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 | 821,460,445 | 10,966 | alexey-milovidov | @azat We've implemented it here: #22359
The old bug in zlib-ng was fixed. | 2021-04-16T18:47:20 | 2021-04-16T18:47:20 | {} | MEMBER |
ClickHouse | ClickHouse | 821,509,387 | 23,088 | akuzm | I looked at the old PR with experiments, didn't try 5s there: https://github.com/ClickHouse/ClickHouse/pull/11401
The graph for MemoryResident looks almost the same. The 6.5% may be not very reliable, it's the difference of medians of metrics from system.asynchronous_metric_log, it's weirdly noisy and I don't trust it... | 2021-04-16T19:15:30 | 2021-04-16T19:15:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,615,566 | 23,155 | mergify[bot] | **Command `rebase`: success**
> **Branch already up to date**
>
| 2021-04-15T17:46:36 | 2021-04-15T17:46:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,029,350 | 15,065 | Slach | Hello @Andr0901
any news about `Integration of S2 geometry library in ClickHouse`? | 2021-04-16T08:56:31 | 2021-04-16T08:56:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,099,191 | 23,057 | cadet354 | Спасибо за ответ,
я посмотрел комментарии, но там нигде не встречается мой случай, в моем случае всегда есть нужная цепочка (последние 3 шага), но алгоритм прохода раньше просто прерывается и не пытается дальнейшую реализацию.
По подсчету в просто windowFunnel тоже есть вопрос, т.к. число промежуточных шагов получа... | 2021-04-16T11:05:30 | 2021-04-16T11:05:30 | {} | NONE |
ClickHouse | ClickHouse | 821,126,149 | 10,441 | xjewer | >If old version created a table with CAST expressions in defaults we will read them from .sql and from the zookeeper. For new tables, we will store defaults without casts also in both places. It should work well.
in case if the table was created on the old version with the CAST, and adding a replica with a new versi... | 2021-04-16T12:02:18 | 2021-04-16T12:02:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,478,107 | 23,151 | ruct | Hi, this feature is currently in development and booked by me | 2021-04-15T14:37:41 | 2021-04-15T14:37:41 | {
"hooray": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,525,965 | 21,770 | Avogar | Oops, I didn't want to close it, it was a double misclick. Sorry, my concentration decreased towards evening :) | 2021-04-15T15:36:21 | 2021-04-15T15:37:45 | {} | MEMBER |
ClickHouse | ClickHouse | 820,530,857 | 21,101 | madm1keru | 21.4.3 revision 54447 is still afftected for dictGetOrDefault:
`dict_net_payable_usd Float64 DEFAULT round(net_payable / dictGetOrDefault('currencies', 'rate', (toDate(event_date), 'USD'), toFloat32(1)), 10)`
`<Error> Application: DB::Exception: Dictionary ('currencies') not found: While processing CAST(round((CA... | 2021-04-15T15:41:38 | 2021-04-15T15:49:40 | {} | NONE |
ClickHouse | ClickHouse | 820,599,236 | 21,101 | kitaisreal | @madm1keru dictGetFloat64, dictGetString. | 2021-04-15T17:20:27 | 2021-04-15T17:20:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,919,740 | 20,994 | onine007 | bitmapAndnot(bitmapBuild(arrayConcat([1111111111111],bitmapToArray(measure))),bitmapBuild([1111111111111])) can convert AggregateFunction(groupBitmap, UInt32) to AggregateFunction(groupBitmap, UInt64) | 2021-04-16T05:37:01 | 2021-04-16T05:37:01 | {} | NONE |
ClickHouse | ClickHouse | 821,138,454 | 23,160 | alexey-milovidov | BTW, we run all our production clusters with `trace` logging, it is by default and it is recommended for production usage (despite the common beliefs). | 2021-04-16T12:26:04 | 2021-04-16T12:26:04 | {} | MEMBER |
ClickHouse | ClickHouse | 821,663,601 | 5,429 | alexey-milovidov | @azat How comes it does not work in version 20.3?
https://github.com/ClickHouse/ClickHouse/issues/16897 | 2021-04-16T22:42:15 | 2021-04-16T22:42:15 | {} | MEMBER |
ClickHouse | ClickHouse | 821,669,798 | 8,655 | alexey-milovidov | @vdimir What do you think about supporting this feature? | 2021-04-16T22:45:22 | 2021-04-16T22:45:22 | {
"rocket": 2
} | MEMBER |
ClickHouse | ClickHouse | 821,705,096 | 15,185 | alexey-milovidov | The version 18.14 is too old, so this issue has lost its relevance. | 2021-04-16T23:02:45 | 2021-04-16T23:03:07 | {} | MEMBER |
ClickHouse | ClickHouse | 820,875,627 | 23,102 | godliness | > Some tests are failing, but I hope we will see great difference in out performance tests..
Thanks!
Test failed has been fixed.
About performance tests, above [information](https://github.com/ClickHouse/ClickHouse/pull/23102#issue-615701038) showed the before and after results, query time and memory usage ha... | 2021-04-16T03:10:26 | 2021-04-16T09:12:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,530,969 | 22,976 | filimonov | > > IMHO when we are executed inside deb we should only use deb standard ways to interact with user (or not interact with him at all).
>
> I'm too tired of Debian ways to do the things as well as RedHat ways to do the same things.
Interacting with user during package installation sucks. In redhat there is no 'off... | 2021-04-16T20:02:42 | 2021-04-16T20:02:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,586,578 | 23,183 | alexey-milovidov | What servers do you use to test these versions? | 2021-04-16T22:02:34 | 2021-04-16T22:02:34 | {} | MEMBER |
ClickHouse | ClickHouse | 821,681,507 | 17,146 | vitlibar | To execute `KILL MUTATION` (without `ON CLUSTER`) for a mutation you must have the exact grant which was required to start that mutation. It means e.g. to kill `alter delete` mutation you must have the `ALTER DELETE` grant.
To execute `KILL MUTATION ON CLUSTER` you must have all the grants `ALTER DELETE, ALTER UPDAT... | 2021-04-16T22:51:02 | 2021-04-16T22:52:59 | {} | MEMBER |
ClickHouse | ClickHouse | 821,701,649 | 15,700 | alexey-milovidov | @den-crane it looks alright.
@CurtizJ please clarify. | 2021-04-16T23:01:04 | 2021-04-16T23:01:04 | {} | MEMBER |
ClickHouse | ClickHouse | 820,452,039 | 22,359 | alexey-milovidov | A patch for TSan is required. | 2021-04-15T14:10:18 | 2021-04-15T14:10:18 | {} | MEMBER |
ClickHouse | ClickHouse | 821,108,166 | 23,105 | KochetovNicolai | Yes.
1. Is indeed partly implemented in #22087, but only for `LEFT` and `INNER` join.
2. It is not very logical indeed. Some legacy stuff. It would be more logical to have two inputs for join step.
3. Now it should not. I fixed in in some patch. But `CreatingSets` is still executed before join, so main query will no... | 2021-04-16T11:24:16 | 2021-04-16T11:24:16 | {} | MEMBER |
ClickHouse | ClickHouse | 821,199,853 | 23,038 | alesapin | Really, it's not possible to write such systems without jepsen. | 2021-04-16T14:02:44 | 2021-04-16T14:02:44 | {} | MEMBER |
ClickHouse | ClickHouse | 821,508,689 | 23,205 | den-crane | I understand that the issue is in my application. Just a ticket for the history & search. | 2021-04-16T19:14:45 | 2021-04-16T19:15:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,685,859 | 17,146 | alexey-milovidov | @MyroTk Please update the documentation. | 2021-04-16T22:53:10 | 2021-04-16T22:53:10 | {} | MEMBER |
ClickHouse | ClickHouse | 820,615,961 | 23,109 | filimonov | Good ORDER BY typically contains 3 to 5 columns.
The most important for filtering (and/or the lowest cardinal) on the left-most position.
Later importance is getting lower, and column cardinality - higher.
Ideally, if you can organize the data in the natural hierarchy order, for ex. `(continent, country, regio... | 2021-04-15T17:47:16 | 2021-04-15T17:47:16 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,626,940 | 23,105 | filimonov | Some optimizations on the pipeline level were merged couple a of days ago, see https://github.com/ClickHouse/ClickHouse/pull/22087
Not sure if it addresses your issue, but it's worth checking the behavior of the same query on the latest testing release.
@KochetovNicolai do you want to add some comments here? | 2021-04-15T18:05:24 | 2021-04-15T18:05:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,729,925 | 23,000 | alexey-milovidov | > Stress test (memory) — Logical error thrown (see clickhouse-server.log)
@tavplubix said, it's related to table loading:
> Mapping for table with UUID=00001192-0000-4000-8000-000000000002 already exists, посмотрю | 2021-04-15T21:05:31 | 2021-04-15T21:05:31 | {} | MEMBER |
ClickHouse | ClickHouse | 820,851,164 | 23,055 | Stephane987 | Thanks. I did not found the culprit yet and the error stopped spanning... but now I know where to look next time.
| 2021-04-16T01:55:03 | 2021-04-16T01:55:03 | {} | NONE |
ClickHouse | ClickHouse | 820,866,953 | 20,463 | dgzdot | This patch didn't really solve #19278 | 2021-04-16T02:39:22 | 2021-04-16T02:39:22 | {} | NONE |
ClickHouse | ClickHouse | 821,413,471 | 23,150 | alexey-milovidov | This information mostly duplicates https://clickhouse.tech/docs/en/getting-started/install/
The motivation adding it in "What is ClickHouse" page is unclear. | 2021-04-16T18:24:12 | 2021-04-16T18:24:12 | {} | MEMBER |
ClickHouse | ClickHouse | 821,425,931 | 23,088 | azat | Right [now](https://clickhouse-test-reports.s3.yandex.net/23088/29281ea6e062294652210d3a3011ce0998956f28/performance_comparison/report.html) (with `muzzy_decay_ms/dirty_decay_ms=5s`) overall 0.1% performance degradation looks normal (other perf tests reports has 0.1..0.4% of degradation), but I still want to collect so... | 2021-04-16T18:30:18 | 2021-04-16T18:31:06 | {} | MEMBER |
ClickHouse | ClickHouse | 821,479,134 | 23,038 | alesapin | For some reason `get` requests became faster:
Get requests with 32B payload:
Keeper
```
read requests 663006, write requests 0, Read RPS: 97260.6, Read MiB/s: 13.0785
Read sampler:
0% 0.000102419 sec.
10% 0.000193944 sec.
20% 0.000216892 sec.
30% 0.000235284 s... | 2021-04-16T18:56:39 | 2021-04-16T18:56:39 | {} | MEMBER |
ClickHouse | ClickHouse | 821,707,217 | 15,185 | alexey-milovidov | ALTER RENAME COLUMN is also implemented in newer versions. | 2021-04-16T23:03:48 | 2021-04-16T23:03:48 | {} | MEMBER |
ClickHouse | ClickHouse | 821,719,396 | 14,810 | alexey-milovidov | @MyroTk I see nothing wrong in this behaviour.
@vitlibar Please clarify. | 2021-04-16T23:18:51 | 2021-04-16T23:18:51 | {} | MEMBER |
ClickHouse | ClickHouse | 820,633,906 | 23,103 | filimonov | See also https://github.com/ClickHouse/ClickHouse/issues/2161#issuecomment-378197593 | 2021-04-15T18:16:18 | 2021-04-15T18:16:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,486,905 | 23,088 | alexey-milovidov | @azat The metrics about memory usage is at the bottom of the report. | 2021-04-16T19:00:28 | 2021-04-16T19:00:28 | {} | MEMBER |
ClickHouse | ClickHouse | 821,522,441 | 9,430 | alexey-milovidov | The essential metrics are:
- load average and the number of processes;
- time spent ("cpu usage" in userspace, kernel, io wait...);
- memory usage (total, cached, buffers, swap, free + cached...);
- disk (filesystem) usage (total, used - across all mounted filesystems);
- network traffic (received, transferred, ... | 2021-04-16T19:44:32 | 2021-04-16T19:44:32 | {} | MEMBER |
ClickHouse | ClickHouse | 820,699,712 | 23,057 | akuzm | @vdimir might know | 2021-04-15T20:08:12 | 2021-04-15T20:08:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,911,603 | 22,976 | filimonov | IMHO when we are executed inside deb we should only use deb standard ways to interact with user (or not interact with him at all). | 2021-04-16T05:15:58 | 2021-04-16T05:15:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,029,121 | 21,701 | keepflooding | 21.4.3.21 - schema quotation is still wrong when creating dictionary with postgresql source.
**Create query**
```
CREATE DICTIONARY dicts.zones
(
`id` UInt64,
`channel_id` UInt64 DEFAULT 0,
`timezone` String DEFAULT '',
`timezone_seconds` Int64 DEFAULT 0
)
PRIMARY KEY id
SOURCE(POSTGRESQL(
... | 2021-04-16T08:56:10 | 2021-04-16T08:56:10 | {} | NONE |
ClickHouse | ClickHouse | 821,489,674 | 23,088 | alexey-milovidov | I'm ok merging this change. | 2021-04-16T19:01:49 | 2021-04-16T19:01:49 | {} | MEMBER |
ClickHouse | ClickHouse | 821,487,800 | 23,038 | alesapin | One example without `fsync`:
### Create requests with 32 bytes payload (30 threads):
Keeper no fsync
```
read requests 0, write requests 399962, Write RPS: 58782.8, Write MiB/s: 8.57712. ... | 2021-04-16T19:00:54 | 2021-04-16T19:27:50 | {} | MEMBER |
ClickHouse | ClickHouse | 820,907,631 | 21,167 | filimonov | @KochetovNicolai any thoughts? | 2021-04-16T05:03:40 | 2021-04-16T05:03:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,133,044 | 23,102 | godliness | > > About performance tests, above [information](https://github.com/ClickHouse/ClickHouse/pull/23102#issue-615701038) showed the before and after results, query time and memory usage has been improved.
>
> Of course, I read it, but you used only 60MB file. Our performance tests use about ~10G of obfuscated data.
... | 2021-04-16T12:15:44 | 2021-04-16T12:19:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,140,814 | 22,976 | alexey-milovidov | > Fontend can be be also kde / gnome / web etc. I don't think they will do proper redirects of stdin / stdout.
Let's check it - it should not be able to use tty. | 2021-04-16T12:30:45 | 2021-04-16T12:30:45 | {} | MEMBER |
ClickHouse | ClickHouse | 821,425,631 | 23,150 | alexey-milovidov | @pavi2803 I don't feel it's good to change this documentation page.
I appreciate your help! Let's add documentation for some of the new features.
For example, there is no documentation of the `executable_pool` dictionary source. | 2021-04-16T18:30:09 | 2021-04-16T18:30:09 | {} | MEMBER |
ClickHouse | ClickHouse | 821,470,117 | 23,038 | alesapin | Single host with batches:
### Create requests with 32 bytes payload (30 threads):
Keeper + fsync
```
read requests 0, write requests 188478, Write RPS: 23723.5, Write MiB/s: 3.46155. ... | 2021-04-16T18:52:08 | 2021-04-16T18:57:16 | {} | MEMBER |
ClickHouse | ClickHouse | 821,652,607 | 17,147 | alexey-milovidov | @vitlibar has confirmed that it works perfectly as expected. | 2021-04-16T22:36:56 | 2021-04-16T22:36:56 | {} | MEMBER |
ClickHouse | ClickHouse | 821,667,252 | 16,897 | alexey-milovidov | It's not allowed to INSERT into MATERIALIZED column.
Resolution:
1. Update ClickHouse to the latest version.
2. Replace MATERIALIZED column to DEFAULT column with ALTER query to be able to insert data to it. | 2021-04-16T22:44:10 | 2021-04-16T22:44:10 | {} | MEMBER |
ClickHouse | ClickHouse | 820,742,329 | 21,670 | vitlibar | I've improved the documentation for `CREATE ROW POLICY`, added a special note about this issue.
https://github.com/ClickHouse/ClickHouse/pull/23164 | 2021-04-15T21:27:39 | 2021-04-15T21:47:04 | {} | MEMBER |
ClickHouse | ClickHouse | 820,890,572 | 22,208 | azat | @vdimir can you take a look? | 2021-04-16T04:05:33 | 2021-04-16T04:05:33 | {} | MEMBER |
ClickHouse | ClickHouse | 821,020,119 | 23,057 | vdimir |
Поведение режимов действительно довольно специальное, кроме того, они могу использоваться совместно.
В режиме `strict` цепочка считается начиная с первого вхождения первого события в цепочке (`Start`) и прерывается любым не ожидаемым событием из цепочки (второе вхождение `Anketa` в примере). События, которых нет... | 2021-04-16T08:41:00 | 2021-04-16T08:41:00 | {} | MEMBER |
ClickHouse | ClickHouse | 821,247,965 | 23,177 | tavplubix | Description check - fixed
Functional stateless tests (release, DatabaseReplicated) - 01305_replica_create_drop_zookeeper - flaky
Functional stateless tests (thread) - #23198
Yandex synchronization check - broken in master | 2021-04-16T15:14:52 | 2021-04-16T15:14:52 | {} | MEMBER |
ClickHouse | ClickHouse | 821,523,940 | 19,649 | vdimir | I'm going to benchmark implementation again, now I use axel, but I think s5cmd can be tested too.
Implementation intended to parallel downloading one big file from s3. If there are many files (using `*`) it will be parallelized on pipeline level.
| 2021-04-16T19:47:46 | 2021-04-16T19:51:55 | {} | MEMBER |
ClickHouse | ClickHouse | 821,711,151 | 15,176 | alexey-milovidov | CC @BohuTANG @zhang2014 | 2021-04-16T23:05:43 | 2021-04-16T23:05:43 | {} | MEMBER |
ClickHouse | ClickHouse | 820,910,439 | 22,976 | filimonov |
> Unfortunately it keeps asking password even if the variable `DEBIAN_FRONTEND=noninteractive` is set.
> (at least when terminal is available)
>
> I can add a check for this variable in `Install.cpp`.
I don't think the solution is safe. Fontend can be be also kde / gnome / web etc. I don't think they will do p... | 2021-04-16T05:12:23 | 2021-04-16T05:16:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,046,672 | 22,706 | tavplubix | AST fuzzer (UBSan) - OOM
| 2021-04-16T09:26:22 | 2021-04-16T09:26:22 | {} | MEMBER |
ClickHouse | ClickHouse | 821,140,348 | 9,703 | Jokser | I've faced the same problem and it was related to request throttling on ZK side due to the high load.
Try to set the system property `zookeeper.globalOutstandingLimit` to higher values for ZK server JVMs (default is 1000). | 2021-04-16T12:29:47 | 2021-04-16T12:37:40 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,582,748 | 23,054 | alexey-milovidov | We have found an issue in our release script.
The corrections are made in release 21.4.4. | 2021-04-16T21:53:36 | 2021-04-16T21:53:36 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 820,541,800 | 21,101 | kitaisreal | @madm1keru currently most reliable solution is to use dict get with type specified. | 2021-04-15T15:56:53 | 2021-04-15T15:56:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,615,220 | 23,155 | nikitamikhaylov | @Mergifyio rebase
| 2021-04-15T17:46:00 | 2021-04-15T17:46:00 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 820,622,220 | 23,110 | filimonov | `move_factor=0.8` means if there are less than 80% of free disk space on the first volume - move **the largest** part to the next volume. So i guess you have some large parts there which makes that difference.
See also
https://altinity.com/blog/2019/11/29/amplifying-clickhouse-capacity-with-multi-volume-storage-p... | 2021-04-15T17:57:44 | 2021-04-15T17:58:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,702,386 | 23,019 | akuzm | Please post the `CREATE TABLE` statements for both tables, `my_db.PI_TRAFFIC` and `my_db.all_PI_TRAFFIC`. | 2021-04-15T20:13:16 | 2021-04-15T20:13:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,281,250 | 21,167 | chyy | it seems the bug is still exists after upgrade
```
2021.04.15 17:09:30.304298 [ 32005 ] {1a75d1e4-5aa9-4c32-a732-b20c647d1d44} <Warning> TCPHandler: Client has gone away.
2021.04.15 17:09:30.576915 [ 24306 ] {} <Fatal> BaseDaemon: ########################################
2021.04.15 17:09:30.577016 [ 24306 ] {} <F... | 2021-04-15T09:34:48 | 2021-04-16T05:02:00 | {} | NONE |
ClickHouse | ClickHouse | 820,989,947 | 23,160 | azat | >[Functional stateless tests (thread) — fail: 6, passed: 517](https://clickhouse-test-reports.s3.yandex.net/23160/d2cf03ea41babefd715436f1f9e23e48fb3e6f8d/functional_stateless_tests_(thread).html#fail1)
~~Unfinished patch - https://github.com/ClickHouse-Extras/zlib-ng/commit/28dfdaa8a3c5add48dcaf56086a9306a357e6e6b~... | 2021-04-16T07:50:51 | 2021-04-16T07:56:51 | {} | MEMBER |
ClickHouse | ClickHouse | 821,076,963 | 23,162 | hexiaoting | I will try to fix this. | 2021-04-16T10:21:46 | 2021-04-16T10:21:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,139,447 | 22,627 | adrian17 | Also reproducible with other join types than CROSS.
Example, reproducible on 21.3.4:
```
with example as (select toDate('2021-01-01') as date, 1 as node, 1 as user)
SELECT extra_data
FROM (
SELECT join1.*
FROM example
LEFT JOIN ( SELECT toDate('2021-01-01') as date, 1 as extra_data ) AS join1
... | 2021-04-16T12:28:00 | 2021-04-16T12:28:00 | {} | NONE |
ClickHouse | ClickHouse | 821,457,679 | 22,359 | alexey-milovidov | #23153 | 2021-04-16T18:45:58 | 2021-04-16T18:45:58 | {} | MEMBER |
ClickHouse | ClickHouse | 821,599,569 | 23,183 | alexey-milovidov | The issue does not reproduce on Amazon Linux (AArch64, Graviton2):
```
__| __|_ )
_| ( / Amazon Linux 2 AMI
___|\___|___|
https://aws.amazon.com/amazon-linux-2/
No packages needed for security; 2 packages available
Run "sudo yum update" to apply all updates.
[ec2-user@ip-172-31-... | 2021-04-16T22:11:28 | 2021-04-16T22:11:28 | {} | MEMBER |
ClickHouse | ClickHouse | 821,653,370 | 17,147 | alexey-milovidov | The KILL grant is only needed to kill other users queries. | 2021-04-16T22:37:18 | 2021-04-16T22:37:18 | {} | MEMBER |
ClickHouse | ClickHouse | 821,678,555 | 16,708 | alexey-milovidov | @CurtizJ Still does not work. | 2021-04-16T22:49:36 | 2021-04-16T22:49:36 | {} | MEMBER |
ClickHouse | ClickHouse | 820,996,887 | 22,679 | ianton-ru | > @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).
I can't reproduce it.
Could you please help me?
I can't s... | 2021-04-16T08:02:25 | 2021-04-16T08:02:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,038,248 | 21,701 | kssenii | @keepflooding, because it was added only for table engine / table function. Ok, I'll add for dictionary source also. | 2021-04-16T09:11:50 | 2021-04-16T09:11:50 | {} | MEMBER |
ClickHouse | ClickHouse | 821,108,140 | 21,589 | akuzm | The test failures are from master. | 2021-04-16T11:24:14 | 2021-04-16T11:24:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,383,854 | 23,160 | azat | >BTW, we run all our production clusters with trace logging, it is by default and it is recommended for production usage (despite the common beliefs).
Me too, but some messages is just too verbose, and I add some suppressions via `logger.levels`. | 2021-04-16T18:09:50 | 2021-04-16T18:09:50 | {} | MEMBER |
ClickHouse | ClickHouse | 821,420,435 | 23,150 | alexey-milovidov | > ClickHouse is written in C++, and thereby supports the same.
It's not clear to me, what does it mean "supports the same"? How exactly ClickHouse *supports* C++... this information is technically correct but it's not essential to be located on this page. | 2021-04-16T18:27:35 | 2021-04-16T18:27:35 | {} | MEMBER |
ClickHouse | ClickHouse | 821,716,158 | 15,165 | alexey-milovidov | @MyroTk You have to setup access rights to the Distributed table.
Internal SELECT or INSERT subqueries that Distributed table is generating are not subjects of access control. | 2021-04-16T23:08:40 | 2021-04-16T23:09:10 | {} | MEMBER |
ClickHouse | ClickHouse | 820,635,320 | 23,059 | filimonov | /cc @abyss7 | 2021-04-15T18:18:24 | 2021-04-15T18:18:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,639,504 | 23,055 | filimonov | It is related to some mutations (ALTER UPDATE / DELETE), those statements are executed in a special manner, so they don't print the context / query in the logs.
You can find more details in system.mutations table. I guess you have some error in one of the statements there. | 2021-04-15T18:25:10 | 2021-04-15T18:26:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,954,328 | 23,110 | FoolishWall | > `move_factor=0.8` means if there are less than 80% of free disk space on the first volume - move **the largest** part to the next volume. So i guess you have some large parts there which makes that difference.
>
> See also
> https://altinity.com/blog/2019/11/29/amplifying-clickhouse-capacity-with-multi-volume-sto... | 2021-04-16T06:47:56 | 2021-04-16T06:57:17 | {} | NONE |
ClickHouse | ClickHouse | 821,131,017 | 23,102 | nikitamikhaylov | > About performance tests, above [information](https://github.com/ClickHouse/ClickHouse/pull/23102#issue-615701038) showed the before and after results, query time and memory usage has been improved.
Of course, I read it, but you used only 60MB file. Our performance tests use about ~10G of obfuscated data.
| 2021-04-16T12:11:38 | 2021-04-16T12:11:38 | {} | MEMBER |
ClickHouse | ClickHouse | 821,141,156 | 22,976 | alexey-milovidov | > IMHO when we are executed inside deb we should only use deb standard ways to interact with user (or not interact with him at all).
I'm too tired of Debian ways to do the things as well as RedHat ways to do the same things. | 2021-04-16T12:31:27 | 2021-04-16T12:31:27 | {} | MEMBER |
ClickHouse | ClickHouse | 821,488,235 | 23,088 | alexey-milovidov | -6.5% RSS. | 2021-04-16T19:01:08 | 2021-04-16T19:01:08 | {} | MEMBER |
ClickHouse | ClickHouse | 821,646,668 | 17,147 | alexey-milovidov | @MyroTk Isn't it alright if user is always being able to kill their own queries? | 2021-04-16T22:34:06 | 2021-04-16T22:34:06 | {} | MEMBER |
ClickHouse | ClickHouse | 820,525,139 | 21,770 | Avogar | @taylor12805 I found the problem, please fix it. | 2021-04-15T15:35:12 | 2021-04-15T15:35:12 | {} | MEMBER |
ClickHouse | ClickHouse | 820,551,105 | 21,101 | madm1keru | @kitaisreal
Do you mean dictGetFloat64 / dictGetString etc ? or some toFloat64(dictGet) ? | 2021-04-15T16:09:27 | 2021-04-15T16:09:27 | {} | NONE |
ClickHouse | ClickHouse | 821,410,652 | 23,198 | kitaisreal | Fixed in scope of https://github.com/ClickHouse/ClickHouse/pull/23153. | 2021-04-16T18:22:50 | 2021-04-16T18:22:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,435,838 | 22,299 | akuzm | > > @nikitamikhaylov you should copy changelog entry here, because the original PR was closed and won't show up in changelog. Also you'll have to mention the author manually in it.
>
> @akuzm Done.
You forgot the CL category :) I changed it to Improvement. | 2021-04-15T13:47:35 | 2021-04-15T13:47:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,613,017 | 23,115 | filimonov | Number of concurrent inserts looks big. That can happen if you start many inserts in parallel at the same time, or you send data to clickhouse slowly, or clickhouse write data to target table slow. Or some combination of those.
Not enough information to provide a better answer. You should provide at least informati... | 2021-04-15T17:42:21 | 2021-04-15T17:42:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,629,349 | 23,104 | filimonov | 20.6 and earlier - works. 20.7 and newer - fails. | 2021-04-15T18:09:15 | 2021-04-15T18:09:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,155,107 | 23,102 | nikitamikhaylov | > > > About performance tests, above [information](https://github.com/ClickHouse/ClickHouse/pull/23102#issue-615701038) showed the before and after results, query time and memory usage has been improved.
> >
> >
> > Of course, I read it, but you used only 60MB file. Our performance tests use about ~10G of obfuscat... | 2021-04-16T12:57:20 | 2021-04-16T12:57:20 | {} | MEMBER |
ClickHouse | ClickHouse | 821,192,397 | 23,038 | alesapin | the code became so optimal, that starts losing data:
```
INFO [2021-04-16 16:49:22,261] jepsen results - jepsen.store Wrote /home/alesap/code/cpp/ClickHouse/tests/jepsen.clickhouse-keeper/store/clickhouse-keeper-quorum=false-set-bridge-partitioner/20210416T164635.000+0300/results.edn
INFO [2021-04-16 16:49:22,305] j... | 2021-04-16T13:51:35 | 2021-04-16T14:06:08 | {
"laugh": 5
} | MEMBER |
ClickHouse | ClickHouse | 821,329,915 | 23,038 | alesapin | Oh yes, I'm giving equal session ID to different clients:
```
[20:17:13]alesap@click:~/.../ClickHouse/tests/jepsen.clickhouse-keeper$ grep 'Received session ID 21' store/latest/*/logs/clickhouse-server.log
store/latest/srv1/logs/clickhouse-server.log:2021.04.16 20:01:14.794681 [ 3649 ] {} <Information> NuKeeperTCPHa... | 2021-04-16T17:34:36 | 2021-04-16T17:34:36 | {} | MEMBER |
ClickHouse | ClickHouse | 821,532,700 | 22,976 | filimonov | For example i can start typing next command (or it's waiting in input buffer because i did copy & paste few commands). That password prompt will eat it accidebtally etc. | 2021-04-16T20:05:57 | 2021-04-16T20:05:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 821,659,228 | 17,146 | alexey-milovidov | @MyroTk
Most likely you are being able to execute your KILL MUTATION query because there are no mutations and there is nothing to kill. | 2021-04-16T22:40:08 | 2021-04-16T22:40:26 | {} | MEMBER |
ClickHouse | ClickHouse | 830,361,935 | 22,222 | Byteceptikus | @agusterodin Thanks!
Works on M1 | 2021-04-30T20:27:12 | 2021-04-30T20:28:04 | {
"laugh": 1
} | NONE |
ClickHouse | ClickHouse | 830,370,797 | 23,800 | filimonov | Marking major (regression) | 2021-04-30T20:39:32 | 2021-04-30T20:39:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 830,385,387 | 23,349 | azat | >Also not a good enough changelog entry.
@alexey-milovidov I've adjusted it a little | 2021-04-30T20:57:01 | 2021-04-30T20:57:01 | {} | MEMBER |
ClickHouse | ClickHouse | 830,538,216 | 19,569 | alexey-milovidov | We also have support to set different levels for different loggers (MergeTree, DDLWorker, etc.) | 2021-05-01T04:56:14 | 2021-05-01T04:56:14 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.