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 | 745,389,915 | 18,105 | abyss7 | Ok! | 2020-12-15T16:05:06 | 2020-12-15T16:05:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,362,322 | 18,100 | den-crane | Please provide create table at least.
What CPU model do you use?
Do you have columns with CODECs ? | 2020-12-15T15:23:00 | 2020-12-15T16:20:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,508,341 | 18,099 | cetver | Thanks for the explanation.
I still think that this is a bug, or the behavior is not specified in the documentation.
Max "true" int is `Int64`, if the value between -9223372036854775808 : 9223372036854775807 extract functions works as expected
Сonclusion: keep int, float values as string | 2020-12-15T19:17:48 | 2020-12-15T19:19:05 | {} | NONE |
ClickHouse | ClickHouse | 744,368,516 | 18,032 | alexey-milovidov | Discussion:
\- all functions cannot work for Nullable types, as the original arraySum was not working;
\- all functions cannot work for big ints, as the original arraySum was not working;
This will be good to improve in subsequent PRs.
\- arrayMin/Max cannot work for Strings and other types while arrayReduce work... | 2020-12-14T11:10:53 | 2020-12-14T11:11:03 | {} | MEMBER |
ClickHouse | ClickHouse | 744,626,749 | 18,067 | gonzalezjo | Apparently #18066 was made in the period between me seeing the issue and checking GitHub, and me actually making this PR. It is currently redundant given 18066, so I'll be closing this.
| 2020-12-14T18:29:15 | 2020-12-14T18:29:15 | {} | NONE |
ClickHouse | ClickHouse | 745,202,530 | 18,007 | alesapin | Close in favor of https://github.com/ClickHouse/ClickHouse/pull/18007. | 2020-12-15T10:36:03 | 2020-12-15T10:36:03 | {} | MEMBER |
ClickHouse | ClickHouse | 745,293,819 | 18,093 | Aureliano-Berlindia | Hi, Here is the latest clickhouse-server.log:
not sure what's the problem
2020.12.15 21:32:48.232592 [ 6420 ] {} <Trace> TCPHandlerFactory: TCP Request. Address: [::1]:40420
2020.12.15 21:32:48.232669 [ 6420 ] {} <Debug> TCPHandler: Connected ClickHouse client version 20.8.0, revision: 54438, user: default.
202... | 2020-12-15T13:42:16 | 2020-12-15T13:42:16 | {} | NONE |
ClickHouse | ClickHouse | 745,526,148 | 17,666 | TigProg | ServOps changed the hypervisor setting from `<cpu mode='host-model'><model fallback='allow'/></cpu>` to `<cpu mode='host-passthrough'/>` and the problem stopped reproducing.
This is a bit strange, because on another cluster (we recently added second CH cluster) the problem is not reproducible with the previous setting... | 2020-12-15T19:49:27 | 2020-12-15T19:49:27 | {} | NONE |
ClickHouse | ClickHouse | 745,600,380 | 17,630 | alexey-milovidov | Reproduced. | 2020-12-15T22:12:13 | 2020-12-15T22:12:13 | {} | MEMBER |
ClickHouse | ClickHouse | 745,689,869 | 17,463 | alexey-milovidov | ```
$ readelf -s -V build/programs/clickhouse | grep -F 2.16
5: 0000000000000000 0 FUNC GLOBAL DEFAULT UND aligned_alloc@GLIBC_2.16 (3)
004: 1 (*global*) 3 (GLIBC_2.16) 2 (GLIBC_2.2.5) 2 (GLIBC_2.2.5)
0x0100: Name: GLIBC_2.16 Flags: none Version: 3
``` | 2020-12-16T00:45:42 | 2020-12-16T00:45:42 | {} | MEMBER |
ClickHouse | ClickHouse | 744,683,547 | 9,609 | alexey-milovidov | So, the timeout is set to `\10\7\0\0\0\0\0\0\0\0\0\0\0\0\0\0` (in struct timeval).
```
\10\7\0\0\0\0\0\0\0\0\0\0\0\0\0\0
^ ^ ^ ^ ^ ^ ^ ^
```
This is the number of seconds in little endian with octal escaping: 8 + 7 * 256 = 1800 seconds.
It looks enough.
Maybe connection timeout is too low? | 2020-12-14T20:14:22 | 2020-12-14T20:14:22 | {} | MEMBER |
ClickHouse | ClickHouse | 745,504,562 | 17,726 | alexey-milovidov | It means that the `normalized_query_hash` is not always initialized.
We can initialize it to zero in the struct... but we also have to check why it's not always set? | 2020-12-15T19:11:05 | 2020-12-15T19:11:05 | {} | MEMBER |
ClickHouse | ClickHouse | 745,618,390 | 17,726 | alexey-milovidov | @nikitamikhaylov What do you think about this tsan failure?
https://clickhouse-test-reports.s3.yandex.net/17726/ab16a0bc533f410efed3cc026b7610c7e23870b9/stress_test_(thread)/stderr.log | 2020-12-15T22:55:10 | 2020-12-15T22:55:10 | {} | MEMBER |
ClickHouse | ClickHouse | 745,819,618 | 7,094 | vivarum | @alexey-milovidov, it looks like the task is a bit more tricky than just to make a matcher that matches any single event except one (i.e. a simple event negation can be added with existing matchers, just by adding a new event with inverted condition).
In my understanding, the new matcher, i.e. (!?123) should match a... | 2020-12-16T07:22:59 | 2020-12-16T07:23:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 744,592,709 | 17,904 | spongedu | Why not support ALTER TTL on tables created with old syntax ? | 2020-12-14T17:29:31 | 2020-12-14T17:29:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,092,631 | 18,056 | zhangjmruc | @alexey-milovidov I have done with code changes according to your two comments and also handled the integration failure related to "SHOW SETTINGS PROFILES". Please review it again and let me know your opinions. | 2020-12-15T06:51:14 | 2020-12-15T06:51:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,110,557 | 18,007 | alesapin | `test_log_family_s3/test.py::test_log_family_s3[StripeLog-1-2]` -- flaky test | 2020-12-15T07:32:47 | 2020-12-15T07:32:47 | {} | MEMBER |
ClickHouse | ClickHouse | 745,342,255 | 17,160 | vitlibar | The current settings can be shown also with the function `getSetting`, for example
```
SELECT getSetting('send_timeout')
```
Although it works only for singular settings because `getSetting()` requires its parameter to be a constant string. | 2020-12-15T14:50:35 | 2020-12-15T15:12:43 | {} | MEMBER |
ClickHouse | ClickHouse | 744,368,831 | 17,849 | alexey-milovidov | Implemented by @kitaisreal. | 2020-12-14T11:11:33 | 2020-12-14T11:11:33 | {} | MEMBER |
ClickHouse | ClickHouse | 744,542,579 | 17,641 | nikitamikhaylov | Yandex synchronization check failed, because I didn't patch Poco in Arcadia :(( | 2020-12-14T16:09:14 | 2020-12-14T16:09:14 | {} | MEMBER |
ClickHouse | ClickHouse | 744,750,999 | 17,884 | l1t1 | > Example here: https://clickhouse.tech/docs/en/operations/performance-test/
cannot find the link mentioned. could you post a picture about this?
Click on the first green check mark or red cross with green “ClickHouse Build Check” and click on the “Details” link near “ClickHouse Build Check”. There is no such link... | 2020-12-14T22:29:16 | 2020-12-14T22:29:16 | {} | NONE |
ClickHouse | ClickHouse | 745,116,991 | 17,656 | bharatnc | @tavplubix and @azat
I have implemented the following as per your review:
- [X] entry String (or entry_name) instead of name
- [X] cluster String - cluster name, we can extract it from query, but it's easier to use where cluster='smth' in select
- [X] host String (or maybe two separate columns with host and p... | 2020-12-15T07:49:12 | 2020-12-15T08:08:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,615,641 | 17,726 | alexey-milovidov | Ok. | 2020-12-15T22:47:54 | 2020-12-15T22:47:54 | {} | MEMBER |
ClickHouse | ClickHouse | 745,730,634 | 17,630 | alexey-milovidov | Not fixed? | 2020-12-16T02:56:58 | 2020-12-16T02:56:58 | {} | MEMBER |
ClickHouse | ClickHouse | 745,862,260 | 18,105 | alesapin | Oh, nice. It works. | 2020-12-16T07:54:55 | 2020-12-16T07:54:55 | {} | MEMBER |
ClickHouse | ClickHouse | 745,879,164 | 18,096 | knight11111 | Hello，This is the result of the order。I didn't see anything unusual， I've turned it on.but use sudo systemctl status clickhouse-server ， I see failure。
[clickhouse-server.err.log](https://github.com/ClickH... | 2020-12-16T08:07:16 | 2020-12-16T08:07:16 | {} | NONE |
ClickHouse | ClickHouse | 744,403,583 | 16,613 | alexey-milovidov | @nikitamikhaylov Bad changelog. | 2020-12-14T12:21:01 | 2020-12-14T12:21:01 | {} | MEMBER |
ClickHouse | ClickHouse | 744,579,521 | 18,058 | akuzm | > Then you could simply close this one :)
Ooops x2, just noticed it wasn't identical and now we are disabling it twice :smile: | 2020-12-14T17:07:36 | 2020-12-14T17:07:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 744,999,501 | 18,061 | wang2222qq | OK, the error log information is as follows:
![image](https://user-images.githubusercontent.com... | 2020-12-15T01:54:44 | 2020-12-15T01:54:44 | {} | NONE |
ClickHouse | ClickHouse | 745,159,511 | 16,954 | pod2metra | Hey Guys. What about this PR? This feature looks great and we waiting for it. When do you plan to merge it? | 2020-12-15T09:17:11 | 2020-12-15T09:17:11 | {} | NONE |
ClickHouse | ClickHouse | 745,582,418 | 18,116 | alexey-milovidov | It's only possible if WITH TOTALS is in the direct subquery in FROM. | 2020-12-15T21:35:45 | 2020-12-15T21:35:45 | {} | MEMBER |
ClickHouse | ClickHouse | 745,722,960 | 17,865 | jingwangfei | I ran into the same problem.
When i remove '1 = 1' condition , it can run.
confused~~~ | 2020-12-16T02:29:40 | 2020-12-16T02:29:40 | {} | NONE |
ClickHouse | ClickHouse | 745,737,462 | 18,123 | taiyang-li | I found this comment in source file: `src/Databases/DatabaseAtomic.h`
```
/// All tables in DatabaseAtomic have persistent UUID and store data in
/// /clickhouse_path/store/xxx/xxxyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyyy/
/// where xxxyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyyy is UUID of the table.
/// RENAMEs are performed witho... | 2020-12-16T03:19:28 | 2020-12-16T03:19:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,765,479 | 17,215 | coreyralph | Closing this because I'm no longer using odbc. My workaround is copying those tables into tsv files and having an external dictionary read those. | 2020-12-16T05:02:21 | 2020-12-16T05:02:21 | {} | NONE |
ClickHouse | ClickHouse | 744,795,457 | 11,617 | nikitamikhaylov | ```bash
2020-12-15 02:21:19 b"../src/IO/WriteIntText.h:27:9: runtime error: applying non-zero offset 40 to null pointer\n
#0 0xf1d576c in void DB::writeIntText<char8_t>(char8_t, DB::WriteBuffer&) /build/obj-x86_64-linux-gnu/../src/IO/WriteIntText.h:27:9\n
#1 0x1a496b8d in DB::IRowOutputFormat::write(std::__1... | 2020-12-14T23:26:11 | 2020-12-14T23:26:11 | {} | MEMBER |
ClickHouse | ClickHouse | 745,005,903 | 17,680 | den-crane | fixed | 2020-12-15T02:12:42 | 2020-12-15T02:12:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,386,839 | 18,105 | alesapin | @abyss7 Just merge this into your branch. I don't want to fix all images here :) | 2020-12-15T16:00:16 | 2020-12-15T16:00:16 | {} | MEMBER |
ClickHouse | ClickHouse | 745,503,467 | 17,726 | alexey-milovidov | @amosbird But MSan failure looks related:
https://clickhouse-test-reports.s3.yandex.net/17726/b3e8067700a361f51351ecb620f17c2ed85ec7cb/stress_test_(memory)/stderr.log
The mechanics:
- uninitialized memory is written to QueryThreadLog;
- checksum of the data part is calculated over uninitialized memory;
- it's ... | 2020-12-15T19:09:05 | 2020-12-15T19:09:05 | {} | MEMBER |
ClickHouse | ClickHouse | 745,302,723 | 10,455 | mrAndersen | Seems like a have the same issue. Consider Following:
1) Create MergeTree table on 4 shards - *track_shard*
2) Create Distributed table on master server - *track_main*
3) Create MergeTree table on master server - *track_old*
track_old has ~ approx 5 Bil rows, i am trying to load the data to track_main and respe... | 2020-12-15T13:59:12 | 2020-12-15T19:19:21 | {} | NONE |
ClickHouse | ClickHouse | 745,676,970 | 17,463 | alexey-milovidov | It is reproduced very easily:
```
$ docker run --rm -it --volume=/home/milovidov/work/ClickHouse/build/programs/clickhouse:/clickhouse centos:5 /clickhouse
/clickhouse: /lib64/libc.so.6: version `GLIBC_2.16' not found (required by /clickhouse)
``` | 2020-12-16T00:25:21 | 2020-12-16T00:25:21 | {} | MEMBER |
ClickHouse | ClickHouse | 744,366,340 | 17,795 | alexey-milovidov | It requires to store additional info in serialized column.
It cannot be implemented without additional changes in design.
Good news that these additional changes are already planned.
Hold before implementation of "move serialization/deserialization from IDataType to IColumn". | 2020-12-14T11:07:01 | 2020-12-14T11:08:06 | {} | MEMBER |
ClickHouse | ClickHouse | 744,517,994 | 18,064 | KochetovNicolai | cc. @akuzm can you check this changes too? | 2020-12-14T15:31:05 | 2020-12-14T15:31:05 | {} | MEMBER |
ClickHouse | ClickHouse | 745,216,965 | 18,086 | mis98zb | I read some related code in master branch.
I guess I found the reason.
In DataTypes/convertMySQLDataType.cpp
```
// we expect mysql_data_type to be either "basic_type" or "type_with_params(param1, param2, ...)"
auto data_type = std::string_view(mysql_data_type);
const auto param_start_pos = data_typ... | 2020-12-15T11:03:27 | 2020-12-15T11:03:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,361,768 | 18,100 | den-crane | How I can reproduce it? | 2020-12-15T15:22:07 | 2020-12-15T15:22:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,730,100 | 17,463 | alexey-milovidov | @amosbird Apologies. I was build it after using TSan build (for #17630) and jemalloc was disabled.
Now I cannot reproduce the issue either. | 2020-12-16T02:54:59 | 2020-12-16T02:54:59 | {} | MEMBER |
ClickHouse | ClickHouse | 745,784,831 | 16,985 | Zhile | > @Zhile please install a package from https://clickhouse-builds.s3.yandex.net/16985/e78010c28cd0152f33ece50ec647b39b1100a3c9/clickhouse_build_check/report.html
> (links in the first line of the table), and test / ensure if that solves the issue.
Confirmed. Fixed. | 2020-12-16T06:00:10 | 2020-12-16T06:00:10 | {} | NONE |
ClickHouse | ClickHouse | 744,533,638 | 17,738 | filimonov | @alexey-milovidov not too much of original changeset survived. | 2020-12-14T15:55:40 | 2020-12-14T15:55:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 744,565,216 | 17,738 | alexey-milovidov | Thank you! Now it's perfect. | 2020-12-14T16:44:24 | 2020-12-14T16:44:24 | {} | MEMBER |
ClickHouse | ClickHouse | 744,245,293 | 17,910 | adevyatova | Ready for merge @alesapin | 2020-12-14T07:46:57 | 2020-12-15T07:22:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,148,640 | 17,639 | Akazz | There is something wrong with `test_ttl_replicated::test_ttl_empty_parts`. And that is due to changes in `MergeTreeData::getPartsCount()`. | 2020-12-15T08:56:42 | 2020-12-15T08:56:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,298,881 | 16,954 | KochetovNicolai | Yes, all tests passed. Merging | 2020-12-15T13:51:51 | 2020-12-15T13:51:51 | {} | MEMBER |
ClickHouse | ClickHouse | 745,453,741 | 18,096 | den-crane | `sudo journalctl -u clickhouse-server `
`sudo cat /var/log/clickhouse-server/stderr.log`
` sudo systemctl status clickhouse-server` | 2020-12-15T17:44:33 | 2020-12-15T17:44:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,518,334 | 17,726 | amosbird | > but we also have to check why it's not always set?
Because we don't always have thread_group...
https://github.com/ClickHouse/ClickHouse/blob/b0d27d6747b32b8edb0c869e14a34615661761c6/src/Interpreters/ThreadStatusExt.cpp#L421 | 2020-12-15T19:35:04 | 2020-12-15T19:35:04 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 745,526,224 | 18,061 | tavplubix | Could you please copy the whole CREATE query and the whole error message with stacktrace from the server log and post it here as a text? | 2020-12-15T19:49:35 | 2020-12-15T19:49:35 | {} | MEMBER |
ClickHouse | ClickHouse | 745,616,220 | 18,118 | den-crane | >What are the recommended ways of implementing ALTER DELETE mutations on Replicated*
>tables that have WHERE clauses depending on other tables?
We don't recommend such WHERE clauses.
But still:
* you can use Replicated table -- records_to_delete
* you can enable allow_nondeterministic_mutations by `set a... | 2020-12-15T22:49:28 | 2020-12-15T22:49:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,738,177 | 18,122 | taiyang-li | `ALTER DELETE` need extra disk space to reserve intermediate result. Maybe you can consider `ALTER DROP PARTITION`, which drop a partition each time. | 2020-12-16T03:21:58 | 2020-12-16T03:21:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,885,957 | 18,122 | gj-zhang | > `ALTER DELETE` need extra disk space to reserve intermediate result. Maybe you can consider `ALTER DROP PARTITION`, which drop a partition each time.
Is that the only way? | 2020-12-16T08:11:59 | 2020-12-16T08:11:59 | {} | NONE |
ClickHouse | ClickHouse | 745,353,111 | 17,983 | alexey-milovidov | Merged prematurally. | 2020-12-15T15:08:00 | 2020-12-15T15:08:00 | {} | MEMBER |
ClickHouse | ClickHouse | 745,358,367 | 18,104 | vitlibar | The second attempt to merge after https://github.com/ClickHouse/ClickHouse/pull/17983 | 2020-12-15T15:16:33 | 2020-12-15T15:16:33 | {} | MEMBER |
ClickHouse | ClickHouse | 745,500,825 | 18,056 | alexey-milovidov | There is a failure with integration tests, looks related to the changes.
Could you please also merge/rebase with master to fix timeouts of functional tests? | 2020-12-15T19:04:33 | 2020-12-15T19:04:33 | {} | MEMBER |
ClickHouse | ClickHouse | 745,537,570 | 18,098 | alesapin | As far as I remember AST Fuzzer already reported. All tests not required for merge. | 2020-12-15T20:09:05 | 2020-12-15T20:09:14 | {} | MEMBER |
ClickHouse | ClickHouse | 745,565,647 | 17,666 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/commit/eed767bbab573f6f72d69f93ddc2c577539bfab2 | 2020-12-15T21:03:12 | 2020-12-15T21:03:12 | {} | MEMBER |
ClickHouse | ClickHouse | 745,594,234 | 18,117 | alexey-milovidov | References:
https://github.com/intel/QATzip
https://github.com/ceph/ceph/pull/19714
Do you have some numbers for comparison?
It's unclear - 30% boost on hadoop is in what scenario? | 2020-12-15T21:59:45 | 2020-12-15T22:00:40 | {} | MEMBER |
ClickHouse | ClickHouse | 745,714,148 | 18,003 | taylor131415 | > Please share log fragments. Nodes don't go readonly when you shutdown another node. Do you also run zookeeper on the same nodes?
At first, zookeeper and clickhouse running on the same node. After I found this issue, I change to run all zookeeper on a separate node.(Without any clickhouse process on this node), but... | 2020-12-16T02:01:50 | 2020-12-16T02:01:50 | {} | NONE |
ClickHouse | ClickHouse | 745,749,516 | 18,095 | spongedu | BTW, I found that for now we don't have concepts like`reserved keyword` in [MySQL](https://dev.mysql.com/doc/refman/5.7/en/keywords.html), so we can execute sql like `CREATE VIEW CREATE AS SELECT * FROM ..` which are illegal in other dbms | 2020-12-16T04:03:53 | 2020-12-16T04:03:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,887,033 | 18,095 | spongedu | @alexey-milovidov PTAL~ thx. | 2020-12-16T08:12:50 | 2020-12-16T08:12:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 744,611,178 | 17,848 | azat | >@nikitamikhaylov nikitamikhaylov self-assigned this 5 days ago
@nikitamikhaylov any news? | 2020-12-14T18:02:06 | 2020-12-14T18:02:06 | {} | MEMBER |
ClickHouse | ClickHouse | 745,118,528 | 18,091 | lenovore | I found that clickhouse has no time data type, maybe MaterializeMySQL can convert the time data type to string | 2020-12-15T07:53:08 | 2020-12-15T07:53:08 | {} | NONE |
ClickHouse | ClickHouse | 745,261,137 | 17,905 | akuzm | Turned out the workaround doesn't work, backported to 20.8 manually. | 2020-12-15T12:36:29 | 2020-12-15T12:36:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,538,889 | 18,065 | alesapin | Wow, the error stress tsan is stable. | 2020-12-15T20:11:44 | 2020-12-15T20:11:44 | {} | MEMBER |
ClickHouse | ClickHouse | 745,719,786 | 17,463 | amosbird | > It is reproduced very easily:
>
> ```
> $ docker run --rm -it --volume=/home/milovidov/work/ClickHouse/build/programs/clickhouse:/clickhouse centos:5 /clickhouse
> /clickhouse: /lib64/libc.so.6: version `GLIBC_2.16' not found (required by /clickhouse)
> ```
... What build variant did you use and what toolch... | 2020-12-16T02:19:43 | 2020-12-16T02:19:43 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 745,747,002 | 17,857 | fastio | > @fastio The only remark - please do not reformat the code here. Git blame will become inconsistent. And these changes are not relevant to the subject of this PR. If you want to reformat code somehow, please, open another PR.
Thanks, I got it. | 2020-12-16T03:53:57 | 2020-12-16T03:53:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,761,085 | 17,977 | azat | @alesapin @alexey-milovidov do you mind if I will try to increase rss limit (`ccache prlimit --rss=10000000 --cpu=600`) ? (or just `-DENABLE_CHECK_HEAVY_BUILDS=0`) | 2020-12-16T04:46:37 | 2020-12-16T04:47:59 | {} | MEMBER |
ClickHouse | ClickHouse | 744,577,907 | 18,058 | azat | Then you could simply close this one :) | 2020-12-14T17:04:53 | 2020-12-14T17:04:53 | {} | MEMBER |
ClickHouse | ClickHouse | 744,700,307 | 17,951 | alexey-milovidov | @filimonov These are new features, they will be available in the first release, that will become prestable after PRs will be merged. | 2020-12-14T20:45:56 | 2020-12-14T20:45:56 | {} | MEMBER |
ClickHouse | ClickHouse | 745,432,962 | 17,905 | akuzm | Double-checked the backported fix under debugger on 20.8, seems to work OK.
```
Thread 51 "clickhouse-serv" hit Breakpoint 1, Poco::RotateBySizeStrategy::mustRotate (this=0x7ffff7016120, pFile=0x7ffff7012300) at /home/akuzm/ch1/ch/contrib/poco/Foundation/src/RotateStrategy.cpp:108
108 {
(gdb) info locals
No loca... | 2020-12-15T17:10:32 | 2020-12-15T17:11:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,487,000 | 16,995 | azat | BTW I can't merge anyway, so you need to ask someone who can, @alexey-milovidov maybe you can take a look? (or point the right person) | 2020-12-15T18:43:07 | 2020-12-15T18:43:07 | {
"heart": 1
} | MEMBER |
ClickHouse | ClickHouse | 745,599,251 | 17,630 | alexey-milovidov | Cannot reproduce TSan report on my machine... | 2020-12-15T22:09:51 | 2020-12-15T22:09:51 | {} | MEMBER |
ClickHouse | ClickHouse | 744,374,274 | 18,056 | alexey-milovidov | I would also like to have `SHOW CHANGED SETTINGS` query that will add `WHERE changed = 1` | 2020-12-14T11:21:50 | 2020-12-14T11:21:50 | {} | MEMBER |
ClickHouse | ClickHouse | 744,482,878 | 17,951 | filimonov | Would be great also to merge those 2 in the next release:
* https://github.com/ClickHouse/ClickHouse/pull/17846
* https://github.com/ClickHouse/ClickHouse/pull/17642
| 2020-12-14T14:37:42 | 2020-12-14T14:37:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 744,494,060 | 18,058 | akuzm | Oops, I just fixed it in master in a bytewise-identical way :) | 2020-12-14T14:54:43 | 2020-12-14T14:54:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,284,392 | 17,857 | nikitamikhaylov | @fastio The only remark - please do not reformat the code here. Git blame will become inconsistent. And these changes are not relevant to the subject of this PR. If you want to reformat code somehow, please, open another PR. | 2020-12-15T13:23:57 | 2020-12-15T13:24:36 | {} | MEMBER |
ClickHouse | ClickHouse | 745,467,819 | 18,099 | tavplubix | `JSON` format does not have `float64` data type, it has implementation-defined `Number` data type. The number `99999999999999999999` is not representable as `Number` in some implementations, so the first row is invalid and `JSONExtract` returns empty string.
```
:) SELECT isValidJSON(json), type FROM (SELECT '{"val... | 2020-12-15T18:08:31 | 2020-12-15T18:08:31 | {} | MEMBER |
ClickHouse | ClickHouse | 745,535,118 | 18,083 | tavplubix | Probably it's because `SELECT count(*) FROM task_score_log` is implicitly converted to `SELECT count(*) FROM task_score_log FINAL where _sign=1`. However, 4 seconds is too much even for SELECT FINAL. Could you please provide output of the following:
```
$ cat /var/lib/clickhouse/db_name/task_score_log.sql
```
and
... | 2020-12-15T20:04:31 | 2020-12-15T20:04:31 | {} | MEMBER |
ClickHouse | ClickHouse | 745,690,403 | 17,463 | alexey-milovidov | ```
$ /usr/local/bin/clang++ -Wl,--trace-symbol=aligned_alloc -fdiagnostics-color=always -std=gnu++2a -fsized-deallocation -gdwarf-aranges -pipe -msse4.1 -msse4.2 -mpopcnt -Wall -Wno-unused-command-line-argument -fdiagnostics-absolute-paths -fexperimental-new-pass-manager -Werror -Wextra -Wframe-larger-than=65536... | 2020-12-16T00:46:55 | 2020-12-16T00:46:55 | {} | MEMBER |
ClickHouse | ClickHouse | 744,437,693 | 18,061 | taiyang-li | Can you provide detail stack trace about this exception? You can find it in clickhouse error log. | 2020-12-14T13:24:42 | 2020-12-14T13:24:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 744,461,328 | 17,962 | allen12921 | close as no one can help | 2020-12-14T14:04:07 | 2020-12-14T14:04:07 | {} | NONE |
ClickHouse | ClickHouse | 745,308,680 | 13,390 | vitlibar | I think I'll implement the following syntax:
```
GRANT ALTER USER ON *.* TO `admin` WITH GRANT OPTION(ALL EXCEPT `admin`, `superuser`)
```
Without parentheses `GRANT OPTION` will mean `GRANT OPTION(ALL)`, i.e. user will be allowed to grant to anyone. | 2020-12-15T14:09:58 | 2020-12-15T14:09:58 | {} | MEMBER |
ClickHouse | ClickHouse | 745,731,770 | 9,523 | nicelulu | https://github.com/ClickHouse/ClickHouse/pull/14684 | 2020-12-16T03:01:00 | 2020-12-16T03:01:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,738,920 | 18,123 | Fullstop000 | @taiyang-li These comments are really helpful! Hope the doc can be updated soon. | 2020-12-16T03:24:29 | 2020-12-16T03:24:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 745,849,592 | 3,157 | Sammion | After execute dpkg-reconfigure tzdata and choose one time zone, it works! | 2020-12-16T07:45:31 | 2020-12-16T07:45:31 | {} | NONE |
ClickHouse | ClickHouse | 745,398,120 | 18,098 | alesapin | Now for some reason child process still not dead.
```
root 2350 0.0 0.0 4108 3380 ? S 16:41 0:00 | \_ /bin/bash /ClickHouse/docker/test/fasttest/run.sh
root 24478 12.0 10.5 64041152 13983316 ? Sl 17:06 15:37 | | \_ clickhouse-server --config-file=/fasttest-workspa... | 2020-12-15T16:17:39 | 2020-12-15T16:17:39 | {} | MEMBER |
ClickHouse | ClickHouse | 745,502,925 | 17,726 | alexey-milovidov | @amosbird Yes, it is. | 2020-12-15T19:08:03 | 2020-12-15T19:08:03 | {} | MEMBER |
ClickHouse | ClickHouse | 745,614,007 | 17,630 | alexey-milovidov | Fixed. | 2020-12-15T22:43:51 | 2020-12-15T22:43:51 | {} | MEMBER |
ClickHouse | ClickHouse | 744,573,428 | 18,011 | dpzmick | Indeed, this does seem to be the cause. I don't quite have an intuition for why compact vs wide would make such a large difference, some extremely naive "open lots of files" tests on this host don't show me numbers that seem to agree with the performance difference, but I'm satisfied now with these numbers so there's n... | 2020-12-14T16:58:02 | 2020-12-14T16:58:02 | {} | NONE |
ClickHouse | ClickHouse | 745,760,523 | 17,977 | azat | #### tsan
```
mv: error writing '/output/clickhouse-common-static_20.13.1.5425+tsan_amd64.deb': No space left on device
mv: error writing '/output/clickhouse-server_20.13.1.5425+tsan_all.deb': No space left on device
mv: error writing '/output/clickhouse-test_20.13.1.5425+tsan_all.deb': No space left on device
`... | 2020-12-16T04:44:35 | 2020-12-16T04:45:53 | {} | MEMBER |
ClickHouse | ClickHouse | 745,790,063 | 16,995 | sundy-li | > does not sounds user friendly, it should be described more from the user point of view, not the developer
Really thanks for your time :) , I've updated the changelog entry.
> Also I was thinking about one more case:
```
create table data (key Int, key_alias Int alias key) engine=MergeTree order by key;
sel... | 2020-12-16T06:13:30 | 2020-12-16T06:15:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 778,509,145 | 20,346 | den-crane | could be related https://github.com/ClickHouse/ClickHouse/issues/20456 | 2021-02-12T23:20:54 | 2021-02-12T23:20:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 778,599,522 | 20,465 | amosbird | Perhaps we should just ignore all errors in dtor as the same way the stdlib does it. If the user cares about errors during close, they must call close() themselves. Otherwise, the dtor calls close and ignores all errors from the call. | 2021-02-13T10:56:27 | 2021-02-13T10:56:27 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 778,644,165 | 20,462 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/pull/19524 does not handle SIGABRT deliberately.
Because there should not be the cases when we cannot print our own diagnostics. | 2021-02-13T16:50:00 | 2021-02-13T16:50:00 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.