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 | 817,199,591 | 22,297 | abyss7 | > The only remaining is Arcadia.
Yes, I'm fixing it right now. | 2021-04-10T20:41:28 | 2021-04-10T20:41:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,265,437 | 7,751 | nvartolomei | I'm closing this as the initial work was done as part of #16193. | 2021-04-11T07:48:56 | 2021-04-11T07:48:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,284,966 | 21,639 | traceon | Performance numbers are published here: https://clickhouse.tech/benchmark/hardware/ (search for "MacBook Air M1".) | 2021-04-11T10:31:44 | 2021-04-11T10:31:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,513,177 | 22,963 | filimonov | > There is too much infrastructure in Debian tools.
👍
Not only debian tools, also lot of things in clickhouse repo spread to many shell / python / docker files are needed to build rpm, tar etc
> To build Debian package we only need to place some files accordingly, run tar, gzip and ar.
👎
We just need ... | 2021-04-12T06:09:16 | 2021-04-12T06:22:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,730,249 | 22,980 | alexey-milovidov | Also try to lower the value of `max_threads` setting. | 2021-04-12T11:25:11 | 2021-04-12T11:25:11 | {} | MEMBER |
ClickHouse | ClickHouse | 818,012,303 | 22,996 | vzakaznikov | @loyd, could you please try specifying `JSONEachRowWithProgress` as the output format? This will force the connection use heartbeats to keep HTTP connection alive. When client abruptly disappears it will force a fault on the next hearbeat and the output stream should close. So the max expected time for the table to be ... | 2021-04-12T18:00:30 | 2021-04-12T18:00:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 818,248,096 | 2,719 | alexey-milovidov | The issue most likely fixed here: #19516 | 2021-04-12T21:24:58 | 2021-04-12T21:24:58 | {} | MEMBER |
ClickHouse | ClickHouse | 817,257,103 | 21,770 | taylor12805 | Hi, filimonov, I have fixed variable not initialized bug.
I tried to add array nullable support in the latest commit, but it ran into "Memory Address not mapped to object" error, which i had no idea why it happens. Could you please take a look at it? Thanks a lot
Test case:
CREATE TABLE test.event_table(buvid Str... | 2021-04-11T06:22:57 | 2021-04-11T06:25:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,362,174 | 22,290 | alexey-milovidov | https://clickhouse-test-reports.s3.yandex.net/22729/680d722f22232e5922477a440140902421b20be2/fuzzer_debug/report.html#fail1 | 2021-04-11T19:47:17 | 2021-04-11T19:47:17 | {} | MEMBER |
ClickHouse | ClickHouse | 818,010,755 | 22,987 | alexey-milovidov | Maybe it's related to old version of `gcc`. I've fixed the warning in https://github.com/ClickHouse/ClickHouse/commit/fafe8c00fe53559eb28d9a1392e8b45c3600e30b. | 2021-04-12T17:58:07 | 2021-04-12T17:58:17 | {} | MEMBER |
ClickHouse | ClickHouse | 818,202,974 | 23,002 | alexey-milovidov | ```
CREATE DICTIONARY dict_flat (`key` UInt64 DEFAULT 0, `a` UInt8 DEFAULT 42, `b` String DEFAULT 'x', `c` Float64 DEFAULT 42.) PRIMARY KEY key SOURCE(CLICKHOUSE(HOST 'localhost' PORT tcpPort() USER 'default' TABLE 't1' PASSWORD '' DB 'db_01115')) LIFETIME(MIN 1 MAX 10) LAYOUT(FLAT())
```
```
SELECT * FROM (SELEC... | 2021-04-12T20:40:13 | 2021-04-12T20:40:13 | {} | MEMBER |
ClickHouse | ClickHouse | 817,192,285 | 22,960 | alexey-milovidov | I'm ok with removal of `01802_dict_sparse_hashed_bytes_allocated`. | 2021-04-10T19:36:44 | 2021-04-10T19:36:44 | {} | MEMBER |
ClickHouse | ClickHouse | 817,558,986 | 13,889 | melin | @den-crane Can be optimized?
| 2021-04-12T07:26:27 | 2021-04-12T07:26:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,574,258 | 22,985 | UnamedRus | There is setting for that:
```
set cast_keep_nullable=1;
``` | 2021-04-12T07:48:49 | 2021-04-12T07:48:49 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,693,836 | 22,673 | KochetovNicolai | ```
zless clickhouse-server.log.gz | grep -a Fatal
2021.04.08 01:47:41.131832 [ 295 ] {} <Fatal> Application: Child process was terminated by signal 9 (KILL). If it is not done by 'forcestop' command or manually, the possible cause is OOM Killer (see 'dmesg' and look at the '/var/log/kern.log' for the details).
``` | 2021-04-12T10:29:02 | 2021-04-12T10:29:02 | {} | MEMBER |
ClickHouse | ClickHouse | 817,840,090 | 22,679 | taiwhite | @ianton-ru I am still receiving the same error from [#22864](https://github.com/ClickHouse/ClickHouse/pull/22864).
The tables I am using are the same as before. Note: This error occurs when there are two S3 buckets (the bucket urls are different).
Query: `SELECT d FROM zero_copy_replication ORDER BY d DESC LIMIT... | 2021-04-12T14:04:31 | 2021-04-12T14:06:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,941,404 | 22,503 | CurtizJ | In general everything is ok. | 2021-04-12T16:14:48 | 2021-04-12T16:14:48 | {} | MEMBER |
ClickHouse | ClickHouse | 817,698,616 | 21,511 | alexey-milovidov | We have one client that will be very happy to enable this option, but they also have very long lists (millions of items). | 2021-04-12T10:35:57 | 2021-04-12T10:35:57 | {} | MEMBER |
ClickHouse | ClickHouse | 817,725,322 | 22,987 | alexey-milovidov | If you compile only `clickhouse` target:
```
ninja clickhouse
```
does it succeed?
Does it help to update to master? | 2021-04-12T11:16:42 | 2021-04-12T11:16:42 | {} | MEMBER |
ClickHouse | ClickHouse | 818,014,301 | 22,974 | CurtizJ | Data parts of `MergeTree`-family tables can be stored on disk in 2 formats: Wide (read optimized) and Compact (write optimized). In Wide format data of each column is stored in one or several files. In Compact format all columns are written to one file: granules are written one by one, while every granule has columnar ... | 2021-04-12T18:03:20 | 2021-04-12T18:03:20 | {
"+1": 3
} | MEMBER |
ClickHouse | ClickHouse | 818,021,961 | 18,918 | alexey-milovidov | @CurtizJ has fixed all the issues. | 2021-04-12T18:14:09 | 2021-04-12T18:14:09 | {} | MEMBER |
ClickHouse | ClickHouse | 818,069,367 | 22,314 | alexey-milovidov | Another example:
`SELECT count() FROM system.numbers WHERE NOT ignore(intDiv(number, 1000000000))`
- 960.49 million rows/s.
`SELECT count() FROM system.numbers WHERE NOT ignore(divide(number, 1000000000))`
- 1.20 billion rows/s.
`SELECT count() FROM system.numbers WHERE NOT ignore(toUInt32(divide(number, 100... | 2021-04-12T19:19:23 | 2021-04-12T19:21:58 | {} | MEMBER |
ClickHouse | ClickHouse | 818,116,160 | 22,980 | gituser | @alexey-milovidov
how do I obfuscate the dataset? | 2021-04-12T19:58:48 | 2021-04-12T19:58:48 | {} | NONE |
ClickHouse | ClickHouse | 817,274,781 | 15,765 | abyss7 | I moved this PR to the draft state because of `[WIP]` prefix in description - if it's ready for review then, please, move it back with a proper title. | 2021-04-11T09:06:18 | 2021-04-11T09:06:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,500,469 | 22,647 | azat | >[Integration tests (release) — fail: 0, passed: 1293, flaky: 1](https://clickhouse-test-reports.s3.yandex.net/22647/fad42f8ac5d8f23c2215a84dff8e61be0dc40495/integration_tests_(release).html)
Now it passed
>[Integration tests (asan) — Timeout, fail: 0, passed: 1290, flaky: 0](https://clickhouse-test-reports.s3.ya... | 2021-04-12T05:44:39 | 2021-04-12T06:51:37 | {} | MEMBER |
ClickHouse | ClickHouse | 817,619,351 | 22,760 | cfroystad | @abyss7 Good point! I should of course have done that from the start. I've added an integration test for the initial replication. Note: due to some system trouble on my part, I've been unable to test the integration test programmatically, but it's representative of the issue which caused ClickHouse to crash/stop replic... | 2021-04-12T08:49:49 | 2021-04-12T08:49:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 818,026,395 | 22,996 | loyd | Yep, with `JSONEachRowWithProgress` it works as expected.
I write a library based on HTTP + `RowBinary`. I can switch to `JSONEachRowWithProgress` for `WATCH`, but it's not very convenient. Maybe I should `shutdown` the connection? | 2021-04-12T18:20:39 | 2021-04-12T18:20:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 818,120,275 | 22,980 | gituser | > Also try to lower the value of `max_threads` setting.
did you mean `<max_thread_pool_size>10000</max_thread_pool_size>` this setting, because I don't see `max_threads` in config.xml. | 2021-04-12T20:00:48 | 2021-04-12T20:00:48 | {} | NONE |
ClickHouse | ClickHouse | 818,129,322 | 21,693 | alexey-milovidov | Version 21.4 has been released. | 2021-04-12T20:04:53 | 2021-04-12T20:04:53 | {} | MEMBER |
ClickHouse | ClickHouse | 818,284,480 | 19,522 | alexey-milovidov | `export CLICKHOUSE_WATCHDOG_ENABLE=0` | 2021-04-12T22:28:40 | 2021-04-12T22:28:40 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 817,265,353 | 22,960 | azat | >@azat I removed test 01802_dict_sparse_hashed_bytes_allocated.sql because it does check that we use right method to calculate size of sparse hash table, but check overall size of hashed dictionary, and it will be inconvenient to update it every time internal implementation will change. What do you think ?
Sure, no ... | 2021-04-11T07:48:04 | 2021-04-11T07:48:04 | {} | MEMBER |
ClickHouse | ClickHouse | 817,305,596 | 21,639 | changvvb | The performance on MacBook Air M1 seems even better than MacBook Pro 2020. | 2021-04-11T13:10:00 | 2021-04-11T13:10:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,613,402 | 22,762 | vdimir | > Can't see in this PR where and how do you use new creator variant, can you show please?
@abyss7
See commit [`f022f07` (#22762)](https://github.com/ClickHouse/ClickHouse/pull/22762/commits/f022f0744b1f2a6bb80da0c116a9124306efb827). But I'm not intended to make `windowFunnel` experimental (now flag true on defau... | 2021-04-12T08:41:52 | 2021-04-12T08:41:52 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 817,663,098 | 22,973 | kreuzerkrieg | > Do you mean that we should not perform this retry?
>
> https://github.com/ClickHouse-Extras/aws-sdk-cpp/blob/7d48b2c8193679cc4516e5bd68ae4a64b94dae7d/aws-cpp-sdk-core/source/client/AWSClient.cpp#L967
yes, of course we should not retry on non-retryable errors, like 404 | 2021-04-12T09:47:56 | 2021-04-12T09:47:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,732,797 | 22,980 | gituser | @alexey-milovidov, I don't think this setting has anything to do with the issue, as the issue is reproduce-able only on versions after `v20.6.11.1` and in both `v20.6.11.1` and `v20.7.2.30` this setting has same value:
* `v20.6.11.1`:
```# grep -r mark_cache_size config.xml
<mark_cache_size>5368709120</mark_c... | 2021-04-12T11:29:06 | 2021-04-12T11:36:07 | {} | NONE |
ClickHouse | ClickHouse | 817,750,236 | 22,987 | filimonov | Looks like a duplicate of https://github.com/ClickHouse/ClickHouse/issues/22862 | 2021-04-12T11:59:12 | 2021-04-12T11:59:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 818,263,655 | 2,719 | alexey-milovidov | I will add a test. | 2021-04-12T21:45:27 | 2021-04-12T21:45:27 | {} | MEMBER |
ClickHouse | ClickHouse | 817,295,291 | 22,647 | azat | >Need to merge with master.
Rebased against upstream/master
>The failure of test_zookeeper_config is suspicious.
Right, multiple clusters had been broken (with correct `docker-compose down`), fixed in https://github.com/ClickHouse/ClickHouse/pull/22647/commits/fad42f8ac5d8f23c2215a84dff8e61be0dc40495 | 2021-04-11T11:56:31 | 2021-04-11T11:56:31 | {} | MEMBER |
ClickHouse | ClickHouse | 817,527,786 | 19,522 | onine007 | How to disable watchdog ? | 2021-04-12T06:36:27 | 2021-04-12T06:36:27 | {} | NONE |
ClickHouse | ClickHouse | 817,599,619 | 20,031 | bobliugq | I got same error | 2021-04-12T08:22:49 | 2021-04-12T08:22:49 | {} | NONE |
ClickHouse | ClickHouse | 817,622,685 | 4,006 | hei- | Env:
clickhouse: 21.1.2.15
mysql: 5.7.9
Execute SQL:
CREATE DATABASE xxx ENGINE = MaterializeMySQL('x.x.x.x:3306', 'db_name', 'xxx', 'xxx.')
Error:
2021.04.12 02:13:05.623932 [ 46 ] {c4ad0e3d-f7d3-47dd-8f38-7dbaa9524ca4} <Error> TCPHandler: Code: 1000, e.displayText() = DB::Exception: m... | 2021-04-12T08:54:17 | 2021-04-12T08:54:17 | {} | NONE |
ClickHouse | ClickHouse | 817,750,575 | 22,162 | filimonov | No information. Most probably a duplicate of #22862 | 2021-04-12T11:59:44 | 2021-04-12T11:59:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 818,009,136 | 22,996 | vzakaznikov | This most likely means that the HTTP connection is not closed on the server and the live output stream never gets destroyed. | 2021-04-12T17:55:41 | 2021-04-12T17:55:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 815,096,499 | 22,490 | filimonov | I've read it. I've answered that
If you feel my answer not satisfy you - please introduce more details / context.
Asking me to reread the question is not constructive. If you think it can help with something - please reread my answer above. 10 times :p | 2021-04-07T17:35:29 | 2021-04-12T18:08:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 818,081,431 | 22,945 | alesapin | Strange things happen:
```
2021.04.12 13:21:50.746597 [ 91081 ] {} <Fatal> BaseDaemon: ########################################
2021.04.12 13:21:50.746926 [ 91081 ] {} <Fatal> BaseDaemon: (version 21.5.1.6535, build id: 4D18A4888711EFBB24A82666D5A7FBB65D2EAC44) (from thread 300) (no query) Received signal Segmentati... | 2021-04-12T19:40:09 | 2021-04-12T19:40:09 | {} | MEMBER |
ClickHouse | ClickHouse | 818,081,682 | 22,945 | alesapin | No idea, failures not related to changes. | 2021-04-12T19:40:35 | 2021-04-12T19:40:35 | {} | MEMBER |
ClickHouse | ClickHouse | 818,227,233 | 22,996 | vzakaznikov | > But emergencies are still possible, so the use of `JSONEachRowWithProgress` seems to be a more reliable method.
Yes, you should only use `JSONEachRowWithProgress` format for `WATCH` queries over the HTTP connection. First, because HTTP connection needs heartbeat messages send periodically from the server to the cl... | 2021-04-12T20:52:00 | 2021-04-12T20:52:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,285,395 | 22,647 | alexey-milovidov | Need to merge with master.
The failure of `test_zookeeper_config` is suspicious. | 2021-04-11T10:35:20 | 2021-04-11T10:35:20 | {} | MEMBER |
ClickHouse | ClickHouse | 817,342,293 | 7,229 | Tasselmi | > @Tasselmi it sound like you are looking for that:
> #1616
> #2741
> #3483
>
> BTW. @BayoNet looks like `from_env` is not documented.
hi，please see my latest comment in Otc. 2019. I think we should extract macros from `remote_servers` configuration. How do u think? thus our configure file will be all the same... | 2021-04-11T17:25:58 | 2021-04-11T17:30:41 | {} | NONE |
ClickHouse | ClickHouse | 817,438,179 | 22,980 | den-crane | Change `<mlock_executable>false</mlock_executable>` in /etc/clickhouse-server/config.xml | 2021-04-12T02:34:10 | 2021-04-12T02:34:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,439,803 | 22,273 | xiaogaozi | > > @vzakaznikov May I ask, do you have any clue about this issue?
>
> That version is "quite old". Do you have ability to try the same action on a newer release as mentioned above?
@vzakaznikov Thanks for your reply, we may try to verify this behavior with newer CH in the future. BTW, as a workaround, we have us... | 2021-04-12T02:40:06 | 2021-04-12T07:47:11 | {} | NONE |
ClickHouse | ClickHouse | 817,199,432 | 22,297 | alexey-milovidov | The only remaining is Arcadia. | 2021-04-10T20:39:45 | 2021-04-10T20:39:45 | {} | MEMBER |
ClickHouse | ClickHouse | 817,322,313 | 18,011 | chenzhiping007 | > enabling/disabling compact parts for a table
what is "enabling/disabling compact parts for a table" ,which setting variables related to
| 2021-04-11T15:04:33 | 2021-04-11T15:04:33 | {} | NONE |
ClickHouse | ClickHouse | 817,647,530 | 22,977 | KochetovNicolai | `Yandex synchronization check` is broken in other pr.
Test with ODBC is flacky. | 2021-04-12T09:26:58 | 2021-04-12T09:26:58 | {} | MEMBER |
ClickHouse | ClickHouse | 818,063,052 | 22,314 | alexey-milovidov | The performance of floating point division has been improved due to migration to clang-11:
https://github.com/ClickHouse/ClickHouse/blob/master/docs/en/whats-new/changelog/2020.md#buildtestingpackaging-improvement-4 | 2021-04-12T19:08:46 | 2021-04-12T19:08:46 | {} | MEMBER |
ClickHouse | ClickHouse | 818,065,768 | 22,314 | alexey-milovidov | `SELECT intDiv(rand64(), 1000000000) FROM numbers(10000000000) FORMAT Null`
- 681 million rows/sec.
`SELECT divide(rand64(), 1000000000) FROM numbers(10000000000) FORMAT Null`
- 750 million rows/sec.
`SELECT toUInt32(divide(rand64(), 1000000000)) FROM numbers(10000000000) FORMAT Null`
- 506 million rows/sec.
| 2021-04-12T19:13:23 | 2021-04-12T19:13:23 | {} | MEMBER |
ClickHouse | ClickHouse | 818,186,918 | 22,980 | alexey-milovidov | `SELECT * FROM system.settings WHERE name = 'max_threads'`
\- to show the current value of `max_threads`
`SET max_threads = ...`
\- to set it to a new value for session before executing the query.
`users.xml`, put it into profiles/default to change the default. | 2021-04-12T20:32:24 | 2021-04-12T20:32:24 | {} | MEMBER |
ClickHouse | ClickHouse | 818,229,808 | 22,996 | vzakaznikov | @alexey-milovidov, we can close this issue as the behavior is not a bug in LIVE VIEW tables. The only potential issue is the inability of the ClickHouse server to detect dropped HTTP connections. | 2021-04-12T20:53:34 | 2021-04-12T20:53:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,286,310 | 22,945 | alesapin | > Ok.
>
> In previous versions task assignment was also performed from multiple threads.
> But it was more complicated and now it is better.
In the previous version, there was no such thing as "task assignment". All threads were just trying to grab a task and execute it immediately. So in previous version we wou... | 2021-04-11T10:43:58 | 2021-04-11T10:43:58 | {} | MEMBER |
ClickHouse | ClickHouse | 817,357,351 | 7,229 | den-crane | > configuration in `remote_servers` -- the hostname and their order actually already tell us what replica and what shard belongs to.
I have 4 clusters in remote_servers with 12 nodes.
1 shard * 12 replicas, 12 shards * 1 replica, 4 shards * 3 replicas, 6 shards * 2 replicas.
The same time a node has 4 different s... | 2021-04-11T19:11:07 | 2021-04-11T19:16:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,551,371 | 22,985 | SE2AI | I try with Nullable type, it solves the problem
`select record_id, note from demo where CAST(note AS Nullable(DECIMAL(18,0))) is not null limit 3;` | 2021-04-12T07:15:17 | 2021-04-12T08:00:58 | {} | NONE |
ClickHouse | ClickHouse | 817,703,798 | 15,621 | thyn | Can you share any information about that issue.
Can' set row filter for distributed table . Row Policy works only on none distributed tables
Clickhouse version is 20.12.4.5 | 2021-04-12T10:43:27 | 2021-04-12T10:43:27 | {} | NONE |
ClickHouse | ClickHouse | 817,727,383 | 22,945 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/22989 | 2021-04-12T11:20:15 | 2021-04-12T11:20:15 | {} | MEMBER |
ClickHouse | ClickHouse | 817,717,775 | 22,760 | cfroystad | Is there any way to get the output of the test (given that I'm currently unable to get the test environment running locally)?
At first glance, I'm unable to see what makes the tests fail when the identical operation works when executed manually... Any guidance would be greatly appreciated!
EDIT 12.04 13:54: Scrat... | 2021-04-12T11:04:45 | 2021-04-12T11:55:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,903,872 | 22,755 | alesapin | ```
2021.04.12 01:25:23.463909 [ 243 ] {} <Error> Application: DB::Exception: Listen [::]:9005 failed: Poco::Exception. Code: 1000, e.code() = 98, e.displayText() = Net Exception: Address already in use: [::]:9005 (version 21.5.1.6533)
``` | 2021-04-12T15:24:55 | 2021-04-12T15:27:44 | {} | MEMBER |
ClickHouse | ClickHouse | 818,027,879 | 22,996 | loyd | But emergencies are still possible, so the use of `JSONEachRowWithProgress` seems to be a more reliable method. | 2021-04-12T18:22:45 | 2021-04-12T18:22:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,342,726 | 7,229 | Tasselmi | @filimonov @alexey-milovidov | 2021-04-11T17:29:20 | 2021-04-11T17:29:20 | {} | NONE |
ClickHouse | ClickHouse | 815,642,537 | 22,831 | godliness | Seem like we had an optimized PR about this, but it is for regular file not stdin: https://github.com/ClickHouse/ClickHouse/pull/10580/files#diff-99d522ea02f81cbff9c5c1490075962f0f695f866e4b6b30e55373f5f66a662fR90
But only import one orc file(~300MB) which cost 55GB memory is very wired, even the entire file load in... | 2021-04-08T10:19:47 | 2021-04-12T00:56:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,861,267 | 4,006 | TCeason | In 5.7.9 the query is not supported. But may be you can try to use this model:
Master 5.7.9->Slave 5.7.28->ClickHouse
Or upgrade your 5.7.9 to a new version. | 2021-04-12T14:29:57 | 2021-04-12T14:29:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 818,018,359 | 20,173 | alexey-milovidov | Done in 21.4. | 2021-04-12T18:08:45 | 2021-04-12T18:08:53 | {} | MEMBER |
ClickHouse | ClickHouse | 818,080,870 | 22,945 | alesapin | ```
2021.04.12 13:21:52.277723 [ 293 ] {} <Debug> Application: Uninitializing subsystem: Logging Subsystem
2021.04.12 13:21:52.277925 [ 294 ] {} <Trace> BaseDaemon: Received signal -2
2021.04.12 13:21:52.278002 [ 294 ] {} <Information> BaseDaemon: Stop SignalListener thread
2021.04.12 13:22:08.129291 [ 471320 ] {} ... | 2021-04-12T19:39:18 | 2021-04-12T19:39:18 | {} | MEMBER |
ClickHouse | ClickHouse | 817,195,815 | 22,942 | alexey-milovidov | Don't understand how `sleep` helps. | 2021-04-10T20:07:38 | 2021-04-10T20:07:38 | {} | MEMBER |
ClickHouse | ClickHouse | 817,314,815 | 22,697 | TCeason | There are some cinflicts. I will fixed them. @kssenii | 2021-04-11T14:14:13 | 2021-04-11T14:14:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,360,549 | 22,529 | alexey-milovidov | Obscure feature. | 2021-04-11T19:34:49 | 2021-04-11T19:34:49 | {} | MEMBER |
ClickHouse | ClickHouse | 817,433,830 | 7,229 | Tasselmi | > > configuration in `remote_servers` -- the hostname and their order actually already tell us what replica and what shard belongs to.
>
> I have 4 clusters in remote_servers with 12 nodes.
> 1 shard * 12 replicas, 12 shards * 1 replica, 4 shards * 3 replicas, 6 shards * 2 replicas.
> The same time a node has 4 di... | 2021-04-12T02:18:46 | 2021-04-12T02:18:46 | {} | NONE |
ClickHouse | ClickHouse | 817,517,313 | 22,963 | filimonov | It's quite straight forward. See https://gitlab.kitware.com/cmake/community/-/wikis/doc/cpack/Packaging-With-CPack
@traceon have some experience with that. | 2021-04-12T06:17:07 | 2021-04-12T06:23:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,748,208 | 22,290 | KochetovNicolai | ```sql
SELECT NULL FROM (SELECT untuple((NULL, dummy)))
``` | 2021-04-12T11:55:51 | 2021-04-12T11:55:51 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 817,775,725 | 22,978 | alexey-milovidov | PVS check:
```
Fail/Info | LibraryBridge.cpp:1 | | V008 | Unable to start the analysis on this file.
-- | -- | -- | -- | --
``` | 2021-04-12T12:38:35 | 2021-04-12T12:38:35 | {} | MEMBER |
ClickHouse | ClickHouse | 818,008,982 | 22,980 | alexey-milovidov | If you will share the dataset (maybe obfuscated) we will be able to investigate the memory usage. | 2021-04-12T17:55:25 | 2021-04-12T17:55:25 | {} | MEMBER |
ClickHouse | ClickHouse | 818,219,515 | 22,976 | alexey-milovidov | Before this change, when user is installing .deb package interactively, it will ask about default password with pink screen, but:
- it works only in debian packages, not in rpm;
- it works only during interactive installation, not in scripted install;
- for some reason, it was completely broken and confused people (... | 2021-04-12T20:48:17 | 2021-04-12T20:51:23 | {} | MEMBER |
ClickHouse | ClickHouse | 817,322,600 | 22,843 | chenzhiping007 | what is "enabling/disabling compact parts for a table" ,which setting variables related to | 2021-04-11T15:06:06 | 2021-04-11T15:06:06 | {} | NONE |
ClickHouse | ClickHouse | 817,519,964 | 22,503 | hexiaoting | @CurtizJ In my test, `alter table xxxx` , And I give the serverError code, Is that enough?
I don't know why test `01788_update_nested_type_subcolumn_check ` is not passed in fast test? Can you help me? | 2021-04-12T06:22:16 | 2021-04-12T06:22:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,715,981 | 22,980 | gituser | @den-crane, just tried this, unfortunately, this option didn't help - the issue is still there. | 2021-04-12T11:02:32 | 2021-04-12T11:14:17 | {} | NONE |
ClickHouse | ClickHouse | 817,728,269 | 22,980 | alexey-milovidov | Lower the value of
```
<mark_cache_size>5368709120</mark_cache_size>
```
in `config.xml` to `1000000000` | 2021-04-12T11:21:42 | 2021-04-12T11:21:57 | {} | MEMBER |
ClickHouse | ClickHouse | 818,063,666 | 22,314 | alexey-milovidov | To compare intDiv with floating point division, we also need to add truncation after floating point division.
Example:
```
SELECT toUInt32(divide(rand64(), 1000000000)) FROM numbers(10000000000) FORMAT Null
``` | 2021-04-12T19:09:50 | 2021-04-12T19:09:50 | {} | MEMBER |
ClickHouse | ClickHouse | 818,111,199 | 21,972 | kssenii | diff from clickhouse-benchmark with ClickHouse DSN (ANSI) looks more than 2 times better:
Poco: https://gist.github.com/kssenii/83eebd1033c340ab8531466a7719aa19
nanodbc: https://gist.github.com/kssenii/27bec1436b7b758913dfc4e50c11b392 | 2021-04-12T19:56:26 | 2021-04-12T19:56:26 | {
"hooray": 2
} | MEMBER |
ClickHouse | ClickHouse | 817,435,878 | 7,229 | den-crane | > so , what you mean is that you run multi clickhouse instances in one node, right ? How could this happen, I've no idea, Is there any document? THx.
No. I mean that Clickhouse instance can be a member of any number of clusters at the same time. | 2021-04-12T02:26:00 | 2021-04-12T02:32:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,649,929 | 4,006 | tavplubix | > mysql: 5.7.9
> find run this sql error : Unknown column 'Value' in 'where clause'
> but , this sql can run in mysql 5.7.28、5.7.17、5.6.16 .
Looks like a bug in MySQL. Maybe it related somehow to https://bugs.mysql.com/bug.php?id=79549, but I'm not sure.
| 2021-04-12T09:30:01 | 2021-04-12T09:30:01 | {} | MEMBER |
ClickHouse | ClickHouse | 817,758,349 | 22,987 | LutzWeischerFujitsu | getting
...
../src/Client/HedgedConnectionsFactory.cpp: In member function 'std::__1::vector<DB::Connection*> DB::HedgedConnectionsFactory::getManyConnections(DB::PoolMode)':
../src/Client/HedgedConnectionsFactory.cpp:98:31: error: 'max_entries' may be used uninitialized in this function [-Werror=maybe-uninitiali... | 2021-04-12T12:11:23 | 2021-04-12T12:11:23 | {} | NONE |
ClickHouse | ClickHouse | 817,295,091 | 17,769 | xjewer | @alexey-milovidov failed tests don't seem to be related to the change | 2021-04-11T11:54:37 | 2021-04-11T11:54:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,340,777 | 7,229 | Tasselmi | > And how are going to set `<shard>` and `<layer>` ?
> I just use `chef` and generate macros.xml using internal cmdb which knows what node is which shard, and substitutes <replica with hostname.
hi，please see my latest comment. I think we should extract macros from `remote_servers ` configuration. How do u think? t... | 2021-04-11T17:15:20 | 2021-04-11T17:15:20 | {} | NONE |
ClickHouse | ClickHouse | 817,440,936 | 7,229 | Tasselmi | > > so , what you mean is that you run multi clickhouse instances in one node, right ? How could this happen, I've no idea, Is there any document? THx.
>
> No. I mean that Clickhouse instance can be a member of any number of clusters at the same time.
ok. I totally understand what you say. In this situation , we ... | 2021-04-12T02:44:09 | 2021-04-12T02:44:09 | {} | NONE |
ClickHouse | ClickHouse | 817,585,620 | 22,985 | SE2AI | @UnamedRus Thanks, it's more convenient, and really a flexible way. | 2021-04-12T08:04:23 | 2021-04-12T08:04:23 | {} | NONE |
ClickHouse | ClickHouse | 817,646,168 | 22,973 | vdimir | Do you mean that we should not perform this retry?
https://github.com/ClickHouse-Extras/aws-sdk-cpp/blob/7d48b2c8193679cc4516e5bd68ae4a64b94dae7d/aws-cpp-sdk-core/source/client/AWSClient.cpp#L967 | 2021-04-12T09:25:12 | 2021-04-12T09:25:31 | {} | MEMBER |
ClickHouse | ClickHouse | 818,008,637 | 22,980 | alexey-milovidov | It may be related due to more efficient (more parallel) processing of queries with multiple UNION ALL.
Does lowering the number of threads helps? | 2021-04-12T17:54:49 | 2021-04-12T17:54:49 | {} | MEMBER |
ClickHouse | ClickHouse | 818,222,712 | 22,976 | alexey-milovidov | https://youtu.be/BBM0QbpPA0Q?t=640 (run at 2x speed) | 2021-04-12T20:49:50 | 2021-04-12T20:49:50 | {} | MEMBER |
ClickHouse | ClickHouse | 852,166,003 | 24,783 | den-crane | > it will establish nullable fields in the distributed table, so the SQL statement above will report an error
Please show the example.
Steps to reproduce.
I cannot repro this
```
CREATE TABLE juba_sessioncount_replica
(
`_id` String,
`resourcedomain` String,
`date_year` Int32,
`keywords`... | 2021-06-01T14:21:09 | 2021-06-01T14:21:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 852,834,143 | 23,964 | azat | > Why do we even have any default value for Settings in program_options? The Settings handle this by itself.
There is no default option for settings in program options right now.
> Implicit value looks weird, e.g. I write clickhouse-client --port and it means --port 9000? Why do we need this?
I was thinking ab... | 2021-06-02T08:08:37 | 2021-06-02T08:08:37 | {} | MEMBER |
ClickHouse | ClickHouse | 852,784,091 | 24,804 | yoavo-datricks | Hi. Thank you for your helpful response.
I tried testing the issue with other types and forms of "Create" statements and it seems PG has a bigger bug when it comes to attndims.
The following three create statements:
CREATE TABLE t1 (
id varchar NULL,
arr_t _int8 NULL
);
CREATE TABLE t2 (
id varchar ... | 2021-06-02T06:46:23 | 2021-06-02T11:13:37 | {} | NONE |
ClickHouse | ClickHouse | 852,030,021 | 23,964 | akuzm | Why do we even have any default value for Settings in program_options? The Settings handle this by itself.
Implicit value looks weird, e.g. I write `clickhouse-client --port` and it means `--port 9000`? Why do we need this? | 2021-06-01T10:55:41 | 2021-06-01T10:55:41 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.