Thanks, that's an unfortunate storing big dict as a mv would be a performance impact.
Is it possible to force/schedule mv update manually anyhow? | 2020-04-23T14:48:15 | 2020-04-23T14:48:15 | {} | NONE |
ClickHouse | ClickHouse | 618,444,850 | 10,397 | maison2710 | this code in `DatabaseOrdinary.cpp` should have some problem
```
void tryAttachTable(
Context & context,
const ASTCreateQuery & query,
DatabaseOrdinary & database,
const String & database_name,
const String & metadata_path,
bool has_force_restore_data_flag)
... | 2020-04-23T14:57:46 | 2020-04-23T17:14:26 | {} | NONE |
ClickHouse | ClickHouse | 618,589,961 | 10,435 | alexey-milovidov | Thank you!
Maybe you can create a new pull request and we will close current one? | 2020-04-23T18:53:01 | 2020-04-23T18:53:01 | {} | MEMBER |
ClickHouse | ClickHouse | 619,482,051 | 10,503 | alexey-milovidov | Yes, I faced similar issues multiple times...
It's a shame that we did not notice it earlier. | 2020-04-26T04:56:41 | 2020-04-26T04:56:48 | {} | MEMBER |
ClickHouse | ClickHouse | 618,462,466 | 9,624 | CurtizJ | Failed 01135_default_and_alter_zookeeper - ok. Format of `SHOW CREATE TABLE` was a bit changed in #10049.
Failed 01116_cross_count_asterisks - ok. Setting `multiple_joins_rewriter_version` was inroduced later than 20.3 | 2020-04-23T15:25:08 | 2020-04-23T15:25:40 | {} | MEMBER |
ClickHouse | ClickHouse | 618,536,185 | 10,326 | den-crane | 1. Yes, MV impacts on system performance and `Insert` works longer, it finishes when all MVs inserted their data.
2. No atomicity. No consistency.
MV can break insert into source table in the middle.
MV can insert part of the rows inserted into the source table.
see explanation https://clickhouse.tech/docs/en/op... | 2020-04-23T17:30:52 | 2020-04-23T17:31:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,536,924 | 8,513 | nauu | @alex-krash Thanks. | 2020-04-23T17:32:07 | 2020-04-23T17:32:07 | {} | MEMBER |
ClickHouse | ClickHouse | 618,797,953 | 10,243 | drook | Thanks to @olevole we now have a working port for 20.4, but I still cannot build 20.3 (as far as I know - @olevole cannot too), @alexey-milovidov could you please take a look at this ?
````crash
FAILED: programs/server/CMakeFiles/clickhouse-server-lib.dir/TCPHandler.cpp.o
/usr/local/bin/clang++90 -DBOOST_SYSTEM_N... | 2020-04-24T04:41:36 | 2020-04-24T04:42:18 | {} | NONE |
ClickHouse | ClickHouse | 618,884,727 | 10,459 | alesapin | test failures are irrelevant to changes. | 2020-04-24T08:42:00 | 2020-04-24T08:42:00 | {} | MEMBER |
ClickHouse | ClickHouse | 618,921,215 | 6,721 | wk198803 | > clichouse source code version: v19.13.3.26-stable
> ubantu for arm version: 18.04
> gcc versioin:8.3.0
> ninja version: 1.8.2-1
> cmake version:3.11.2
>
> build failed，Details are as follows：
>
> during RTL pass: expand
> In file included from ../dbms/src/Functions/bitRotateRight.cpp:2:
> ../dbms/src/Func... | 2020-04-24T10:02:14 | 2020-04-24T10:02:14 | {} | NONE |
ClickHouse | ClickHouse | 618,958,786 | 10,479 | filimonov | thanks :) | 2020-04-24T11:35:10 | 2020-04-24T11:35:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,801,885 | 10,470 | YiuRULE | With more investigation, it seems we have a limit, who is `2^63 - offset_size_clause` | 2020-04-24T04:58:10 | 2020-04-24T12:11:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,271,075 | 10,484 | alexey-milovidov | FYI the tests for external models are not automated. It's time to convert them to integration tests. | 2020-04-24T22:53:06 | 2020-04-24T22:53:06 | {} | MEMBER |
ClickHouse | ClickHouse | 618,299,812 | 9,702 | 4ertus2 | Partial MergeJoin use max_bytes_in_join/max_rows_in_join (and if they are not pressent - default_max_bytes_in_join) as limit when it should start join-on-disk. It's the same as `max_bytes_before_external_group_by` and `max_bytes_before_external_sort` for `GROUP BY` and `ORDER BY`.
If you set them bigger then allowed m... | 2020-04-23T09:45:05 | 2020-04-23T09:48:27 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,555,759 | 10,461 | brenoarosa | Thanks for the suggestion, I'm trying it right now.
The table has around 40 columns, which 2 are partition columns and other 2 are sorting columns.
There are around 900M rows in total. | 2020-04-23T18:02:47 | 2020-04-23T18:02:47 | {} | NONE |
ClickHouse | ClickHouse | 618,615,135 | 10,454 | den-crane | > Is it possible to force/schedule mv update manually anyhow?
No, it's not possible.
Try Live View https://www.altinity.com/blog/tag/Live+Views | 2020-04-23T19:31:22 | 2020-04-23T19:31:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,781,823 | 2,278 | den-crane | @winter7 The default value max_part_loading_threads = auto (=== cpu cores)
https://github.com/ClickHouse/ClickHouse/issues/10087#issuecomment-610966025 | 2020-04-24T03:29:14 | 2020-04-24T03:29:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,885,333 | 10,450 | zhouxiujue | Yes, I sent my data in URL. Maybe it is the reason.
But in java, I can't send the insert data in body, do you know why?
I hava wrote another issue at #10475. | 2020-04-24T08:43:08 | 2020-04-24T08:43:08 | {} | NONE |
ClickHouse | ClickHouse | 619,006,598 | 10,405 | hombit | Thank you, @CurtizJ. I tried on other machine, you are right, everything works fine. Can it be connected with the specific CPU version? Head of `/proc/cpuinfo`:
```
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 85
model name : Intel(R) Xeon(R) Gold 5118 CPU @ 2.30GHz
stepping : 4
microcode : 0x... | 2020-04-24T13:23:08 | 2020-04-24T13:25:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,404,620 | 10,454 | qoega | 1) I guess not.
2) You can try to do something like source table join with this dictionary. And store last updated version in this source table.
``` sql
INSERT INTO local_table
key, dictGet('dict_name', 'value', key) as value, dictGet('dict_name', 'version', key) as version
(
SELECT key, max(version)
FROM... | 2020-04-23T13:49:29 | 2020-04-23T13:49:29 | {} | MEMBER |
ClickHouse | ClickHouse | 619,090,703 | 10,455 | den-crane | The root cause still unclear. It needs a reproduction. | 2020-04-24T15:45:29 | 2020-04-24T15:45:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,271,540 | 9,545 | alexey-milovidov | > Well, one performance test is failing anyway :(
That's well known flacky test, it fails sporadically. So we can say that performance tests are passed! | 2020-04-24T22:54:56 | 2020-04-24T22:54:56 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 619,443,027 | 10,357 | alexey-milovidov | @fibersel Yes, it's yours! | 2020-04-25T21:27:52 | 2020-04-25T21:27:52 | {} | MEMBER |
ClickHouse | ClickHouse | 618,352,943 | 9,545 | alexey-milovidov | You can use thread local variable as a flag to check if perf-events are initialized. | 2020-04-23T11:41:12 | 2020-04-23T11:41:12 | {} | MEMBER |
ClickHouse | ClickHouse | 618,480,669 | 10,460 | vzakaznikov | In the materialized views we have the following https://github.com/ClickHouse/ClickHouse/blob/82a71e95986c46bae157749515af7b5c88f8ba97/src/Storages/StorageMaterializedView.h#L31. Maybe we need to add the same?
| 2020-04-23T15:53:55 | 2020-04-23T15:55:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,891,766 | 10,468 | filimonov | Added ca-certificates to yandex/clickhouse-server image.
In the meanwhile you can do smth like
```
cat 'apt-get update && apt-get install ca-certificates' > update_ca.sh
chmod +x update_ca.sh
docker run -v $(pwd)/update_ca.sh:/docker-entrypoint-initdb.d/update_ca.sh yandex/clickhouse-server
``` | 2020-04-24T08:57:09 | 2020-04-24T08:57:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,352,033 | 10,493 | byx313 | > it will send the query to all shard.
>
> if the `channel` is the PARTATION KEY or PK column. other shards that didn't contain the value will return empty result very soon.
What if `channel` is not the partition key or pk?Obviously,send the query to the direct shard will speed up the query time,i can not figure ... | 2020-04-25T09:42:51 | 2020-04-25T09:42:51 | {} | NONE |
ClickHouse | ClickHouse | 619,475,351 | 10,495 | xiashuijun | clickhouse-server --daemon --config-file=/etc/clickhouse-server/config.xml,兄弟你把deamon放到前面试试
| 2020-04-26T03:31:54 | 2020-04-26T03:31:54 | {} | NONE |
ClickHouse | ClickHouse | 619,483,586 | 10,024 | alexey-milovidov | We cannot use it if we don't understand the name of the variable.
I can guess that D is probably "desktop" and X is probably "extensible" but I don't know what is G. | 2020-04-26T05:14:47 | 2020-04-26T05:14:47 | {} | MEMBER |
ClickHouse | ClickHouse | 618,358,637 | 5,666 | alexey-milovidov | @nikitamikhaylov Cool! Let's add a test and close. | 2020-04-23T11:54:42 | 2020-04-23T11:54:42 | {} | MEMBER |
ClickHouse | ClickHouse | 618,853,737 | 10,243 | alexey-milovidov | We won't backport build fixes (for build variants that are not covered by CI) to previous releases. | 2020-04-24T07:32:29 | 2020-04-24T07:32:29 | {} | MEMBER |
ClickHouse | ClickHouse | 618,869,098 | 10,468 | filimonov | related #10055 | 2020-04-24T08:07:59 | 2020-04-24T11:31:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 618,987,622 | 10,481 | alex-zaitsev | Probably fixed by #9739 in 20.4. Need to confirm | 2020-04-24T12:47:22 | 2020-04-24T12:47:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 619,177,971 | 10,477 | alesapin | Actually the whole idea is wrong. virtual columns is a constant property of Storage class, not Storage object. | 2020-04-24T18:34:46 | 2020-04-24T18:34:46 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 619,369,338 | 8,915 | zhang2014 | Do you have any settings on? I can't reproduce it in default settings with master branch.
BTW: Perhaps you could execute the following SQL and copy the result here ?
``` SQL
SET enable_debug_queries= 1;
ANALYZE SELECT alias AS value FROM system.settings ALL INNER JOIN (SELECT value AS alias, name FROM system.... | 2020-04-25T12:05:39 | 2020-04-25T12:05:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,186,964 | 10,638 | alexey-milovidov | Simply disable this test in Sandbox. | 2020-05-03T21:42:28 | 2020-05-03T21:42:28 | {} | MEMBER |
ClickHouse | ClickHouse | 623,224,908 | 10,434 | alexey-milovidov | It did not help. | 2020-05-04T01:51:56 | 2020-05-04T01:51:56 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.