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 | 811,037,185 | 19,766 | vdimir | @achimbab
Yes, I'm going to look at code.
Could you describe in general what changes you made since last version?
| 2021-03-31T12:41:39 | 2021-03-31T12:41:39 | {} | MEMBER |
ClickHouse | ClickHouse | 811,176,490 | 22,409 | ilejn | @tangxueming
What you are doing is correct, registry_url is needed for AvroConfluent, not for Avro.
The issue you faced should be fixed by https://github.com/ClickHouse/ClickHouse/issues/21437
Please, repeat your test against a version it is backported to and let me know if it works for you. | 2021-03-31T15:54:39 | 2021-03-31T15:54:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,234,939 | 16,947 | KochetovNicolai | > @KochetovNicolai what do you think about this?
I think that names, types, and column positions should be the same and checked.
If we have query `SELECT _shard_num, count() FROM remote('127.0.0.{1,2}', system.one) GROUP BY _shard_num`, remote server should return columns with names `_shard_num, countState`, with t... | 2021-03-31T16:42:11 | 2021-03-31T16:42:11 | {} | MEMBER |
ClickHouse | ClickHouse | 811,536,902 | 19,233 | alexey-milovidov | Something is really wrong with this index: https://clickhouse-test-reports.s3.yandex.net/22411/5cd1dc485fbc5b8767f76e3abbb81c63218a6532/fuzzer_tsan/report.html#fail1
| 2021-03-31T23:49:03 | 2021-03-31T23:49:03 | {} | MEMBER |
ClickHouse | ClickHouse | 811,537,080 | 22,411 | alexey-milovidov | > AST fuzzer (TSan) — SUMMARY: ThreadSanitizer: heap-use-after-free obj-x86_64-linux-gnu/../src/Storages/MergeTree/MergeTreeIndexFullText.cpp:647:34 in DB::SplitT
https://github.com/ClickHouse/ClickHouse/issues/19233 | 2021-03-31T23:49:35 | 2021-03-31T23:49:35 | {} | MEMBER |
ClickHouse | ClickHouse | 810,269,235 | 22,239 | MaxJPowers | To further clarify, here is the discrepancy I'm dealing with:
Queries on table of 100000 rows and 25 columns with 69999 cached rows.
Expected hit rate: 0.6999992857142857
Hit rate: 0.9999997435897765
Mean cached time: 0.155
All entries queried were cached
Queries on table of 100000 rows and 25 columns wit... | 2021-03-30T13:52:03 | 2021-03-30T13:54:06 | {} | NONE |
ClickHouse | ClickHouse | 810,451,006 | 21,852 | excitoon | Idea of this PR is to address this issue: aws/aws-cli#4682 , inside docker AWS SDK can't access `http://169.254.169.254/latest/api/token` and waits too long before switching to `http://169.254.169.254/latest/meta-data/iam/security-credentials/`. This PR allows to manually specify which type of request is preferred. | 2021-03-30T17:41:28 | 2021-03-30T17:41:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,580,302 | 22,273 | filimonov | Mutation was still running / in progress when you was checking? (check system.mutations) | 2021-03-30T21:10:55 | 2021-03-30T21:11:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,613,345 | 22,365 | filimonov | Also can be useful to distinguish running queries faster, when there are lot of databases with similar set of tables, and the query don't list dbname explicitly, or when currentDatabase() function is used. | 2021-03-30T22:13:17 | 2021-03-30T22:13:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,599,796 | 22,381 | filimonov | Is it possible both for vertical and horizontal merges? Does it guarantee that long merge will still be getting some cpu time even if new tiny merges are pushed with high frequency?
Now heuristics which picks the parts to merge rely on pool size and how busy it is. Does it stay the same?
If number of merges will ... | 2021-03-30T21:48:13 | 2021-03-30T22:17:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,680,192 | 22,358 | alexey-milovidov | Alright. | 2021-03-31T01:12:51 | 2021-03-31T01:12:51 | {} | MEMBER |
ClickHouse | ClickHouse | 810,704,844 | 16,184 | alexey-milovidov | I think our codebase is just not ready for this function.
Correct solution will be the following:
1. @KochetovNicolai will add ordering properties on data streams.
2. Extend the interface of `IAggregateFunction` to support specialized implementation on ordered data.
3. If the data stream is ordered by arguments of ... | 2021-03-31T02:14:09 | 2021-03-31T02:14:09 | {} | MEMBER |
ClickHouse | ClickHouse | 810,872,914 | 22,399 | hexiaoting | try:
```
SELECT
id,
name,
sum(m1)
FROM
(
SELECT
d_array_id,
m1
FROM f
LEFT ARRAY JOIN d_ids AS d_array_id
) AS tmp
LEFT JOIN d ON d.id = tmp.d_array_id
GROUP BY
id,
name
``` | 2021-03-31T08:15:45 | 2021-03-31T08:15:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,307,658 | 22,404 | KochetovNicolai | Hi!
Not all stages of query are easily cancellable. But, also, it may be a bug.
If you have slow query, which cannot be killed, you may dump trace of query threads using query:
```sql
SELECT arrayStringConcat(arrayMap(x -> demangle(addressToSymbol(x)), trace), '\n')
FROM system.stack_trace
WHERE query_id = '24cf2... | 2021-03-31T18:22:10 | 2021-03-31T18:22:10 | {} | MEMBER |
ClickHouse | ClickHouse | 810,590,590 | 22,260 | filimonov | Yes, I know that, it's really obscure in mysql :)
Actually there is a really easy way to make it clear: use Unix socket only when it's path is provided as connection string, like in redis
```
redis-cli -h 127.0.0.1 -p 6379
vs
redis-cli -s /tmp/redis.sock
```
One more alternative is to probe both stacks. But ... | 2021-03-30T21:30:18 | 2021-03-30T21:30:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,687,768 | 20,994 | onine007 | > We had ran into that situation, you can try to delete all the data stored in the table my_sdap.crowd_bitmap_result_replicated and upgrade to version 21.2.3 and write data to it again. The error message won't appear.
> The reason for this error is in old version ClickHouse AggregateFunction(groupBitmap, UInt32) encod... | 2021-03-31T01:31:48 | 2021-03-31T01:31:48 | {} | NONE |
ClickHouse | ClickHouse | 810,912,580 | 11,741 | laowangriji | same crash error: 21.2.3.15 | 2021-03-31T09:16:16 | 2021-03-31T09:16:16 | {} | NONE |
ClickHouse | ClickHouse | 810,945,363 | 22,156 | olgarev | @nikitamikhaylov Hi, will you merge this PR, pls? So we can close the ticket. | 2021-03-31T10:06:22 | 2021-03-31T10:06:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,782,082 | 20,193 | Zhile | Has this been fixed? I think it's a bug. This can be still reproduced in 21.2 version.
@alexey-milovidov Can you help to take a look? | 2021-03-31T05:42:38 | 2021-03-31T05:42:38 | {} | NONE |
ClickHouse | ClickHouse | 811,431,903 | 22,405 | den-crane | Try to move Timeslot to the end of index
```
CREATE TABLE IF NOT EXISTS ttl_test
(
Timeslot DateTime,
Device String,
EType String,
Proto String,
Bytes UInt64,
Packets UInt64
) ENGINE = SummingMergeTree()
PARTITION BY toYYYYMMDD(Timeslot)
ORDER BY (Device,EType,Proto, toStartOfHour(Timeslot), T... | 2021-03-31T20:15:05 | 2021-03-31T20:15:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,545,233 | 22,343 | tavplubix | Integration tests (thread) - https://clickhouse-test-reports.s3.yandex.net/22343/5f3cb525eea8f512ebc4e0bab38861c0a79d5dfa/integration_tests_(thread).html - OK
There are two `CLICKHOUSE_INTEGRATION_TESTS_TSAN` tasks with `commit_sha=5f3cb525eea8f512ebc4e0bab38861c0a79d5dfa`: https://nda.ya.ru/t/rigd56N_3qdKne. The firs... | 2021-03-30T20:10:23 | 2021-03-30T20:10:23 | {} | MEMBER |
ClickHouse | ClickHouse | 810,713,912 | 22,304 | maskshell | > It's not neccessary that the reason of
> `ZooKeeper session has been expired`
> is expiration by timeout.
>
> It may become 'expired' after some failures too, check if you see some other zookeeper related exceptions (also in ZooKeeper logs) around the time you see the 'expired' exception.
>
> Maybe we should ... | 2021-03-31T02:38:47 | 2021-03-31T02:40:33 | {} | NONE |
ClickHouse | ClickHouse | 810,837,319 | 22,370 | alesapin | No related failures. | 2021-03-31T07:20:10 | 2021-03-31T07:20:10 | {} | MEMBER |
ClickHouse | ClickHouse | 810,627,399 | 22,340 | den-crane | https://github.com/ClickHouse/ClickHouse/issues/14648 | 2021-03-30T22:45:14 | 2021-03-30T22:45:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,732,207 | 22,326 | alexey-milovidov | Clear win. | 2021-03-31T03:31:55 | 2021-03-31T03:31:55 | {} | MEMBER |
ClickHouse | ClickHouse | 810,758,876 | 22,398 | alexey-milovidov | CERN ClickHouse vs InfluxDB: http://cds.cern.ch/record/2667383/ | 2021-03-31T04:51:11 | 2021-03-31T04:51:11 | {} | MEMBER |
ClickHouse | ClickHouse | 810,870,708 | 22,385 | UnamedRus | I would expect it to build tuple as result.
SELECT argMaxSimpleState(1, 2) = (1,2)
SELECT argMaxSimpleState((1, 1), 'anything') = ((1,2), 'anything')
| 2021-03-31T08:12:27 | 2021-03-31T08:12:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,148,799 | 21,939 | kitaisreal | Was included in scope of https://github.com/ClickHouse/ClickHouse/pull/22097. | 2021-03-31T15:17:35 | 2021-03-31T15:17:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,752,737 | 14,893 | fpacifici | +1 We would be very interested in this as well. Specifically being able to exchange sketches with clickhouse.
Our use case includes a very high load data ingestion pipeline where it would be desirable to aggregate sets or quantiles close to the client to reduce the amount of data transferred and being able to stor... | 2021-03-31T04:33:17 | 2021-03-31T04:33:17 | {} | NONE |
ClickHouse | ClickHouse | 810,803,099 | 16,108 | alexey-milovidov | I thought it's already fixed. | 2021-03-31T06:18:28 | 2021-03-31T06:18:28 | {} | MEMBER |
ClickHouse | ClickHouse | 810,962,410 | 22,115 | Lishengchao | > Looks similar to #13406.
During this period of time, I have been observing, and I found that clikchouse would crash from time to time when reading HDFS.
The first case: too many queries;
The second case: the time of the crash is uncertain, sometimes one or two hours, sometimes three or four days.
The above cr... | 2021-03-31T10:35:15 | 2021-03-31T10:35:15 | {} | NONE |
ClickHouse | ClickHouse | 811,062,471 | 22,411 | nvartolomei | This is required for #20088 to work properly. | 2021-03-31T13:20:42 | 2021-03-31T13:20:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,098,314 | 22,113 | filimonov | BTW it's a nice task for an intern. Very independent and requires diving into one particular subject - how binary files are organized / how compilers work etc. | 2021-03-31T14:09:31 | 2021-03-31T14:09:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,196,646 | 22,414 | ilejn | Plan to handle it. | 2021-03-31T16:18:39 | 2021-03-31T16:18:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,144,795 | 22,406 | nikitamikhaylov | What will be if you set `input_format_parallel_parsing=0` ? | 2021-03-31T15:12:21 | 2021-03-31T21:31:24 | {} | MEMBER |
ClickHouse | ClickHouse | 810,566,340 | 22,359 | alexey-milovidov | https://github.com/zlib-ng/zlib-ng/pull/462 ? | 2021-03-30T20:46:10 | 2021-03-30T20:46:57 | {} | MEMBER |
ClickHouse | ClickHouse | 810,701,701 | 21,694 | alexey-milovidov | It should work on this database: https://github.com/matomo-org/device-detector/tree/master/regexes
| 2021-03-31T02:06:15 | 2021-03-31T02:06:15 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 810,709,699 | 16,947 | alexey-milovidov | Maybe @akuzm can take it for review? | 2021-03-31T02:27:24 | 2021-03-31T02:27:24 | {} | MEMBER |
ClickHouse | ClickHouse | 810,716,767 | 22,384 | alexey-milovidov | If I understand, the expectation was that argMax with tuple argument will return tuple. | 2021-03-31T02:46:54 | 2021-03-31T02:46:54 | {} | MEMBER |
ClickHouse | ClickHouse | 811,099,887 | 22,408 | filimonov | One more test - here also the error of quantile is drastic:
```
SELECT
quantile(0.95)(v) AS quantile,
quantileTiming(0.95)(v) AS timing,
quantileExact(0.95)(v) AS exact,
quantileTDigest(0.95)(v) AS tdigest
FROM
(
SELECT
multiIf(number < 15, 1, 2) AS k,
multiIf(number < 10... | 2021-03-31T14:11:40 | 2021-03-31T14:13:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,702,032 | 21,694 | alexey-milovidov | @filimonov
> I'm afraid that tree may be not enough.
Our experience in [Yandex.Metrica](https://metrica.yandex.com/) shows that tree is enough. | 2021-03-31T02:07:02 | 2021-03-31T02:07:02 | {} | MEMBER |
ClickHouse | ClickHouse | 810,702,305 | 11,160 | alexey-milovidov | Close in favor of #21694. | 2021-03-31T02:07:43 | 2021-03-31T02:07:43 | {} | MEMBER |
ClickHouse | ClickHouse | 810,889,809 | 20,193 | hexiaoting | This query can run in version : 21.4.1.1
```
dell123 :) create table test (id Int) Engine=Memory;;
CREATE TABLE test
(
`id` Int
)
ENGINE = Memory
Query id: bea26518-a41a-4441-b4c2-e5cd01299810
Ok.
0 rows in set. Elapsed: 0.030 sec.
dell123 :) create table test1 (id Int,name String) Engine=Memor... | 2021-03-31T08:41:29 | 2021-03-31T08:41:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,911,856 | 21,361 | Zhile | Hi @cyberhuman Which version do you use now?
> I think #21264 may be related. We've just upgraded to a version that contains the fix and it looks good so far 🤞🏻
| 2021-03-31T09:15:08 | 2021-03-31T09:15:57 | {} | NONE |
ClickHouse | ClickHouse | 811,164,738 | 22,373 | alesapin | For some reason sometimes clickhouse keeper is not able to start on nodes after binary download. When we run from already downloaded everything looks good. | 2021-03-31T15:39:01 | 2021-03-31T15:41:26 | {} | MEMBER |
ClickHouse | ClickHouse | 811,486,839 | 22,008 | nikitamikhaylov | @alexey-milovidov
> Because `new`/`delete` hooks are disabled with sanitizers.
BTW, It is only disabled for gcc compiler https://github.com/ClickHouse/ClickHouse/blob/6fb444f732b6c6315e5d511690949175ca21d6c8/src/CMakeLists.txt#L143-L144
And do you mean these operator overloading under the word 'hook'? https:/... | 2021-03-31T21:45:53 | 2021-03-31T21:45:53 | {} | MEMBER |
ClickHouse | ClickHouse | 810,632,895 | 22,260 | alexey-milovidov | Note that the bridges are also using HTTP with chunked transfer encoding.
This is expected to have higher overhead than TCP vs unix domain socket (noticeable at several GB/sec per core).
And clickhouse-odbc-bridge is slow not for this reason. | 2021-03-30T22:59:32 | 2021-03-30T22:59:32 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 810,625,969 | 22,376 | den-crane | Yes, it reads data in the table and materialize ttl.txt. Use `set materialize_ttl_after_modify=0` to avoid. | 2021-03-30T22:41:32 | 2021-03-30T23:17:25 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,666,424 | 22,157 | alexey-milovidov | It's trivially reproducible and I have MSan build at this moment of time... Will dig it. | 2021-03-31T00:32:44 | 2021-03-31T00:32:44 | {} | MEMBER |
ClickHouse | ClickHouse | 810,708,626 | 19,242 | alexey-milovidov | @abyss7 Is it abandoned? | 2021-03-31T02:24:25 | 2021-03-31T02:24:25 | {} | MEMBER |
ClickHouse | ClickHouse | 810,709,153 | 20,151 | alexey-milovidov | Depends on #20495, #20499, #18205. | 2021-03-31T02:25:51 | 2021-03-31T02:25:51 | {} | MEMBER |
ClickHouse | ClickHouse | 810,717,083 | 22,384 | alexey-milovidov | @ildus @Avogar Could you please clarify? | 2021-03-31T02:47:52 | 2021-03-31T02:47:52 | {} | MEMBER |
ClickHouse | ClickHouse | 810,826,979 | 17,495 | simPod | What is the alternative? | 2021-03-31T07:03:18 | 2021-03-31T07:03:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,866,315 | 22,399 | Zhile | @4ertus2 @hexiaoting Can you help to take a look? Thanks. | 2021-03-31T08:05:48 | 2021-03-31T08:05:48 | {} | NONE |
ClickHouse | ClickHouse | 810,891,960 | 22,399 | hexiaoting | Join ON is not supported completely in the latest version, It need to be improved to support more situations. like:https://github.com/ClickHouse/ClickHouse/issues/21868 | 2021-03-31T08:44:53 | 2021-03-31T08:44:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,028,661 | 22,384 | filimonov | +1 for rollback #17359. It's too easy to do it with simple max/min by putting the second argument at the first position in tuple.
> rollback ... #4629
It's not funny. It's widely used. | 2021-03-31T12:26:38 | 2021-03-31T12:26:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,030,332 | 19,766 | ghost | @vdimir
It seems that failed testcases are irrelavent with this PR.
May I get your code-review again? | 2021-03-31T12:29:29 | 2021-03-31T12:30:52 | {} | NONE |
ClickHouse | ClickHouse | 811,077,364 | 16,947 | akuzm | > Also don't miss @akuzm suggestion of matching columns by index for the remote queries ([#16947 (comment)](https://github.com/ClickHouse/ClickHouse/pull/16947#issuecomment-735808270)), I'm afraid ([#16947 (comment)](https://github.com/ClickHouse/ClickHouse/pull/16947#issuecomment-736839876)) that it will be less robus... | 2021-03-31T13:41:08 | 2021-03-31T13:41:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,088,825 | 22,408 | filimonov | https://github.com/ClickHouse/ClickHouse/issues/1158 ? | 2021-03-31T13:56:35 | 2021-03-31T13:56:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,663,408 | 22,357 | alexey-milovidov | Ok. | 2021-03-31T00:23:59 | 2021-03-31T00:23:59 | {} | MEMBER |
ClickHouse | ClickHouse | 810,706,657 | 17,769 | alexey-milovidov | > Was thinking what is the best approach here ignore a missing setting or add IF EXISTS statement?
Ignore if setting was not set is Ok. | 2021-03-31T02:19:17 | 2021-03-31T02:19:17 | {} | MEMBER |
ClickHouse | ClickHouse | 810,790,607 | 21,956 | Zhile | We don't need to backport anything to 20.12. We just need to build a release for 20.12. If this is not possible, please help to fix 21.3. Many thanks. | 2021-03-31T06:00:53 | 2021-03-31T06:00:53 | {} | NONE |
ClickHouse | ClickHouse | 809,956,991 | 22,110 | excitoon | @azat May be some common runtime check which will be turned on in Debug and throw
exception if at moment of destruction there was written some data but not
finalized?
On Tue, Mar 30, 2021, 8:39 AM Azat Khuzhin ***@***.***> wrote:
> User must call finalize() after they write somthing and the reason we
> introdu... | 2021-03-30T06:43:07 | 2021-03-31T07:47:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,182,880 | 22,373 | alesapin | Sigh https://clickhouse-test-reports.s3.yandex.net/22373/019d4b3ebd0c5a04e49add5af39631592cecdf95/clickhouse_keeper_jepsen/jepsen_run_all_tests.log
```
7 successes
0 unknown
3 crashed <-- something wrong with Jepsen or startup
0 failures
``` | 2021-03-31T16:03:32 | 2021-03-31T16:04:10 | {} | MEMBER |
ClickHouse | ClickHouse | 811,311,237 | 22,405 | KochetovNicolai | @CurtizJ could you help, please?
At lease, it is not good that we have an exception not after table creation, but only when TTL is applied. | 2021-03-31T18:27:31 | 2021-03-31T18:27:31 | {} | MEMBER |
ClickHouse | ClickHouse | 810,438,550 | 21,852 | excitoon | Hi @nikitamikhaylov . I reworked my PR. Can you please re-review it? | 2021-03-30T17:21:50 | 2021-03-30T17:21:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,666,750 | 22,157 | alexey-milovidov | As UInt256 is involved, I change the classification from "bug" to "bug experimental". | 2021-03-31T00:33:41 | 2021-03-31T00:33:41 | {} | MEMBER |
ClickHouse | ClickHouse | 810,667,590 | 22,157 | alexey-milovidov | `SELECT range(toUInt256(1), 1)` | 2021-03-31T00:36:11 | 2021-03-31T00:36:11 | {} | MEMBER |
ClickHouse | ClickHouse | 811,138,365 | 22,217 | UnamedRus | Does it support case when we have only one dns name with multiple ip's for mysql servers for it? | 2021-03-31T15:03:28 | 2021-03-31T15:03:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,663,599 | 22,339 | bohutang | This should be a generic requirement, [MySQl engine](https://clickhouse.tech/docs/en/engines/table-engines/integrations/mysql/) also has this problem, it would be better to call a function, such as Mask. | 2021-03-31T00:24:34 | 2021-03-31T00:24:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,679,407 | 22,384 | alexey-milovidov | @amosbird We should either fix it or rollback one of #4629 #17359 | 2021-03-31T01:10:39 | 2021-03-31T01:11:16 | {} | MEMBER |
ClickHouse | ClickHouse | 810,706,731 | 17,769 | alexey-milovidov | @xjewer Let's revive it. | 2021-03-31T02:19:29 | 2021-03-31T02:19:29 | {} | MEMBER |
ClickHouse | ClickHouse | 810,786,904 | 21,956 | alexey-milovidov | We already stopped backporting changes to version 20.12, it will be more easy to ship fixed to 21.3 than to revive 20.12. | 2021-03-31T05:52:44 | 2021-03-31T05:52:44 | {} | MEMBER |
ClickHouse | ClickHouse | 810,837,120 | 22,368 | alesapin | No related failures | 2021-03-31T07:19:52 | 2021-03-31T07:19:52 | {} | MEMBER |
ClickHouse | ClickHouse | 811,167,313 | 22,353 | CurtizJ | There should be a check like it's done at inserts (see `Nested::validateArraySizes`). Looks like this behaviour always has been (I checked at 20.8). | 2021-03-31T15:42:29 | 2021-03-31T15:42:58 | {} | MEMBER |
ClickHouse | ClickHouse | 810,705,375 | 17,495 | alexey-milovidov | No need for Etcd, let's drop it :tada: | 2021-03-31T02:15:38 | 2021-03-31T02:15:38 | {
"confused": 1
} | MEMBER |
ClickHouse | ClickHouse | 810,709,964 | 22,384 | amosbird | What's the difference between `argMax((c1, c2))` and `max((c1, c2))` ? I think we can rollback https://github.com/ClickHouse/ClickHouse/pull/17359 as it introduces too many specifications to aggregators. | 2021-03-31T02:28:07 | 2021-03-31T02:28:07 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 810,710,351 | 22,385 | alexey-milovidov | How SimpleState is supposed to work for aggregate functions with multiple arguments? | 2021-03-31T02:29:08 | 2021-03-31T02:29:08 | {} | MEMBER |
ClickHouse | ClickHouse | 810,716,936 | 22,273 | xiaogaozi | > Mutation was still running / in progress when you was checking? (check system.mutations)
@filimonov No, all mutations have already done. Actually, I've waited several days after execute `MODIFY TTL` query.
```sql
SELECT count(*)
FROM system.mutations
WHERE (table = 'xxx') AND (command = 'MATERIALIZE TTL') AN... | 2021-03-31T02:47:26 | 2021-03-31T02:48:50 | {} | NONE |
ClickHouse | ClickHouse | 810,743,602 | 16,562 | nvartolomei | @alexey-milovidov this is an RFC. While the code is outdated (and was never meant to be merged in this state), the RFC/proposal is still valid.
I remember you saying in a chat that it doesn't make a lot of sense to introduce protobuf because it is not applicable everywhere and we'll just introduce one more way of S... | 2021-03-31T04:07:34 | 2021-03-31T04:07:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,758,301 | 22,398 | alexey-milovidov | ClickHouse vs Elasticsearch: https://my.oschina.net/taogang/blog/4965018 | 2021-03-31T04:49:33 | 2021-03-31T04:49:33 | {} | MEMBER |
ClickHouse | ClickHouse | 810,814,062 | 22,381 | amosbird | Is it possible that huge merges get starving to death? | 2021-03-31T06:40:06 | 2021-03-31T06:40:06 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 810,836,674 | 22,369 | alesapin | `00067_union_all` not related | 2021-03-31T07:19:13 | 2021-03-31T07:19:13 | {} | MEMBER |
ClickHouse | ClickHouse | 810,703,422 | 16,562 | alexey-milovidov | @alesapin @nvartolomei Is it outdated? | 2021-03-31T02:10:17 | 2021-03-31T02:10:17 | {} | MEMBER |
ClickHouse | ClickHouse | 810,708,459 | 20,189 | alexey-milovidov | This PR is not ready (nothing works).
Plase open another PR. | 2021-03-31T02:24:01 | 2021-03-31T02:24:01 | {} | MEMBER |
ClickHouse | ClickHouse | 810,708,787 | 16,401 | alexey-milovidov | Maybe reopen later. | 2021-03-31T02:24:52 | 2021-03-31T02:24:52 | {} | MEMBER |
ClickHouse | ClickHouse | 810,630,136 | 22,381 | alexey-milovidov | > Is it possible both for vertical and horizontal merges?
Yes.
> Does it guarantee that long merge will still be getting some cpu time even if new tiny merges are pushed with high frequency?
No and this is as expected. When there are too many small merges, large merges will not be processed.
> Now heuristic... | 2021-03-30T22:52:16 | 2021-03-30T22:55:02 | {} | MEMBER |
ClickHouse | ClickHouse | 810,678,357 | 19,575 | alexey-milovidov | Hi, how is it going? | 2021-03-31T01:07:38 | 2021-03-31T01:07:38 | {} | MEMBER |
ClickHouse | ClickHouse | 810,792,562 | 22,387 | alexey-milovidov | > ThreadSanitizer: clock allocator overflow (65536*1024). Dying. | 2021-03-31T06:04:50 | 2021-03-31T06:04:50 | {} | MEMBER |
ClickHouse | ClickHouse | 810,854,621 | 22,110 | excitoon | This PR is ready for review, failed tests are unrelated. | 2021-03-31T07:48:00 | 2021-03-31T07:48:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 810,944,564 | 21,268 | olgarev | @kssenii Hi, wil you review this PR, pls? | 2021-03-31T10:04:59 | 2021-03-31T10:04:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,101,597 | 22,271 | kozlovalex | RabbitMQ 3.8.14
Erlang 23.3
everything works!
but
RabbitMQ 3.3.5
Erlang R16B03-1
in production (((
| 2021-03-31T14:13:58 | 2021-03-31T14:13:58 | {} | NONE |
ClickHouse | ClickHouse | 811,176,812 | 22,318 | akuzm | The test failure looks unrelated. | 2021-03-31T15:55:04 | 2021-03-31T15:55:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,212,723 | 22,373 | alesapin | This check should happen each second but something happens with VM:
```
INFO [2021-03-31 19:27:47,035] jepsen node n1 - jepsen.clickhouse-keeper.utils Checking server alive on
??? what is going on here?
INFO [2021-03-31 19:28:10,370] jepsen node n1 - jepsen.clickhouse-keeper.utils Checking server alive on
``` | 2021-03-31T16:29:05 | 2021-03-31T16:29:28 | {} | MEMBER |
ClickHouse | ClickHouse | 811,498,797 | 2,030 | ramazanpolat | > > max_bytes_before_external_group_by
>
> How to set max_bytes_before_external_group_by because this configuration is not present in config.xml
Use `users.xml` file which resides in `/etc/clickhouse-server/users.xml`.
Example `users.xml` config:
```xml
<yandex>
<!-- Profiles of settings. -->
<pr... | 2021-03-31T22:09:36 | 2021-03-31T22:09:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 811,544,614 | 22,299 | alexey-milovidov | Also it's possible to add a perf test. | 2021-04-01T00:08:55 | 2021-04-01T00:08:55 | {} | MEMBER |
ClickHouse | ClickHouse | 824,412,036 | 22,711 | nikitamikhaylov | @stephane-moreau Also I noticed that you have `<table_pull>yellow_tripdata</table_pull>` in config. You have to specify the original table (*MergeTree engine). So try `<table_pull>yellow_tripdata_staging</table_pull>`
You chosen wrong `PARTITION KEY`, because even in my tests (I use a very small piece of the data fr... | 2021-04-21T22:43:37 | 2021-04-23T00:24:43 | {} | MEMBER |
ClickHouse | ClickHouse | 825,337,098 | 19,981 | hexiaoting | This feature will be discussed in https://github.com/ClickHouse/ClickHouse/issues/23516 with a new datatype `Object`. | 2021-04-23T02:15:46 | 2021-04-23T02:15:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 825,383,362 | 23,539 | l1t1 | are there methods to materialize the subquery t1, so that we only scan the numbers once? | 2021-04-23T04:49:26 | 2021-04-23T04:49:26 | {} | NONE |
ClickHouse | ClickHouse | 825,421,030 | 23,514 | akuzm | It works. | 2021-04-23T06:25:08 | 2021-04-23T06:25:08 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.