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 | 642,387,036 | 11,551 | alexey-milovidov | @YiuRULE To attach existing data, you can simply use ATTACH TABLE query instead of CREATE TABLE. | 2020-06-11T03:40:39 | 2020-06-11T03:40:39 | {} | MEMBER |
ClickHouse | ClickHouse | 642,420,576 | 11,555 | zhuyoucui0909 | @den-crane @alexey-milovidov Thanks very much for all your response.
Your comments are precious for us.
How I wish ClickHouse English official document can be updated in time! | 2020-06-11T05:40:44 | 2020-06-11T05:40:44 | {} | NONE |
ClickHouse | ClickHouse | 642,497,532 | 11,590 | azat | >Performance — 1 errors, 8 faster, 232 slower, 215 unstable
[right.svg.zip](https://github.com/ClickHouse/ClickHouse/files/4763546/right.svg.zip)
So it does used, but slower, need to dig | 2020-06-11T08:31:13 | 2020-06-11T08:31:13 | {} | MEMBER |
ClickHouse | ClickHouse | 642,562,083 | 10,601 | 4ertus2 | Honest MergeJoin logic removed in favour of #11596. Current prototype shows that it's hard to parallelyze MergeJoin of big tables without buckets. | 2020-06-11T10:38:07 | 2020-06-11T10:38:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,475,931 | 11,546 | dmitryikh | @den-crane , Yes! you right! I missed that.
Thank you.
I'm going to close the issue. | 2020-06-09T17:54:07 | 2020-06-09T17:54:07 | {} | NONE |
ClickHouse | ClickHouse | 641,678,100 | 11,508 | etah000 | For my test case in integration test, it reported "RequestsDependencyWarning: urllib3 (1.23) or chardet (3.0.4) doesn't match a supported version!", this is related with "requests" package in python.
For other failed tests, I can see no relation with the changes I made, so what shall I do next? | 2020-06-10T02:15:05 | 2020-06-10T02:15:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,955,238 | 11,561 | Jokser | @nikitamikhaylov It was related with the destruction (alloc-dealloc-mismatch) of default response stream created by streamFactory. We actually don't need this default response stream because it set explicitly in HTTPClient. I turned off creation of this stream using streamFactory that returns nullptr. After that the pr... | 2020-06-10T11:58:12 | 2020-06-10T11:58:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,989,930 | 11,555 | den-crane | I don't know. Maybe it was transparent some YEARS(?) ago. But I saw an evidence that it's not in 19.... maybe it just broken starting with 19.... | 2020-06-10T12:59:41 | 2020-06-10T13:00:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,061,170 | 10,839 | cailin-lai | Hi @filimonov ,
I've faced similiar case, unfortunately. Are there workarounds to avoid this issue?
Appreciate! | 2020-06-10T14:51:26 | 2020-06-10T14:51:26 | {} | NONE |
ClickHouse | ClickHouse | 642,183,958 | 11,564 | bobrik | Another question I have is whether this stage is supposed to be parallelized. From what I see it's a single thread that does the primary key mark filtering.
In addition to that I'm seeing that `DB::MergeTreeDataSelectExecutor::filterMarksUsingIndex()` also runs in the same thread for all parts. I'd be happier if it ... | 2020-06-10T18:29:48 | 2020-06-10T18:29:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,989,040 | 11,563 | azat | >What will be in case that there is no query_log or metric_log in config.xml, but we execute SYSTEM FLUSH LOGS query?
Nothing (and this is covered in the test, for `part_log`) | 2020-06-10T12:57:55 | 2020-06-10T12:57:55 | {} | MEMBER |
ClickHouse | ClickHouse | 642,234,743 | 11,028 | vitlibar | > @vitlibar It looks ready. Let's merge with master and then merge into master...
Yes, we can merge it as is
However I suppose there is something that is to be improved:
- Check ranges for numeric types, for example for `d` in `TINYINT(d)` ([comment]( https://github.com/ClickHouse/ClickHouse/pull/11028#discuss... | 2020-06-10T20:17:43 | 2020-06-10T20:23:04 | {} | MEMBER |
ClickHouse | ClickHouse | 642,289,456 | 10,787 | alexey-milovidov | @bobrik `copy_user` is also eliminated with:
- Linux AIO that is enabled with `min_bytes_to_use_direct_io` setting;
- mmap that is enabled with `min_bytes_to_use_mmap_io` setting.
| 2020-06-10T21:58:50 | 2020-06-10T21:58:50 | {} | MEMBER |
ClickHouse | ClickHouse | 642,509,097 | 11,162 | KochetovNicolai | I've decided to not backport this pr at all. | 2020-06-11T08:54:14 | 2020-06-11T08:54:14 | {} | MEMBER |
ClickHouse | ClickHouse | 642,515,034 | 11,288 | Enmk | `extractAllGroupsVertical` + `extractAllGroupsHorizontal`: https://github.com/ClickHouse/ClickHouse/pull/11554 | 2020-06-11T09:03:40 | 2020-06-11T09:03:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,578,062 | 11,581 | vzakaznikov | My recommendation is leave exit codes as is and close this issue. Any changes might break more things that it will fix. For example, all the CI tests that check exit codes. It is very unfortunate that due to legacy stuff exit codes are limited to 256. | 2020-06-11T11:16:50 | 2020-06-11T11:16:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,379,359 | 11,557 | blinkov | сс @zhang2014 @sundy-li | 2020-06-09T15:29:36 | 2020-06-09T15:29:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,397,466 | 11,505 | zhbdesign | We have a lot of tables that need to be inserted. This error occurs when a large number of tables are inserted. The server load is also high | 2020-06-09T15:55:54 | 2020-06-09T15:55:54 | {} | NONE |
ClickHouse | ClickHouse | 641,564,748 | 11,401 | akuzm | > > Experiment #3: muzzy decay 10 ms, MADV_FREE, no MADV_DONTNEED
>
> @akuzm AFAIK flamegraphs are build automatically, and they are in archive, which one you posted in the [#11401 (comment)](https://github.com/ClickHouse/ClickHouse/pull/11401#issuecomment-641314974) (tried play with SHA in the link - did not work)?... | 2020-06-09T20:55:42 | 2020-06-09T20:55:42 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,722,045 | 11,555 | zhuyoucui0909 | @den-crane Thanks for your response.
If it is not transparent from other session perspective, what does it mean that "All tables are renamed under global locking." in the official document https://clickhouse.tech/docs/en/sql-reference/statements/misc/#misc_operations-rename
What's the meaning to support multiple ta... | 2020-06-10T05:00:24 | 2020-06-10T05:00:24 | {} | NONE |
ClickHouse | ClickHouse | 642,110,817 | 11,567 | filimonov | It is already fixed in https://github.com/ClickHouse/ClickHouse/pull/11388
Will be released in upcoming 20.5.
In the meanwhile you can set `format_avro_schema_registry_url` in default profile in users.xml and restart the server/ | 2020-06-10T16:11:23 | 2020-06-10T16:11:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,187,187 | 11,574 | alexey-milovidov | Most likely, the `Release` and `Release.gpg` files got out of sync due to caching by CDN. | 2020-06-10T18:36:29 | 2020-06-10T18:36:29 | {} | MEMBER |
ClickHouse | ClickHouse | 642,378,308 | 6,359 | sergiuhlihor | Is there any plan to address caching issue in the future with an out of the box solution like all other major databases? Relying on OS for page cache works many times but is not a reliable solution, specially for non dedicated servers. | 2020-06-11T03:08:14 | 2020-06-11T03:08:14 | {} | NONE |
ClickHouse | ClickHouse | 642,583,745 | 4,646 | filimonov | Does it reproduces on newer ClickHouse versions? | 2020-06-11T11:30:30 | 2020-06-11T11:30:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,705,694 | 11,505 | den-crane | Still unclear what is going on. Need more info about queries, tables, environment.
>DB::NetException: Timeout: connect timed out: ...*:9002 (hadoop-master2:9002)
How this error is related to inserts? What is `hadoop-master2:9002`? | 2020-06-10T03:59:18 | 2020-06-10T03:59:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,999,438 | 11,555 | den-crane | it really easy to repro.
20.4.3.16, 19.16.16.72
```
create table X1(A Int64) Engine=Log;
create table X2(A Int64) Engine=Log;
=> cat rename_test.sql
rename table X1 to X3, X2 to X1;
rename table X1 to X2, X3 to X1;
while true; do clickhouse-client -n < rename_test.sql; sleep .05; done
```
other window
... | 2020-06-10T13:13:22 | 2020-06-10T13:26:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,088,814 | 11,484 | alesapin | `test_version_update_after_mutation` already fixed, this failure with old version: `2020.05.27 07:38:09.851100 [ 1 ] {} <Information> : Starting ClickHouse 20.1.6.30 with revision 54431` Now we use https://github.com/clickhouse/ClickHouse/blob/master/tests/integration/test_version_update_after_mutation/test.py#L9 and i... | 2020-06-10T15:34:57 | 2020-06-10T15:35:07 | {} | MEMBER |
ClickHouse | ClickHouse | 642,200,605 | 10,787 | bobrik | Looking at flamegraphs for some queries I see quite a bit of `copy_user_enhanced_fast_string`, which is supposed to be alleviated with `io_uring`, as far as I understand:
| 2020-06-10T19:06:33 | 2020-06-10T19:06:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,441,038 | 11,401 | akuzm | Experiment #3: muzzy decay 10 ms, MADV_FREE, no MADV_DONTNEED
[Result](https://clickhouse-test-reports.s3.yandex.net/11401/2f9f93c69d3423b79adcce1805c2e5661fb74baa/performance_comparison/report.html#fail1)
Some queries are faster, but some are slower. Interestingly, no mmap calls from jemaloc registered by the prof... | 2020-06-09T16:54:21 | 2020-06-10T20:19:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,264,591 | 11,581 | filimonov | I'm wondering is there to much sense to return different exit codes if they can't be mapped in injective manner to clickhouse exception codes. May be we should just return constant value on exception always? Or just group errors in few groups - as it happen for http status, and return one of few possible error codes?
... | 2020-06-10T21:07:37 | 2020-06-10T21:07:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,315,293 | 8,520 | bobrik | @alexey-milovidov should it work for skipping indices as well? I'm trying:
```sql
set min_bytes_to_use_mmap_io = 1024
```
```sql
select *
from system.settings
where name = 'min_bytes_to_use_mmap_io'
```
```
┌─name─────────────────────┬─value─┬─changed─┬─description────────────────────────────────────... | 2020-06-10T23:16:10 | 2020-06-10T23:16:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,817,613 | 11,516 | alesapin | Performance seems broken on master. | 2020-06-10T08:06:06 | 2020-06-10T08:06:06 | {} | MEMBER |
ClickHouse | ClickHouse | 642,052,718 | 10,763 | alexey-milovidov | We can simply make `--split-binary` by default for unbundled build.
@alesapin What do you think? | 2020-06-10T14:37:34 | 2020-06-10T14:37:34 | {} | MEMBER |
ClickHouse | ClickHouse | 641,393,475 | 11,529 | kamalov-ruslan | > The same should be done for `anyLast`.
>
> What will happen if we have `any(a + b)`? Would it be transformed to `any(a) + any(b)` and is it beneficial?
1)So, ok, I will do it to `anyLast`
2) Yes, any(a + b) converts to any(a) + any(b)
3)I think yes, because, local tests shows better results in any(a) + any(b)... | 2020-06-09T15:49:47 | 2020-06-09T15:49:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,400,090 | 11,459 | vzakaznikov | > Yeah - we might just need to throw some more ram at this either way. `LiveView` is an abstraction over @den-crane's solution since they're both storing results in memory for live querying. Appreciate the response
@jakesyl, if you did want to try to get live views to work with your case, we would first need to crea... | 2020-06-09T16:00:17 | 2020-06-09T16:00:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,622,336 | 11,515 | alexey-milovidov | > Performance — Errors while building the report.
I've fixed at least one of the errors in c95d578. | 2020-06-09T22:53:00 | 2020-06-09T22:53:00 | {} | MEMBER |
ClickHouse | ClickHouse | 641,956,881 | 11,571 | alexey-milovidov | @blinkov Let's prevent this data rot in future. | 2020-06-10T12:00:57 | 2020-06-10T12:00:57 | {} | MEMBER |
ClickHouse | ClickHouse | 642,044,004 | 11,575 | alesapin | it works!
```
:) select comment_body from issues where number=11575 order by comment_updated_at desc limit 1;
SELECT comment_body
FROM issues
WHERE number = 11575
ORDER BY comment_updated_at DESC
LIMIT 1
┌─comment_body─┐
│ it works! │
└──────────────┘
``` | 2020-06-10T14:23:39 | 2020-06-10T14:26:44 | {} | MEMBER |
ClickHouse | ClickHouse | 642,192,641 | 11,574 | blinkov | I've just tuned the caching to be less aggressive.
@AbelVM, please let us know if you happen to see this again in the future. | 2020-06-10T18:48:32 | 2020-06-10T18:48:32 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,417,797 | 6,757 | zhuyoucui0909 | @alexey-milovidov Could you explain more about `EXCHANGE TABLES`? From which version it is supported? How to create a database of Atomic engine? How to exchange two tables? Could you help to give some examples?
Thanks | 2020-06-11T05:31:49 | 2020-06-11T05:31:49 | {} | NONE |
ClickHouse | ClickHouse | 642,010,939 | 6,757 | alexey-milovidov | @winter7 It's not possible.
As an alternative solution we have developed `Atomic` database engine that is going to be used by default. And this engine has support for dedicated `EXCHANGE TABLES` query. | 2020-06-10T13:30:38 | 2020-06-10T13:30:38 | {} | MEMBER |
ClickHouse | ClickHouse | 642,051,943 | 11,574 | alexey-milovidov | What Linux distribution and version do you use?
Is there a way to reproduce this issue in Docker? | 2020-06-10T14:36:18 | 2020-06-10T14:36:18 | {} | MEMBER |
ClickHouse | ClickHouse | 641,563,422 | 11,540 | alexey-milovidov | Tests Ok. | 2020-06-09T20:53:05 | 2020-06-09T20:53:05 | {} | MEMBER |
ClickHouse | ClickHouse | 641,853,142 | 8,968 | liwanxue | I already figure out the root cause of my issue, it caused by wrong configuration of zookeeper, which lead to zookeeper can't elected leader or brain split, which caused too many parts. Thanks all of you. | 2020-06-10T08:52:17 | 2020-06-10T08:52:17 | {} | NONE |
ClickHouse | ClickHouse | 642,062,117 | 11,574 | alexey-milovidov | Yes, it is reproduced by the following commands:
```
docker run -it --rm ubuntu:18.04 /bin/bash
apt update
apt-get install apt-transport-https ca-certificates dirmngr
apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv E0C56BD4
echo "deb https://repo.clickhouse.tech/deb/stable/ main/" | tee /etc/apt/... | 2020-06-10T14:52:53 | 2020-06-10T14:52:53 | {} | MEMBER |
ClickHouse | ClickHouse | 642,400,959 | 8,520 | bobrik | I'm struggling to make it work for data files either:
```sql
SET min_bytes_to_use_mmap_io = 1024
```
```sql
SELECT *
FROM system.settings
WHERE name = 'min_bytes_to_use_mmap_io'
```
```
┌─name─────────────────────┬─value─┬─changed─┬─description────────────────────────────────────────────────────────────... | 2020-06-11T04:36:39 | 2020-06-11T04:36:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,402,970 | 11,247 | mdianjun | +1
```
:) insert into hive_data SELECT number, number+100 FROM numbers(10);
INSERT INTO hive_data SELECT
number,
number + 100
FROM numbers(10)
Received exception from server (version 20.4.2):
Code: 243. DB::Exception: Received from localhost:9000. DB::Exception: Cannot reserve 1.00 MiB, not enough... | 2020-06-09T16:05:07 | 2020-06-09T16:24:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,876,363 | 11,566 | qoega | Which version do you use?
What is the current `max_memory_usage` on your cluster?
We have significantly improved memory tracking in 20.4 release.
max_bytes_before_external_sort/max_bytes_before_external_group_by flushed data on disk if sort and group by can't allocate more memory. Probably it is not used as your ... | 2020-06-10T09:28:35 | 2020-06-10T09:28:35 | {} | MEMBER |
ClickHouse | ClickHouse | 641,883,108 | 11,563 | azat | Test failures looks unrelated (sanitize trap for 01069_database_memory) | 2020-06-10T09:38:20 | 2020-06-10T09:38:20 | {} | MEMBER |
ClickHouse | ClickHouse | 642,011,345 | 11,555 | alexey-milovidov | Documentation is wrong. | 2020-06-10T13:31:16 | 2020-06-10T13:31:16 | {} | MEMBER |
ClickHouse | ClickHouse | 642,064,321 | 11,574 | alexey-milovidov | Also reproduces on Ubuntu 20.04. | 2020-06-10T14:56:18 | 2020-06-10T14:56:18 | {} | MEMBER |
ClickHouse | ClickHouse | 642,298,722 | 11,498 | azat | @abyss7 @filimonov can you take a look? This is more "complete" replacement for #11391
Note, that it has a little bit awkward syntax (I guess), you can find usage example in the PR description. | 2020-06-10T22:23:08 | 2020-06-10T22:23:08 | {} | MEMBER |
ClickHouse | ClickHouse | 642,511,940 | 11,432 | KochetovNicolai | I suppose the correct solution is to not create any metadata on disk before checks.
Or create it in temporary directory and move after checks passed.
Or remove after checks, but only if there is no data in table. | 2020-06-11T08:59:07 | 2020-06-11T08:59:07 | {} | MEMBER |
ClickHouse | ClickHouse | 642,585,730 | 5,147 | filimonov | You can adjsut that behavior by putting few lines in config.xml, like in that sample:
https://github.com/ClickHouse/ClickHouse/blob/18ce66bb3e1616a47e438b703c5d511b674bb605/docs/en/engines/table-engines/integrations/kafka.md#configuration-configuration | 2020-06-11T11:35:42 | 2020-06-11T11:35:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,387,739 | 11,555 | den-crane | Interesting question, but it seems it was never transparent from other sessions perspective.
Other sessions could get `Table doesn't exist.` probably it is not a big deal, and clients can do re-try.
And it was never atomic `RENAME TABLE a TO b, NOTEXISTINGTABLE TO a` will fail and table `a` will be renamed. | 2020-06-09T15:40:08 | 2020-06-09T15:40:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,799,669 | 11,466 | sneakythr0ws | I had two copies running in parallel | 2020-06-10T07:44:02 | 2020-06-10T07:44:02 | {} | NONE |
ClickHouse | ClickHouse | 641,099,261 | 8,639 | KochetovNicolai | **stateless tests**
1. `00810_in_operators_segfault` - `Unknown function findClusterIndex` - ok
2. `00834_cancel_http_readonly_queries_on_client_close`- `Operation timed out after 10001 milliseconds with 0 bytes received` - ??? (failed in prev release too)
3. `00933_test_fix_extra_seek_on_compressed_cache` - was not... | 2020-06-09T07:46:08 | 2020-06-10T11:50:16 | {} | MEMBER |
ClickHouse | ClickHouse | 641,987,343 | 10,763 | azat | >https://github.com/ClickHouse/ClickHouse/pull/10763#issuecomment-626132285
@alesapin @alexey-milovidov do you think that it worth add the following build to the matrix?
```
docker/packager/packager --unbundled --split-binary
``` | 2020-06-10T12:54:47 | 2020-06-10T12:54:47 | {} | MEMBER |
ClickHouse | ClickHouse | 641,988,153 | 11,563 | nikitamikhaylov | What will be in case that there is no `query_log` or `metric_log` in config.xml, but we execute `SYSTEM FLUSH LOGS` query? | 2020-06-10T12:56:13 | 2020-06-10T12:56:48 | {} | MEMBER |
ClickHouse | ClickHouse | 642,059,001 | 11,574 | AbelVM | It's a vanilla AWS Ubuntu 18.04 image. And clickhouse is installed following the guide at https://clickhouse.tech/#quick-start
I have run maintenance tasks today, as checking for new versions and a simple `sudo apt update` gave me those messages.
As the listed key (`C8F1E19FE0C56BD4`) is different from the one on... | 2020-06-10T14:47:50 | 2020-06-10T14:47:50 | {} | NONE |
ClickHouse | ClickHouse | 642,066,984 | 11,579 | tavplubix | Yes, other engines throw exception if there is no create query and `throw_on_error = true` | 2020-06-10T15:00:23 | 2020-06-10T15:00:23 | {} | MEMBER |
ClickHouse | ClickHouse | 642,066,976 | 11,574 | alexey-milovidov | I don't see what's wrong with the key:
http://keyserver.ubuntu.com/pks/lookup?op=get&search=0xc8f1e19fe0c56bd4
```
$ pgpdump pgp
New: Public Key Packet(tag 6)(525 bytes)
Ver 4 - new
Public key creation time - Tue May 24 18:30:45 MSK 2016
Pub alg - RSA Encrypt or Sign(pub 1)
R... | 2020-06-10T15:00:23 | 2020-06-10T15:00:52 | {} | MEMBER |
ClickHouse | ClickHouse | 642,172,811 | 11,581 | alexey-milovidov | @filimonov Let's commit that? | 2020-06-10T18:06:57 | 2020-06-10T18:06:57 | {} | MEMBER |
ClickHouse | ClickHouse | 642,217,858 | 11,488 | alexey-milovidov | `Should select * work as select * from system.one`
It's ok if this query will work but it doesn't have to.
> `select xyz.*`
This query should not work, because there is no `xyz`. | 2020-06-10T19:42:19 | 2020-06-10T19:42:19 | {} | MEMBER |
ClickHouse | ClickHouse | 642,482,250 | 11,162 | azat | This PR should not be backported without #11259 | 2020-06-11T08:01:32 | 2020-06-11T08:01:32 | {} | MEMBER |
ClickHouse | ClickHouse | 642,555,296 | 11,563 | alesapin | @azat Deadlock introduced in this PR: https://gist.github.com/alesapin/d915d1deaa27d49aa31223daded02be2 | 2020-06-11T10:22:20 | 2020-06-11T10:22:20 | {} | MEMBER |
ClickHouse | ClickHouse | 641,817,035 | 4,631 | MrSsunlight | I also encountered the same problem, what is the final solution? What caused it? | 2020-06-10T08:04:57 | 2020-06-10T08:04:57 | {} | NONE |
ClickHouse | ClickHouse | 641,840,434 | 11,568 | qoega | Is there additional text in stderr?
assertion is raised from /usr/lib64/libc-2.17.so. Can you install debug symbols for libc and find what is in 0x35228 address or just rerun server? | 2020-06-10T08:35:05 | 2020-06-10T08:35:05 | {} | MEMBER |
ClickHouse | ClickHouse | 641,882,747 | 11,484 | azat | 01069_database_memory TSAN - https://clickhouse-test-reports.s3.yandex.net/11563/8a78cffe5c3e8b9ac2e26e0731c40306cad4c9e1/functional_stateless_tests_(thread).html#fail1 | 2020-06-10T09:37:43 | 2020-06-10T09:37:43 | {} | MEMBER |
ClickHouse | ClickHouse | 641,908,618 | 11,401 | azat | @akuzm correct me if I'm wrong:
- master - no lazy purging - ok
- muzzy decay enabled - half lazy, half forced - ok
>MADV_DONTNEED is still called a lot
>and the total number of madvise calls is higher.
Indeed, since now it has lazy purging, which is `madvise(MADV_FREE)`
- [muzzy decay 10 ms, MADV... | 2020-06-10T10:23:26 | 2020-06-10T10:29:13 | {} | MEMBER |
ClickHouse | ClickHouse | 642,011,592 | 11,555 | alexey-milovidov | There is no global lock, it was removed in https://github.com/ClickHouse/ClickHouse/issues/6757 | 2020-06-10T13:31:42 | 2020-06-10T13:31:42 | {} | MEMBER |
ClickHouse | ClickHouse | 642,120,108 | 11,566 | maison2710 | currently, we are using version 19.5.3 and plan to upgrade to latest version. The `max_memory_usage` is same default, which is 10GB.
By the way, I would like to know in case there are several heavy queries come at same time, which will put ClickHouse into high, then will ClickHouse somehow handle the number of conc... | 2020-06-10T16:26:53 | 2020-06-10T16:27:14 | {} | NONE |
ClickHouse | ClickHouse | 642,197,924 | 11,582 | alesapin | Only integration tests are important | 2020-06-10T19:01:00 | 2020-06-10T19:01:00 | {} | MEMBER |
ClickHouse | ClickHouse | 642,572,109 | 11,592 | alexey-milovidov | PVS Ok. | 2020-06-11T11:02:26 | 2020-06-11T11:02:26 | {} | MEMBER |
ClickHouse | ClickHouse | 641,347,945 | 9,624 | KochetovNicolai | **stateless tests**
`01249_bad_arguments_for_bloom_filter` - query formatting has changed
`01116_cross_count_asterisks` - `Unknown setting multiple_joins_rewriter_version`
**integration tests**
`test_server_initialization` - works locally
`test_settings_constraints_distributed` (x2) - new syntax in tests (`Synta... | 2020-06-09T14:50:30 | 2020-06-10T11:23:55 | {} | MEMBER |
ClickHouse | ClickHouse | 642,102,650 | 11,484 | alesapin | Fix for test_always_fetch_merged/test.py https://github.com/ClickHouse/ClickHouse/pull/11582 | 2020-06-10T15:57:58 | 2020-06-10T15:57:58 | {} | MEMBER |
ClickHouse | ClickHouse | 642,391,220 | 8,520 | alexey-milovidov | It works only for data files, not indices. | 2020-06-11T03:58:08 | 2020-06-11T03:58:08 | {} | MEMBER |
ClickHouse | ClickHouse | 642,577,853 | 11,288 | filimonov | More Kafka improvements: #11599 | 2020-06-11T11:16:18 | 2020-06-11T11:16:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,434,361 | 11,562 | alexey-milovidov | FYI it is required because Docker images are shared between different releases and older releases are not ready to use newer Docker image. | 2020-06-09T16:48:50 | 2020-06-09T16:48:50 | {} | MEMBER |
ClickHouse | ClickHouse | 642,140,204 | 11,581 | akuzm | ERROR: type should be string, got "https://tldp.org/LDP/abs/html/exitcodes.html\r\n\r\nRespecting all conventional codes, it would be `3 + e.code() % 123` -- not human-readable, but can be useful to distinguish some errors in scripts. Or maybe even `10 + code % 100` so that it's possible to check by hand." | 2020-06-10T17:04:42 | 2020-06-10T17:11:27 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,383,598 | 5,746 | 2hangchen | have you solved this problem？ | 2020-06-11T03:27:36 | 2020-06-11T03:27:36 | {} | NONE |
ClickHouse | ClickHouse | 642,390,180 | 11,580 | alexey-milovidov | > Functional stateless tests (thread) — fail: 1, passed: 1909, skipped: 9
https://github.com/ClickHouse/ClickHouse/pull/11592 | 2020-06-11T03:53:31 | 2020-06-11T03:53:31 | {} | MEMBER |
ClickHouse | ClickHouse | 642,479,498 | 11,259 | azat | This PR also contains "fix" and test for #11162 | 2020-06-11T07:56:10 | 2020-06-11T07:56:10 | {} | MEMBER |
ClickHouse | ClickHouse | 642,585,142 | 5,147 | filimonov | > cause a different group name will not pull All the data.
That is not true (at least now). When you use some new group name engine Kafka by default will start reading the topic from the beginning.
https://github.com/ClickHouse/ClickHouse/blob/18ce66bb3e1616a47e438b703c5d511b674bb605/src/Storages/Kafka/StorageKa... | 2020-06-11T11:34:06 | 2020-06-11T11:34:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,588,531 | 5,128 | filimonov | The situation with logs improved in https://github.com/ClickHouse/ClickHouse/pull/10983 and https://github.com/ClickHouse/ClickHouse/pull/11261 (will be released in 20.5)
Stats & introspection is in TODO | 2020-06-11T11:41:32 | 2020-06-11T11:41:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 641,945,267 | 10,561 | KochetovNicolai | Only `test_default_role` integration tests failed. | 2020-06-10T11:38:54 | 2020-06-10T11:38:54 | {} | MEMBER |
ClickHouse | ClickHouse | 642,056,754 | 10,763 | alesapin | Yes, it's ok, but we still need both unbundled and split + unbundled because we cannot pack split build into deb package, but it's required by functional tests in unbundled mode. | 2020-06-10T14:44:13 | 2020-06-10T14:44:13 | {} | MEMBER |
ClickHouse | ClickHouse | 642,154,059 | 11,432 | tavplubix | It's dangerous to automatically remove a directory in case of failed CREATE query, because we cannot determine if the directory contains garbage or data of some detached table for example. Probably it's better to disable this check and reopen #7005. Or move the directory to some place like `detached/` instead of removi... | 2020-06-10T17:32:15 | 2020-06-10T17:32:15 | {
"+1": 3
} | MEMBER |
ClickHouse | ClickHouse | 642,180,180 | 11,071 | r-zenine | Hi @Akazz,
I just pushed a new version.
Thank you for the review. I did my best to take every comment you made into account.
Have a good day. | 2020-06-10T18:21:52 | 2020-06-10T18:21:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,222,975 | 11,497 | vitlibar | > > I don't think we should allow every user to know other users' names.
>
> This is Ok when access management is turned off.
> This info is already available in SHOW PROCESSLIST and query_log.
>
> > Also we can provide another option show_access in users.xml which will enable readonly access to access managemen... | 2020-06-10T19:53:18 | 2020-06-10T19:53:18 | {} | MEMBER |
ClickHouse | ClickHouse | 642,286,727 | 11,581 | alexey-milovidov | No need to worry too much, let's apply non-injective mapping.
The only important for me is that exit code is non-zero on error. | 2020-06-10T21:51:43 | 2020-06-10T21:51:43 | {} | MEMBER |
ClickHouse | ClickHouse | 641,621,060 | 11,550 | alexey-milovidov | test_always_fetch_merged - known flaky test: #11484
Description check — Changelog entry is not found - already fixed. | 2020-06-09T22:48:36 | 2020-06-09T22:48:36 | {} | MEMBER |
ClickHouse | ClickHouse | 641,625,645 | 11,536 | alexey-milovidov | RapidJSON is accidentially get used instead of SIMDJSON. | 2020-06-09T23:04:14 | 2020-06-09T23:04:14 | {} | MEMBER |
ClickHouse | ClickHouse | 642,031,566 | 11,575 | alesapin | test comment | 2020-06-10T14:02:37 | 2020-06-10T14:02:37 | {} | MEMBER |
ClickHouse | ClickHouse | 642,132,794 | 11,581 | alexey-milovidov | `1 + e.code() % 255` or `-(e.code() % 128)` ? | 2020-06-10T16:50:35 | 2020-06-10T16:50:35 | {} | MEMBER |
ClickHouse | ClickHouse | 642,186,376 | 11,574 | alexey-milovidov | It is strange but everything is ok right now.
Here is a "single command" test:
```
docker run -it --rm ubuntu:18.04 /bin/bash -c '
apt-get update
apt-get install -y apt-transport-https ca-certificates dirmngr
apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv E0C56BD4
echo "deb https:... | 2020-06-10T18:34:45 | 2020-06-10T18:36:41 | {} | MEMBER |
ClickHouse | ClickHouse | 642,216,020 | 11,569 | alexey-milovidov | > Should I manually add a docker_compose.yml in the node1 dir?
No, it should be added automatically and work out of the box.
I'm not sure what is the reason, let's ask @alesapin. | 2020-06-10T19:38:27 | 2020-06-10T19:38:27 | {} | MEMBER |
ClickHouse | ClickHouse | 642,582,043 | 2,163 | filimonov | Fixed long ago. Now Kafka engine do commits only if able to parse all the messages and flush the data to destination tables. | 2020-06-11T11:26:22 | 2020-06-11T11:26:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 639,027,485 | 9,479 | alexey-milovidov | > but the watch_count of zk will increase significantly
The number of active watches in ZK is not an issue. There are some scenarios when watches may "leak" but you should not worry.
| 2020-06-04T18:24:36 | 2020-06-04T18:24:36 | {} | MEMBER |
ClickHouse | ClickHouse | 639,041,505 | 11,434 | den-crane | CH does not `delete` only `alter table .. delete` https://clickhouse.tech/docs/en/sql-reference/statements/alter/#alter-mutations | 2020-06-04T18:40:43 | 2020-06-04T18:40:43 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.