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 | 764,487,413 | 19,349 | qoega | Have you tried to run your queries several times? It can be explained for example with data placed in memory by fs after first data read.
And what is the size of your table? Deduplicated result can be much smaller than complete result. | 2021-01-21T09:02:04 | 2021-01-21T09:02:04 | {} | MEMBER |
ClickHouse | ClickHouse | 764,647,463 | 19,341 | ucasfl | Maybe you can add the test to the same `branch` in #19310. | 2021-01-21T13:37:18 | 2021-01-21T13:37:18 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 763,575,855 | 19,309 | den-crane | 1. upgrade to 20.8.12.2
2. Check system.merges -- long merge can block `drop column` | 2021-01-20T12:32:38 | 2021-01-20T12:32:50 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,646,667 | 19,312 | qoega | #19318 | 2021-01-20T14:33:38 | 2021-01-20T14:33:38 | {} | MEMBER |
ClickHouse | ClickHouse | 764,187,335 | 18,684 | amosbird | all `test_materialize_mysql_database` are still failing. @zhang2014 perhaps you can help... | 2021-01-21T02:19:36 | 2021-01-21T02:19:36 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 764,533,140 | 19,257 | nikitamikhaylov | Need to get rid of `std::stringstream` and add `GeographicPoint` support in every possible function. Also need to do some cosmetics. | 2021-01-21T10:21:03 | 2021-01-21T10:21:03 | {} | MEMBER |
ClickHouse | ClickHouse | 764,581,607 | 19,354 | qoega | Have you checked, that there is a relevant perf test for this function? | 2021-01-21T11:37:52 | 2021-01-21T11:37:52 | {} | MEMBER |
ClickHouse | ClickHouse | 763,506,078 | 19,303 | filimonov | also #19279 | 2021-01-20T10:30:50 | 2021-01-20T10:30:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,641,429 | 19,312 | qoega | > @qoega If the behaviour of ClickHouse release version is not affected - never place a "bug" label.
> We have "fuzz" and "testing" labels for the purpose to highlight these issues.
Lets add separate issue template for Sanitizer reports. It is not an abort() found by fuzzing/stress tests. | 2021-01-20T14:25:39 | 2021-01-20T14:25:39 | {} | MEMBER |
ClickHouse | ClickHouse | 764,109,790 | 19,296 | lucasguo | My fault. I didn't deal with closing connection correctly. After correct closing connection after use, everything is fine. | 2021-01-21T00:49:47 | 2021-01-21T00:49:47 | {} | NONE |
ClickHouse | ClickHouse | 764,461,373 | 19,336 | alexey-milovidov | @tavplubix I've made another fix in https://github.com/ClickHouse/ClickHouse/pull/19347. It looks more complete. | 2021-01-21T08:13:16 | 2021-01-21T08:13:16 | {} | MEMBER |
ClickHouse | ClickHouse | 764,462,110 | 19,336 | alexey-milovidov | Please review my PR then :) | 2021-01-21T08:14:46 | 2021-01-21T08:14:46 | {} | MEMBER |
ClickHouse | ClickHouse | 764,541,674 | 15,889 | qoega | @KochetovNicolai
https://clickhouse-test-reports.s3.yandex.net/19346/5f3059555a039181cdb9b558fb073fc7998ad496/fuzzer_debug/report.html#fail1
```
2021.01.21 12:35:44.409077 [ 80 ] {f28a3642-27c7-4d51-be46-b0ea06062068} <Fatal> : Logical error: 'Block structure mismatch in PipelineExecuting stream: different columns:... | 2021-01-21T10:35:21 | 2021-01-21T10:35:21 | {} | MEMBER |
ClickHouse | ClickHouse | 764,599,418 | 19,204 | keen-wolf | > > Also I found other file access behavior in clickhouse as shown in attached screenshot. In this simple test the “outfile “ command seems to having the authority to write file anywhere without limitation, but when read with the file function, it was refused. I don’t know if this is the primitive design or is some kin... | 2021-01-21T12:07:46 | 2021-01-21T12:07:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,416,171 | 19,229 | filimonov | >> you reach the maximum possible speed of memcpy (typically it's something around 20-30 Gbyte/sec).
>
> This is wrong.
OK. Removed the number. But even if the real number can vary quite significantly and can be worse or better (especially on multithreaded CPUs / multiple CPU systems - multichannel, NUMA, etc) but... | 2021-01-20T08:05:52 | 2021-01-20T11:50:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,768,115 | 19,229 | alexey-milovidov | Also need to activate hardware CPU counters with
```
sudo setcap cap_sys_admin=+ep /usr/bin/clickhouse
```
And check ProfileEvents again. | 2021-01-20T16:34:29 | 2021-01-20T16:34:29 | {} | MEMBER |
ClickHouse | ClickHouse | 763,790,223 | 19,310 | qoega | Can you add a test as well? | 2021-01-20T17:03:27 | 2021-01-20T17:03:27 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 763,964,145 | 11,003 | rrati | Yep. RunAsUser can not be set by users by default on openshift. That requires elevated privs. | 2021-01-20T21:35:59 | 2021-01-20T21:35:59 | {} | NONE |
ClickHouse | ClickHouse | 764,311,392 | 16,264 | amosbird | > @amosbird out of curiosity: how did you discover the source of the leak here?
With some reproduceable query and code analysis. | 2021-01-21T04:53:07 | 2021-01-21T04:53:07 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 764,494,387 | 19,349 | nnsgmsone | Also I use the client to execute the sql, my understanding client will automatically add the limit, so the output should not be the reason for the performance gap. | 2021-01-21T09:14:27 | 2021-01-21T09:14:27 | {} | NONE |
ClickHouse | ClickHouse | 764,566,934 | 19,340 | qoega | Should be fixed there https://github.com/ClickHouse/ClickHouse/commit/fe6b964b32c94b85b001cb40372d8cdc780367f5 | 2021-01-21T11:17:23 | 2021-01-21T11:17:23 | {} | MEMBER |
ClickHouse | ClickHouse | 763,613,500 | 19,314 | qoega | Looks similar
```
2021.01.14 05:00:38.868049 [ 219259 ] {b83e3522-376c-4189-845b-53aa646b112d} <Debug> executeQuery: (from [::1]:58536, using experimental parser) SELECT dictGetUUID();
2021.01.14 05:00:38.869723 [ 220929 ] {} <Fatal> BaseDaemon: ########################################
2021.01.14 05:00:38.870095 [ ... | 2021-01-20T13:42:40 | 2021-01-20T13:42:40 | {} | MEMBER |
ClickHouse | ClickHouse | 763,851,791 | 16,264 | nvartolomei | @amosbird out of curiosity: how did you discover the source of the leak here? | 2021-01-20T18:40:05 | 2021-01-20T18:40:05 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,953,166 | 19,284 | alesapin | Ok, let's merge and see what happen. | 2021-01-20T21:19:02 | 2021-01-20T21:19:02 | {} | MEMBER |
ClickHouse | ClickHouse | 764,472,657 | 19,108 | qoega | same issue in debug stateless tests
https://clickhouse-test-reports.s3.yandex.net/0/d8b9278193405512bc96bd74d3f30d281c481bb9/functional_stateless_tests_(debug).html | 2021-01-21T08:35:30 | 2021-01-21T08:35:30 | {} | MEMBER |
ClickHouse | ClickHouse | 764,513,382 | 19,204 | alexey-milovidov | > Also I found other file access behavior in clickhouse as shown in attached screenshot. In this simple test the “outfile “ command seems to having the authority to write file anywhere without limitation, but when read with the file function, it was refused. I don’t know if this is the primitive design or is some kind ... | 2021-01-21T09:47:00 | 2021-01-21T09:47:00 | {} | MEMBER |
ClickHouse | ClickHouse | 764,586,492 | 19,258 | MaxJPowers | > What do you mean by `awk -F:',' '{print $0}' ` ?
> You probably should use `awk -F ','`
>
> And for cached dictionaries executable script should be much more complicated.
> You need to read STDIN. Because CH passes the requested keys using STDIN.
Hey so any updates? I've run that command and it still is an em... | 2021-01-21T11:42:02 | 2021-01-21T11:42:02 | {} | NONE |
ClickHouse | ClickHouse | 764,594,011 | 19,356 | hanzhongzi | This will work
```sql
CREATE TABLE tutorial.visits_v1 ( `CounterID` UInt32, `StartDate` Date, `Sign` Int8, `IsNew` UInt8, `VisitID` UInt64, `UserID` UInt64, `StartTime` DateTime, `Duration` UInt32, `UTCStartTime` DateTime, `PageViews` Int32, `Hits` Int32, `IsBounce`... | 2021-01-21T11:57:05 | 2021-01-21T11:57:05 | {} | NONE |
ClickHouse | ClickHouse | 764,606,642 | 19,217 | l1t1 | how to display 1.00 ,1.10,… | 2021-01-21T12:21:52 | 2021-01-21T12:21:52 | {} | NONE |
ClickHouse | ClickHouse | 764,615,151 | 19,273 | CurtizJ | > Ok, and this not works in where query.
Fixed: #19358. | 2021-01-21T12:38:17 | 2021-01-21T12:38:17 | {} | MEMBER |
ClickHouse | ClickHouse | 764,628,250 | 19,335 | qoega | same
https://clickhouse-test-reports.s3.yandex.net/0/89c64cbdfc1db9899268780bbe03363770c98fa1/functional_stateful_tests_(debug).html | 2021-01-21T13:02:58 | 2021-01-21T13:02:58 | {} | MEMBER |
ClickHouse | ClickHouse | 763,776,617 | 19,229 | den-crane | >but the RAM still have it's own maximum throughput, which is reachable and can be limitation in some scenarios
But I think it will be reflected as 100% CPU utilization.
Though may be I am mistaken in the HT case (56 physical CPU cores (112 logical cores with HT enabled)) | 2021-01-20T16:45:46 | 2021-01-20T16:50:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,879,660 | 11,003 | rrati | You can run a container that requires root using scc on openshift, however that is generally not the way to run containers on openshift. The creation of an scc requires elevated privileges and in a secure environment may not be given. The better way to do this is for the container image not to require root in order t... | 2021-01-20T19:28:38 | 2021-01-20T19:38:01 | {} | NONE |
ClickHouse | ClickHouse | 763,992,333 | 11,003 | rrati | Think of openshift kind of like linux when running pods/containers. It runs pods as a non-root user. Anything a non-root user would be prevented from doing on linux will probably also be disallowed by default on openshift.
I did a quick test with a dockerfile and it seems that the solutions I provided won't work. ... | 2021-01-20T22:26:28 | 2021-01-20T22:26:28 | {} | NONE |
ClickHouse | ClickHouse | 764,032,333 | 18,672 | akuzm | I am somewhat concerned by the amount of changes. I imagined it could be done by adding a new virtual column `CAST(right-key, Nullable(left-type))` and substituting it for the right join key, and then performing the join normally. And this would be done somewhere at the level of `ExpressionAnalyzer`, and wouldn't requi... | 2021-01-20T23:43:24 | 2021-01-20T23:43:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 764,060,244 | 19,311 | l1t1 | if i change `case when age between lo and hi then id end` to
`case when age between lo and hi then id else 0 end`, no error reports, but its logic is wrong. | 2021-01-21T00:16:41 | 2021-01-21T00:16:41 | {} | NONE |
ClickHouse | ClickHouse | 764,573,904 | 18,945 | qoega | 01600_detach_permanently https://clickhouse-test-reports.s3.yandex.net/0/f8830e0f112298ae2388cc42e7de05ca9edbb982/functional_stateless_tests_(release,_databaseordinary).html | 2021-01-21T11:25:44 | 2021-01-21T11:25:44 | {} | MEMBER |
ClickHouse | ClickHouse | 764,639,027 | 18,945 | qoega | 01115_join_with_dictionary https://clickhouse-test-reports.s3.yandex.net/0/6022311a374092968bb3f1981fabc17dcbd402f4/functional_stateless_tests_(debug).html | 2021-01-21T13:22:25 | 2021-01-21T13:22:25 | {} | MEMBER |
ClickHouse | ClickHouse | 750,041,829 | 18,213 | sunny19930321 | > Having the same issue as well. Please advise.
This problem was found in mysql version 5.7.31 and above，I plan to fix it
fixed https://github.com/ClickHouse/ClickHouse/pull/19310 | 2020-12-23T09:39:29 | 2021-01-20T12:17:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 764,538,258 | 19,353 | alexey-milovidov | All stable versions don't have this issue.
It was introduced in recent commits to master. | 2021-01-21T10:29:47 | 2021-01-21T10:29:47 | {} | MEMBER |
ClickHouse | ClickHouse | 764,584,987 | 19,252 | alexey-milovidov | Regression in master #19353 | 2021-01-21T11:40:11 | 2021-01-21T11:40:11 | {} | MEMBER |
ClickHouse | ClickHouse | 764,588,041 | 17,623 | olafbuitelaar | are there any plans to integrate an apache pulsar consumer? | 2021-01-21T11:45:01 | 2021-01-21T11:45:01 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 764,605,843 | 19,339 | alexey-milovidov | UBSan: #19357 | 2021-01-21T12:20:18 | 2021-01-21T12:20:18 | {} | MEMBER |
ClickHouse | ClickHouse | 763,628,027 | 17,886 | alexey-milovidov | Fixed. | 2021-01-20T14:06:45 | 2021-01-20T14:06:45 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 763,628,240 | 19,312 | alexey-milovidov | #19234 | 2021-01-20T14:07:05 | 2021-01-20T14:07:05 | {} | MEMBER |
ClickHouse | ClickHouse | 763,913,381 | 19,204 | alex-zaitsev | > How do we control that this function will not allow to read data from other databases and private configs? It is very powerful function as it runs with the same user as clickhouse-server.
https://clickhouse.tech/docs/en/operations/server-configuration-parameters/settings/#server_configuration_parameters-user_files... | 2021-01-20T20:30:52 | 2021-01-20T20:30:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 764,489,762 | 19,349 | nnsgmsone | I have executed the sql after clearing the page cache and restarting the service。Data volume of 100 million rows (3GB) | 2021-01-21T09:06:13 | 2021-01-21T09:06:13 | {} | NONE |
ClickHouse | ClickHouse | 764,583,106 | 19,346 | alesapin | Race fixed in https://github.com/ClickHouse/ClickHouse/pull/19355 | 2021-01-21T11:38:54 | 2021-01-21T11:38:54 | {} | MEMBER |
ClickHouse | ClickHouse | 764,590,652 | 19,346 | alesapin | `01079_parallel_alter_add_drop_column_zookeeper` suspicious failure... | 2021-01-21T11:50:37 | 2021-01-21T11:50:37 | {} | MEMBER |
ClickHouse | ClickHouse | 764,604,051 | 19,353 | akuzm | Ohhh the second insert... | 2021-01-21T12:16:41 | 2021-01-21T12:16:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 764,604,311 | 19,346 | alesapin | Didn't fail before:
```
└────────────────────────────────────────────────┴─────────────────────┴──────────────────────────────────────────┴─────────────────────┘
┌─test_name──────────────────────────────────────┬────check_start_time─┬─commit_sha───────────────────────────────┬─pull_request_number─┐
│ 01079_parallel... | 2021-01-21T12:17:08 | 2021-01-21T12:17:08 | {} | MEMBER |
ClickHouse | ClickHouse | 763,495,300 | 19,273 | KochetovNicolai | #19305 | 2021-01-20T10:12:39 | 2021-01-20T10:12:39 | {} | MEMBER |
ClickHouse | ClickHouse | 763,627,563 | 19,312 | alexey-milovidov | @qoega If the behaviour of ClickHouse release version is not affected - never place a "bug" label.
We have "fuzz" and "testing" labels for the purpose to highlight these issues. | 2021-01-20T14:06:01 | 2021-01-20T14:06:01 | {} | MEMBER |
ClickHouse | ClickHouse | 763,766,392 | 19,229 | alexey-milovidov | @filimonov It spent two times more CPU time with 56 threads than with 28 threads.
Now we need to check Query Profiler:
https://clickhouse.tech/docs/en/operations/optimizing-performance/sampling-query-profiler/
And look with https://github.com/laplab/clickhouse-speedscope | 2021-01-20T16:32:24 | 2021-01-20T16:32:24 | {} | MEMBER |
ClickHouse | ClickHouse | 764,214,683 | 19,272 | slimtom95 | I tried to debug the code in `PushingToViewsBlockOutputStream`, It seems the order of MV list of a table maybe random.After we got the view list, before view execution, we don't order them by name. The view order is actually found random between debug tests.
If we design the order of MV execution is sequential, alph... | 2021-01-21T03:43:16 | 2021-01-21T03:55:12 | {} | NONE |
ClickHouse | ClickHouse | 764,253,607 | 19,217 | l1t1 | seems fixed in version 21.1.1.5745, but i has a more question, if i round to 2 decimal, but it only shows digits is not 0
```sql
SELECT round(10 / number, 2) FROM numbers(1, 10)
┌─round(divide(10, number), 2)─┐
│ 10 │
│ 5 │
│ 3.33 │
│... | 2021-01-21T04:19:00 | 2021-01-21T04:19:00 | {} | NONE |
ClickHouse | ClickHouse | 763,983,858 | 19,229 | filimonov | @chirag-applift can you collect the requested info?
Also: How many memory modules do you have installed on the server? What `tiptop` is showing during both tests? `perf stats`? | 2021-01-20T22:12:38 | 2021-01-21T05:56:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 764,462,318 | 17,041 | vdimir | Seems that failed test not connected to changes:
Integration tests (asan): `test_zookeeper_config/test.py::test_secure_connection`
Functional stateless tests (thread): `01085_max_distributed_connections_http` | 2021-01-21T08:15:11 | 2021-01-21T08:15:11 | {} | MEMBER |
ClickHouse | ClickHouse | 764,466,694 | 19,327 | KochetovNicolai | Cannot reproduce this error for query from fuzz tests.
I expect this will be fixed by #19316 | 2021-01-21T08:23:57 | 2021-01-21T08:23:57 | {} | MEMBER |
ClickHouse | ClickHouse | 764,514,296 | 19,204 | alexey-milovidov | Also there is the case:
\- if the query is processed by `clickhouse-local`, restriction should not apply.
(you can see similar logic for `file` table function and storage)
Upd. I see you already implemented this. | 2021-01-21T09:48:39 | 2021-01-21T09:50:11 | {} | MEMBER |
ClickHouse | ClickHouse | 763,944,137 | 11,003 | filimonov | ClickHouse container can work rootless.
Just set user uid 101 / gid 101 and mount in the folders writable for uid/gid 101.
| 2021-01-20T21:02:06 | 2021-01-20T21:02:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,967,572 | 19,273 | CurtizJ | Now subcolumns are supported only while reading from several table engines. See `IStorage::supportsSubcolumns`. | 2021-01-20T21:42:24 | 2021-01-20T21:42:24 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 764,349,459 | 18,684 | zhang2014 | > all test_materialize_mysql_database are still failing. @zhang2014 perhaps you can help...
fixes in https://github.com/ClickHouse/ClickHouse/pull/19344 | 2021-01-21T05:12:27 | 2021-01-21T05:12:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 764,436,728 | 3,141 | alexey-milovidov | Yes, 100% of reproduced cases of this issue has been fixed.
If you encounter similar issue on newer versions, please let us know. | 2021-01-21T07:18:06 | 2021-01-21T07:18:06 | {} | MEMBER |
ClickHouse | ClickHouse | 764,556,990 | 18,945 | qoega | 01079_parallel_alter_add_drop_column_zookeeper
https://clickhouse-test-reports.s3.yandex.net/19346/5f3059555a039181cdb9b558fb073fc7998ad496/functional_stateless_tests_(release).html#fail1 | 2021-01-21T11:03:19 | 2021-01-21T11:03:19 | {} | MEMBER |
ClickHouse | ClickHouse | 764,588,542 | 19,348 | alexey-milovidov | @amosbird Maybe you will be interested? | 2021-01-21T11:46:11 | 2021-01-21T11:46:11 | {} | MEMBER |
ClickHouse | ClickHouse | 764,601,370 | 19,353 | akuzm | ```
*2* :) INSERT INTO t0(c0, c1) VALUES ('1',1) ;
Ok.
```
It's a typo -- string literal use single quotes. | 2021-01-21T12:11:28 | 2021-01-21T12:11:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,550,332 | 19,270 | alesapin | `test_multiple_disks/test.py::test_start_stop_moves` flaky in this branch. | 2021-01-20T11:48:50 | 2021-01-20T11:48:50 | {} | MEMBER |
ClickHouse | ClickHouse | 763,946,937 | 11,003 | rrati | That won't work on openshift without creating an scc, which isn't a good practice.
From my previous message:
When openshift runs a container, it assigns a random UID to the user that is running the container.
Setting a UID explicitly requires elevated privs which are not available by default on openshift. | 2021-01-20T21:07:29 | 2021-01-20T21:07:29 | {} | NONE |
ClickHouse | ClickHouse | 763,974,564 | 11,003 | filimonov | Ok. Will fix that tomorrow. It seems only single chmod should be added to dockerfile. Other parts are already there. | 2021-01-20T21:55:22 | 2021-01-20T21:55:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 764,316,611 | 17,886 | spz0507 | @alexey-milovidov thanks a lot. | 2021-01-21T04:55:46 | 2021-01-21T04:55:46 | {} | NONE |
ClickHouse | ClickHouse | 764,489,570 | 18,945 | qoega | test_server_initialization flapped
https://clickhouse-test-reports.s3.yandex.net/0/7b4d1fba6aeeed0a395dd135ef662883f84a0855/integration_tests_(memory).html | 2021-01-21T09:05:51 | 2021-01-21T09:05:51 | {} | MEMBER |
ClickHouse | ClickHouse | 764,508,489 | 19,352 | alexey-milovidov | There is no technical reasons to not support transactions for single large INSERT into ReplicatedMergeTree table. It is in our roadmap and @tavplubix will implement it.
We have plans to do the following:
- large INSERT in single transaction;
- transactional INSERT into table and all dependent materialized views;
... | 2021-01-21T09:38:12 | 2021-01-21T09:38:12 | {
"+1": 16
} | MEMBER |
ClickHouse | ClickHouse | 764,519,167 | 14,100 | buyology | We see this intermittently on 20.11.3.3. One example query that triggered it
```sql
SELECT x1 AS x11, x2 AS x22, x3 AS x33, x4 AS x44, x5 AS x55, x6 AS x66, x7 AS x77, x8 AS x88, x9 AS x99
FROM view
WHERE x2 IN CAST(['value'], 'Array(String)'))
``` | 2021-01-21T09:57:08 | 2021-01-21T10:07:54 | {} | NONE |
ClickHouse | ClickHouse | 763,626,600 | 19,312 | alexey-milovidov | It was fixed yesterday. | 2021-01-20T14:04:30 | 2021-01-20T14:04:30 | {} | MEMBER |
ClickHouse | ClickHouse | 763,903,883 | 19,245 | qoega | Why we backport not-for-changelog PR? | 2021-01-20T20:13:03 | 2021-01-20T20:13:03 | {} | MEMBER |
ClickHouse | ClickHouse | 763,904,324 | 19,245 | qoega | > Why we backport not-for-changelog PR?
Found must-backport flag. OK | 2021-01-20T20:13:52 | 2021-01-20T20:13:52 | {} | MEMBER |
ClickHouse | ClickHouse | 764,532,417 | 19,217 | alexey-milovidov | @l1t1 It works as expected. Trailing zeros are not displayed. | 2021-01-21T10:19:49 | 2021-01-21T10:20:10 | {} | MEMBER |
ClickHouse | ClickHouse | 764,587,697 | 19,349 | qoega | > Also I use the client to execute the sql, my understanding client will automatically add the limit, so the output should not be the reason for the performance gap.
You can add limit manually and check that this is not an issue.
Additionaly there is additional information in query log.
https://clickhouse.tech/... | 2021-01-21T11:44:16 | 2021-01-21T11:44:16 | {} | MEMBER |
ClickHouse | ClickHouse | 764,195,276 | 19,342 | weeway | errlog:
<img width="1423" alt="image" src="https://user-images.githubusercontent.com/12637868/105273147-b18e8080-5bd5-11eb-80a9-65ba2ded8a9a.png">
| 2021-01-21T02:45:00 | 2021-01-21T02:45:00 | {} | NONE |
ClickHouse | ClickHouse | 764,433,156 | 3,141 | elloooooo | Did this issue fixed? | 2021-01-21T07:09:21 | 2021-01-21T07:09:21 | {} | NONE |
ClickHouse | ClickHouse | 764,571,419 | 14,100 | buyology | The problem seems to still exist on 21.1.2:
```sql
Connected to ClickHouse server version 21.1.2 revision 54443.
8257a695ab1a :) CREATE TABLE t ( a String ) ENGINE = Memory()
Ok.
0 rows in set. Elapsed: 0.007 sec.
8257a695ab1a :) CREATE VIEW t_v AS SELECT * FROM t
Ok.
0 rows in set. Elapsed: 0.007... | 2021-01-21T11:21:30 | 2021-01-21T11:21:30 | {} | NONE |
ClickHouse | ClickHouse | 764,602,138 | 19,355 | KochetovNicolai | Discussed in chat.
Race was in callback which used already removed response object.
Now response removes it's own callback after finish. | 2021-01-21T12:12:53 | 2021-01-21T12:12:53 | {} | MEMBER |
ClickHouse | ClickHouse | 763,952,236 | 11,003 | filimonov | I mean smth like
```yml
apiVersion: v1
kind: Pod
metadata:
name: ch
spec:
securityContext:
runAsUser: 101
runAsGroup: 101
fsGroup: 101
containers:
- name: ch-cont
image: yandex/clickhouse-server:v21.1.2.15-stable # on older versions you may need to mount in the entrypo... | 2021-01-20T21:17:11 | 2021-01-20T21:17:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 764,218,965 | 19,273 | sundy-li | > Now subcolumns are supported only while reading from several table engines. See `IStorage::supportsSubcolumns`.
Ok, and this not works in where query.
```
:) SELECT m.keys FROM t_map;
┌─m.keys────┐
│ ['a','b'] │
│ ['a','c'] │
│ ['b','c'] │
└───────────┘
3 rows in set. Elapsed: 0.055 sec.
:) select... | 2021-01-21T03:54:16 | 2021-01-21T03:54:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 764,407,651 | 19,272 | slimtom95 | by the way I cannot reproduce randomness on 20.4.4.18 linux build, but it shows on mac with 20.13 | 2021-01-21T06:15:00 | 2021-01-21T06:15:00 | {} | NONE |
ClickHouse | ClickHouse | 763,720,122 | 18,945 | qoega | `test_live_view_over_distributed/test.py::TestLiveViewOverDistributedSuite::test_watch_live_view_order_by_node[lv_over_distributed_table-node0]` timeout in test with sanitizer
https://clickhouse-test-reports.s3.yandex.net/0/a15092eeb77b7d35d8f068f99bfeb26b0470ae9d/integration_tests_(memory).html | 2021-01-20T15:41:14 | 2021-01-20T15:41:14 | {} | MEMBER |
ClickHouse | ClickHouse | 763,724,589 | 18,945 | qoega | 01149_zookeeper_mutation_stuck_after_replace_partition (release, Wide Parts Enabled) https://clickhouse-test-reports.s3.yandex.net/0/f76846b7c5c7804c10d22616593f54ae09641e77/functional_stateless_tests_(release,_wide_parts_enabled).html | 2021-01-20T15:47:41 | 2021-01-20T15:47:41 | {} | MEMBER |
ClickHouse | ClickHouse | 763,999,843 | 11,003 | haarchri | this is exact at the moment the problem:
`I did a quick test with a dockerfile and it seems that the solutions I provided won't work. The container is checking that the owner of /var/lib/clickhouse` | 2021-01-20T22:40:07 | 2021-01-20T22:40:07 | {} | NONE |
ClickHouse | ClickHouse | 764,511,608 | 19,345 | alexey-milovidov | It's not calculated for *compact* parts. The reason is that multiple columns are compressed as a whole and you cannot tell where compressed data of one column ends and another column starts.
Compact parts are used for small amount of data. | 2021-01-21T09:43:49 | 2021-01-21T09:44:37 | {} | MEMBER |
ClickHouse | ClickHouse | 763,158,170 | 19,229 | filimonov | Maybe you 28 threads reached the maximum possible IO speed? So your disk can not deliver data fast enough to fully load > 30 threads?
Actually even ram has it's maximum throughput, so it's also possible that you reach the maximum possible speed of memcpy. | 2021-01-19T21:39:00 | 2021-01-20T11:28:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,566,469 | 19,310 | sunny19930321 | LGTM | 2021-01-20T12:18:12 | 2021-01-20T12:18:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,738,714 | 19,273 | sundy-li | I am integrating it in system tables, but tests failed in
```
:) select m, m.keys from t_map;
SELECT
m,
m.keys
FROM t_map
Query id: f1e2764a-2b70-437f-8123-cd80848d5e24
┌─m─────────────┬─m.keys────┐
│ {'a':1,'b':2} │ ['a','b'] │
│ {'a':3,'c':4} │ ['a','c'] │
│ {'b':5,'c':6} │ ['b','c'] │
└... | 2021-01-20T16:07:36 | 2021-01-20T16:13:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,805,967 | 19,298 | akuzm | Fuzzer https://github.com/ClickHouse/ClickHouse/issues/19330
Integration tests are not affected so we don't have to wait for them. | 2021-01-20T17:22:36 | 2021-01-20T17:22:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 763,866,366 | 19,091 | tavplubix | AST fuzzer (UBSan) - #19336
Functional stateful tests (debug) - #19335 | 2021-01-20T19:05:20 | 2021-01-20T19:05:20 | {} | MEMBER |
ClickHouse | ClickHouse | 764,539,798 | 19,343 | qoega | AST Fuzzer fail is unrelated #19329 | 2021-01-21T10:32:22 | 2021-01-21T10:32:22 | {} | MEMBER |
ClickHouse | ClickHouse | 764,628,358 | 19,356 | alexey-milovidov | It works perfectly:
```
CREATE TABLE tutorial.visits_v1
(
`CounterID` UInt32,
`StartDate` Date,
`Sign` Int8,
`IsNew` UInt8,
`VisitID` UInt64,
`UserID` UInt64,
`StartTime` DateTime,
`Duration` UInt32,
`UTCStartTime` DateTime,
`PageViews` Int32,
`Hits` Int32,
... | 2021-01-21T13:03:11 | 2021-01-21T13:03:11 | {} | MEMBER |
ClickHouse | ClickHouse | 754,537,724 | 18,761 | xinge-1 | 全量同步没有问题，但是当源数据库MySQL有增量数据写入，错误日志就会出现上面错误。 | 2021-01-05T10:04:47 | 2021-01-05T10:04:47 | {} | NONE |
ClickHouse | ClickHouse | 754,957,435 | 15,173 | alexey-milovidov | Does it reproduce on recent versions (20.12 or newer)? | 2021-01-05T23:12:11 | 2021-01-05T23:12:11 | {} | MEMBER |
ClickHouse | ClickHouse | 755,332,749 | 18,783 | den-crane | ah, OK.
CH does not support JOINs for distributed tables in circle-replication.
https://github.com/ClickHouse/ClickHouse/issues/6571
https://github.com/ClickHouse/ClickHouse/issues/3196
The best option for you to do not use circle-replication. | 2021-01-06T14:33:07 | 2021-01-06T14:33:07 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.