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 | 655,590,762 | 12,298 | filimonov | 19.16 - works, 19.17 and newer - broken | 2020-07-08T15:28:21 | 2020-07-08T15:28:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,896,971 | 11,815 | alexey-milovidov | Fix: #12315 | 2020-07-09T04:46:56 | 2020-07-09T04:46:56 | {} | MEMBER |
ClickHouse | ClickHouse | 655,907,599 | 12,110 | filimonov | Relates https://github.com/ClickHouse/ClickHouse/issues/4609 | 2020-07-09T05:21:47 | 2020-07-09T05:21:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,987,704 | 4,746 | xmariachi | Hi @nvartolomei, was this fixed or worked around? Could you please share any insights? Thanks! | 2020-07-09T08:29:01 | 2020-07-09T08:29:01 | {} | NONE |
ClickHouse | ClickHouse | 656,202,411 | 12,339 | akuzm | It may be interesting to add a scoped class like `getCurrentMemoryTrackerActionLock` that disables memory allocation at all, i.e. makes all new() calls throw. This way we can be sure that there is no allocation. | 2020-07-09T15:42:52 | 2020-07-09T15:42:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 656,223,646 | 12,302 | azat | >Otherwise you will have no data from not stalled buffer, and error will happen in another place.
What error? And is it a problem? | 2020-07-09T16:21:53 | 2020-07-09T16:21:53 | {} | MEMBER |
ClickHouse | ClickHouse | 655,616,103 | 9,075 | miggi | Thanks, I'm using one copier of course.
Yes, but we had a similar option:
<max_workers>4</max_workers>
for that. As far I understood workers = number of import threads, ideally equal to number of shards.
| 2020-07-08T16:13:28 | 2020-07-08T16:13:28 | {} | NONE |
ClickHouse | ClickHouse | 655,814,414 | 12,166 | alexey-milovidov | `SELECT count() FROM system.parts WHERE table = NULL AND database = currentDatabase()` | 2020-07-08T23:46:44 | 2020-07-08T23:46:44 | {} | MEMBER |
ClickHouse | ClickHouse | 655,858,365 | 12,181 | alexey-milovidov | @azat Now we have in-source autogenerated file
```
modified: ../contrib/arrow-cmake/cpp/src/arrow/util/config.h
```
That is constantly modified. It's annoying. | 2020-07-09T02:26:54 | 2020-07-09T02:26:54 | {} | MEMBER |
ClickHouse | ClickHouse | 655,954,519 | 12,181 | azat | >@azat Now we have in-source autogenerated file
@alexey-milovidov fixed in 0aa97af108e80f114be6542f8aaa2b07892ed4e6 (#12319) | 2020-07-09T07:28:15 | 2020-07-09T07:28:15 | {} | MEMBER |
ClickHouse | ClickHouse | 656,168,941 | 12,329 | hiliuxg | @realyota @den-crane Is this the standard sql syntax? This is different from the syntax of hive and presto. It is also different from the public's usage habits. I just thought it was a bug. | 2020-07-09T14:41:30 | 2020-07-09T14:41:30 | {} | NONE |
ClickHouse | ClickHouse | 655,409,217 | 12,261 | filimonov | ```
CREATE TABLE test
(
`ext1` UInt16,
`ext2` AggregateFunction(groupBitmap, UInt32),
`ds` UInt32
)
ENGINE = AggregatingMergeTree
PARTITION BY (ds, ext1)
ORDER BY ext1
SETTINGS index_granularity = 8192
Ok.
0 rows in set. Elapsed: 0.008 sec.
INSERT INTO test SELECT
toUInt16(1) AS ex... | 2020-07-08T09:37:40 | 2020-07-08T09:39:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,476,876 | 11,867 | alexey-milovidov | If you group by single column, you can just replace WITH ROLLUP to WITH TOTALS to get the desired result. | 2020-07-08T12:04:57 | 2020-07-08T12:04:57 | {} | MEMBER |
ClickHouse | ClickHouse | 655,757,148 | 12,178 | azat | @alexey-milovidov friendly ping (pinging because build fixes requires rebase of other things on top of them to compile locally, and rebase against upstream for submitting) | 2020-07-08T21:05:46 | 2020-07-08T21:05:46 | {} | MEMBER |
ClickHouse | ClickHouse | 656,174,238 | 11,815 | alexey-milovidov | 00975_indices_mutation_replicated_zookeeper.sh
https://clickhouse-test-reports.s3.yandex.net/0/dd907b2ee98af1920c6a990fcc5fcb669098cdba/functional_stateless_tests_(debug).html#fail1
Fix #12336 | 2020-07-09T14:50:41 | 2020-07-09T14:52:33 | {} | MEMBER |
ClickHouse | ClickHouse | 656,215,555 | 12,302 | filimonov | Interesting. Didn't meet the situation when poll return a mix of errors & messages. Can it be reproduced somehow?
I think this fix can leave buffer in an incorrect state (it will not be stalled but will message attributes will not be accessible).
The idea is poll should either return true and real message, eithe... | 2020-07-09T16:06:55 | 2020-07-09T20:55:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,610,692 | 12,284 | healiseu | Great news, thank you @blinkov, by the way I have just made a cross-reference with a relevant issue that I opened some time ago at mymarilyn/clickhouse-driver#128. In case someone is willing to help @xzkostyan to support ClickHouse Arrow arrays format I volunteer to test the new feature.
My plan is to support ClickH... | 2020-07-08T16:03:34 | 2020-07-08T18:09:26 | {} | NONE |
ClickHouse | ClickHouse | 655,767,774 | 12,303 | CurtizJ | Maybe related #12301, #12299. | 2020-07-08T21:29:15 | 2020-07-08T21:29:15 | {} | MEMBER |
ClickHouse | ClickHouse | 655,809,730 | 12,304 | CurtizJ | The problem appeared, because `definition_ast` wasn't cloned and it was accidently updated earlier, than it was expeceted, and TTL expressions weren't recalculated. | 2020-07-08T23:29:25 | 2020-07-08T23:29:51 | {} | MEMBER |
ClickHouse | ClickHouse | 655,843,272 | 11,234 | vzakaznikov | @alesapin, @vitlibar, LDAP tests in TestFlows are now passing. I think this PR is ready for final review and merge. | 2020-07-09T01:34:02 | 2020-07-09T01:34:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,857,935 | 12,313 | alexey-milovidov | Tested manually. | 2020-07-09T02:25:21 | 2020-07-09T02:25:21 | {} | MEMBER |
ClickHouse | ClickHouse | 655,886,653 | 11,815 | alexey-milovidov | Race condition in ReplicatedMergeTreeQueue: https://clickhouse-test-reports.s3.yandex.net/12313/e17995cb75316417326b1f882409a571cbdf91c2/functional_stateless_tests_(thread)/stderr.log | 2020-07-09T04:10:56 | 2020-07-09T04:10:56 | {} | MEMBER |
ClickHouse | ClickHouse | 656,033,774 | 12,066 | qoega | Original PR #2760 | 2020-07-09T09:59:37 | 2020-07-09T09:59:37 | {} | MEMBER |
ClickHouse | ClickHouse | 656,131,859 | 1,118 | czwanglei | > I'm fairly new to Clickhouse and as I'm developing queries I'd like to be able to optimize them. With a traditional RDBMS I can examine the query plan to get a sense of what my query will do without running it. I'd like for Clickhouse to be able to give me some kind of diagnostic output that would guide my optimizati... | 2020-07-09T13:36:40 | 2020-07-09T13:36:40 | {} | NONE |
ClickHouse | ClickHouse | 655,402,745 | 12,261 | jessicacheun | > Please show how you get that result (exact queries you run), where that 190 Mb number cames from. And you didn't answer other questions above.
select
table as "table name",
sum(rows) as "total lines",
formatReadableSize(sum(data_uncompressed_bytes)) as "origin size",
formatReadableSize(sum(data... | 2020-07-08T09:25:27 | 2020-07-08T09:25:27 | {} | NONE |
ClickHouse | ClickHouse | 655,423,968 | 11,432 | alexeyerm | Reproduced in v20.5.2.7:
```
CREATE TABLE db.tbl ...
ENGINE = ReplicatedReplacingMergeTree('/clickhouse/tables/{shard}/db.tbl', 'r01')
```
Received exception from server (version 20.5.2):
Code: 62. DB::Exception: No macro 'shard' in config while processing substitutions in '/clickhouse/tables/{shard}/db.tbl'... | 2020-07-08T10:06:48 | 2020-07-08T10:06:48 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,479,318 | 12,181 | alexey-milovidov | ```
Code: 243. DB::Exception: Received from 172.19.0.5:9000. DB::Exception: Cannot reserve 1.00 MiB, not enough space. Stack trace:
E
E 0. /build/obj-x86_64-linux-gnu/../contrib/libcxx/include/exception:129: Poco::Exception::Exception(std::__1::basic_string<char, std::__1::char_traits<char>, std... | 2020-07-08T12:10:10 | 2020-07-08T12:10:27 | {} | MEMBER |
ClickHouse | ClickHouse | 655,755,582 | 12,302 | azat | >Description check — Changelog entry is not found
Added `h4` and it cannot find the entry anymore (with `h4` the description looks better, worth updating the parser and template?) | 2020-07-08T21:02:13 | 2020-07-08T21:02:44 | {} | MEMBER |
ClickHouse | ClickHouse | 655,769,959 | 12,299 | CurtizJ | TTL doesn't work at replicated tables at all after renaming a column.
```
DROP TABLE IF EXISTS table_rename_with_ttl;
CREATE TABLE table_rename_with_ttl
(
date1 Date,
value1 String
)
ENGINE = ReplicatedMergeTree('/clickhouse/test/table_rename_with_ttl', '1')
ORDER BY tuple();
INSERT INTO table_renam... | 2020-07-08T21:34:25 | 2020-07-08T21:34:40 | {} | MEMBER |
ClickHouse | ClickHouse | 655,841,580 | 12,103 | alexey-milovidov | > Performance test: 4 faster
Great result! | 2020-07-09T01:28:19 | 2020-07-09T01:28:19 | {} | MEMBER |
ClickHouse | ClickHouse | 654,412,962 | 10,461 | brenoarosa | Hi,
I'm giving some updates since there were relevant merges in the most recent version (20.5.2.7).
Since then my table raised from 900M rows to 2200M rows, its size as compacted parquet files raised from 20GB to 30GB, still split between 200 files.
I'm running this in the same server with 8 cores and 32GB.
Even ... | 2020-07-06T19:04:43 | 2020-07-09T13:30:39 | {
"+1": 4
} | NONE |
ClickHouse | ClickHouse | 656,283,452 | 12,322 | den-crane | >cluster with 3 nodes
Do you mean 1shard with 3 replicas or you mean cross-replication (circle replication)? | 2020-07-09T18:29:44 | 2020-07-09T18:29:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,422,105 | 12,286 | gbakthavatchalam | @filimonov thank you. I will ensure it going forward. | 2020-07-08T10:03:20 | 2020-07-08T10:03:20 | {} | NONE |
ClickHouse | ClickHouse | 655,432,605 | 10,697 | alesapin | Performance seems not related to changes. | 2020-07-08T10:25:41 | 2020-07-08T10:25:41 | {} | MEMBER |
ClickHouse | ClickHouse | 655,475,801 | 12,287 | alesapin | Okay, we found that tests with `DatabaseAtomic` and `polymorphic parts enabled` actually run without these settings. Broken since https://github.com/ClickHouse/ClickHouse/pull/11302.
In logs:
```
/bin/sh: 1: [[: not found
/bin/sh: 1: [[: not found
/bin/sh: 1: [[: not found
``` | 2020-07-08T12:02:57 | 2020-07-08T12:03:53 | {} | MEMBER |
ClickHouse | ClickHouse | 655,791,889 | 12,109 | alexey-milovidov | @CurtizJ Need to check if it works for functions with zero arguments (`count`). | 2020-07-08T22:33:23 | 2020-07-08T22:33:23 | {} | MEMBER |
ClickHouse | ClickHouse | 655,818,944 | 12,305 | alexey-milovidov | Also briefly checked all other usages of `->filter(`. | 2020-07-09T00:03:35 | 2020-07-09T00:03:35 | {} | MEMBER |
ClickHouse | ClickHouse | 655,064,400 | 11,116 | excitoon | Superseded by <s>#12270</s>#11883 | 2020-07-07T19:08:00 | 2020-07-09T06:48:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,941,898 | 12,026 | excitoon | I moved `min_multi_part_upload_size` to disk configuration, it may be useful for performance tuning of different disks depending on cluster location. | 2020-07-09T07:03:00 | 2020-07-09T07:03:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 656,098,987 | 12,329 | realyota | It's not a bug, after you aliased the column it is used globally. Have a look at: https://clickhouse.tech/docs/en/sql-reference/syntax/#syntax-expression_aliases | 2020-07-09T12:30:33 | 2020-07-09T12:30:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,988,909 | 12,314 | bohutang | > It's Ok to have `database` as a case-insensitive synonim to `currentDatabase` function.
Updated. | 2020-07-09T08:31:24 | 2020-07-09T14:29:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 656,175,918 | 12,336 | alexey-milovidov | We were doing SELECT from replica just after INSERT without quorum.
But replication is quick enough (faster than tens of ms needed to run clickhouse-client) and the race condition in test only reproduces once in a few days. | 2020-07-09T14:53:33 | 2020-07-09T14:53:45 | {} | MEMBER |
ClickHouse | ClickHouse | 656,176,175 | 12,308 | alexey-milovidov | #12336 | 2020-07-09T14:54:02 | 2020-07-09T14:54:02 | {} | MEMBER |
ClickHouse | ClickHouse | 655,410,862 | 11,867 | igla123 | Hi guys,
Any news about this topic?
Thanks! | 2020-07-08T09:40:59 | 2020-07-08T09:40:59 | {} | NONE |
ClickHouse | ClickHouse | 655,406,930 | 12,286 | filimonov | > Attached the PDF
Why not paste the queries in the ticket itself? :)
```sql
create table t1 (id UInt32, name String) engine=MergeTree order by id;
create table t2 (id UInt32, name String) engine=MergeTree order by id;
insert into t1 (id, name) values (1, 'a'), (2, 'b'), (3, 'c');
insert into t2 (id, name)... | 2020-07-08T09:33:19 | 2020-07-08T09:45:30 | {
"heart": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,415,625 | 12,261 | jessicacheun | `CREATE TABLE test
(
`ext1` UInt16,
`ext2` AggregateFunction(groupBitmap, UInt32),
`ds` UInt32
)
ENGINE = AggregatingMergeTree
PARTITION BY (ds, ext1)
ORDER BY ext1
SETTINGS index_granularity = 8192
Ok.
0 rows in set. Elapsed: 0.002 sec.
INSERT INTO test SELECT
toUInt16(1) AS ext1... | 2020-07-08T09:50:29 | 2020-07-08T09:51:20 | {} | NONE |
ClickHouse | ClickHouse | 655,490,934 | 2,451 | vitlibar | > > The REVOKE SELECT on system.* does not work (maybe some tables are required to be visible).
>
> @vitlibar Why not?
The `REVOKE SELECT on system.*` command does not work because it's granted implictly. And it's granted implicitly in order to use those "SHOW *" commands to control access. | 2020-07-08T12:33:51 | 2020-07-08T12:33:51 | {} | MEMBER |
ClickHouse | ClickHouse | 655,808,191 | 12,172 | alexey-milovidov | `SELECT [1, NULL][toNullable(1)]`
`SELECT [toNullable(1)][toNullable(1)]`
`SELECT [NULL][toNullable(1)]` | 2020-07-08T23:24:43 | 2020-07-08T23:26:04 | {} | MEMBER |
ClickHouse | ClickHouse | 655,487,320 | 2,451 | vitlibar | There are privileges `SHOW_COLUMNS`, `SHOW_TABLES`, `SHOW_DATABASES` which control visibility of columns, tables, and databases to user. For example, if you don't have `SHOW_COLUMNS` privilege for some column that column isn't displayed in the `system.columns` table for you.
The SHOW_COLUMNS privilege is special: it... | 2020-07-08T12:25:58 | 2020-07-08T12:26:48 | {} | MEMBER |
ClickHouse | ClickHouse | 655,562,950 | 12,295 | filimonov | Hm...
https://github.com/mfontanini/cppkafka/blob/2b66fd3a22a0b33458a27029df55cd8bca10cc7c/include/cppkafka/utils/backoff_committer.h#L132-L147
| 2020-07-08T14:41:36 | 2020-07-08T14:42:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,590,050 | 9,075 | nikitamikhaylov | > @nikitamikhaylov @Akazz hi, already asked in Telegram channel, but they recommend me to ask here. New version of the ClickHouse Copier has started spawning series of tables which seems pretty weird:
>
>
>
> │ tbl_local_piece_1 │
>
> │ tbl_local_piece_2 │
>
> │ tbl_local_piece_3 │
>
> │ tbl... | 2020-07-08T15:27:15 | 2020-07-08T15:27:15 | {} | MEMBER |
ClickHouse | ClickHouse | 655,622,306 | 9,075 | nikitamikhaylov | > Thanks, I'm using one copier of course.
>
> Yes, but we had a similar option:
> <max_workers>4</max_workers>
> for that. As far I understood workers = number of import threads, ideally equal to number of shards.
No, this setting is needed to control the number of copier processes for each task. A bit useless.... | 2020-07-08T16:25:15 | 2020-07-08T16:25:15 | {} | MEMBER |
ClickHouse | ClickHouse | 655,445,197 | 12,261 | jessicacheun | > 19.5.2.6 is too old.
> The serialization of groupBitmapState has been adjusted, additionally, previous aggregation data format requires to be escaped with Base64 encoding, which leads to remarkable overhead.
Done!!!!THX!!! | 2020-07-08T10:56:01 | 2020-07-08T11:04:42 | {} | NONE |
ClickHouse | ClickHouse | 655,488,412 | 2,451 | vitlibar | Also there is privilege `SHOW USERS` which controls visibility of user names and grants to other users. | 2020-07-08T12:28:19 | 2020-07-08T12:38:53 | {} | MEMBER |
ClickHouse | ClickHouse | 655,586,731 | 9,075 | miggi | @nikitamikhaylov @Akazz hi, already asked in Telegram channel, but they recommend me to ask here. New version of the ClickHouse Copier has started spawning series of tables which seems pretty weird:
│ tbl_local_piece_1 │
│ tbl_local_piece_2 │
│ tbl_local_piece_3 │
│ tbl_local_piece_4 ... | 2020-07-08T15:21:30 | 2020-07-08T15:21:30 | {} | NONE |
ClickHouse | ClickHouse | 655,595,539 | 9,075 | miggi | And.. unfortunately as I see the speed of import ~3x times slower that might be on one-node insert-from remote(..). I had to use non-copier version now to migrate data | 2020-07-08T15:36:34 | 2020-07-08T15:36:34 | {} | NONE |
ClickHouse | ClickHouse | 655,652,948 | 9,075 | miggi | Alright thanks. I use default behaviour with number_of_splits = 10, but the clickhouse-copier seems to be too slow comparing to previous version which I used a lot... and now I had to use now insert into with screen session.
Also I've cleaned up all the tables manually and had to stop this.
That's seems to m... | 2020-07-08T17:25:21 | 2020-07-08T17:25:21 | {} | NONE |
ClickHouse | ClickHouse | 655,419,396 | 12,261 | jessicacheun | > ```
> CREATE TABLE test
> (
> `ext1` UInt16,
> `ext2` AggregateFunction(groupBitmap, UInt32),
> `ds` UInt32
> )
> ENGINE = AggregatingMergeTree
> PARTITION BY (ds, ext1)
> ORDER BY ext1
> SETTINGS index_granularity = 8192
>
> Ok.
>
> 0 rows in set. Elapsed: 0.008 sec.
>
> INSERT INTO t... | 2020-07-08T09:57:55 | 2020-07-08T09:57:55 | {} | NONE |
ClickHouse | ClickHouse | 655,434,203 | 12,261 | yingfeng | 19.5.2.6 is too old.
The serialization of groupBitmapState has been adjusted, additionally, previous aggregation data format requires to be escaped with Base64 encoding, which leads to remarkable overhead. | 2020-07-08T10:29:35 | 2020-07-08T10:29:35 | {} | NONE |
ClickHouse | ClickHouse | 655,478,143 | 11,867 | igla123 | Thanks for the info! Unfortunately I group by several columns, so I need rollup(
Any plans for development this feature for WITH ROLLUP? | 2020-07-08T12:07:41 | 2020-07-08T12:07:41 | {} | NONE |
ClickHouse | ClickHouse | 655,662,831 | 12,300 | nvartolomei | @alex-zaitsev is this really an option for log capturing including crashes? | 2020-07-08T17:45:09 | 2020-07-08T17:45:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,698,271 | 12,269 | inakivb | Thanks @vitlibar !
| 2020-07-08T19:00:16 | 2020-07-08T19:00:16 | {} | NONE |
ClickHouse | ClickHouse | 655,792,356 | 12,109 | alexey-milovidov | Also what shall we do with a function to initialize empty aggregation state with given types of arguments?
If we will have a sepate function for that purpose, what will be the name of this function? | 2020-07-08T22:34:45 | 2020-07-08T22:34:45 | {} | MEMBER |
ClickHouse | ClickHouse | 655,877,204 | 11,815 | alexey-milovidov | multiSearch, MSan:
https://clickhouse-test-reports.s3.yandex.net/12292/eaa13d81769af71d6c620881e7064ff47285c9aa/stress_test_(memory).html#fail1
(something related to Hyperscan?) | 2020-07-09T03:29:48 | 2020-07-09T03:29:48 | {} | MEMBER |
ClickHouse | ClickHouse | 656,362,558 | 12,372 | s-mx | #10429 | 2020-07-09T21:34:42 | 2020-07-09T21:34:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,403,958 | 12,261 | jessicacheun | > 1. Why ReplacingMergeTree not AggregatingMergeTree?
> 2. How do you examine the size? Do you do optimize before, how many states (i.e. different ds,ext1 combinations) do your table store?
>
> ```
> clickhouse-local -q 'select groupBitmapState(number) as state from numbers(50000000)' | clickhouse-local --structu... | 2020-07-08T09:27:39 | 2020-07-08T09:27:39 | {} | NONE |
ClickHouse | ClickHouse | 655,820,415 | 12,110 | alexey-milovidov | I can get logical error with this query:
```
CREATE TABLE delta_codec_synthetic
(
`id` UInt64 NULL CODEC(Delta, ZSTD(22))
)
ENGINE = MergeTree()
ORDER BY tuple()
Received exception from server (version 20.6.1):
Code: 49. DB::Exception: Received from localhost:9000. DB::Exception: Value of type Nullable... | 2020-07-09T00:08:33 | 2020-07-09T00:08:33 | {} | MEMBER |
ClickHouse | ClickHouse | 656,170,206 | 12,294 | alesapin | Why it's so slow? Why tests failed? | 2020-07-09T14:43:42 | 2020-07-09T14:43:42 | {} | MEMBER |
ClickHouse | ClickHouse | 656,178,830 | 11,811 | 4ertus2 | Closed in favour of #12337 | 2020-07-09T14:58:35 | 2020-07-09T14:58:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,591,639 | 9,075 | miggi | Cool, which parameter I should use for that ? And probably subjective, but currently copier works significantly slower that "old one" insert from remote(..) | 2020-07-08T15:29:47 | 2020-07-08T15:29:47 | {} | NONE |
ClickHouse | ClickHouse | 654,032,327 | 11,815 | alexey-milovidov | 00721_force_by_identical_result_after_merge_zookeeper: not fixed.
Obvious race condition: mutation vs. OPTIMIZE.
Upd. Most likely fixed in #12157 | 2020-07-06T06:01:26 | 2020-07-09T03:30:23 | {} | MEMBER |
ClickHouse | ClickHouse | 655,897,107 | 12,313 | alexey-milovidov | TSan report, fix: #12315 | 2020-07-09T04:47:22 | 2020-07-09T04:47:22 | {} | MEMBER |
ClickHouse | ClickHouse | 656,354,279 | 12,277 | bobrik | My first version was to allow a prefix of a PK to engage this. E.g. if PK is `(a, b, c)`, then the following is allowed:
* `(a, b, c)`
* `(a, b)`
* `(a)`
But the following isn't:
* `(a, c)`
* `(b, c)`
* `(b)`
* `(c)`
Is that what you mean? If so, I can totally make it happen. | 2020-07-09T21:13:27 | 2020-07-09T21:13:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,436,326 | 12,269 | vitlibar | There was an error while that dictionary was loading:
```
Unknown field found in CSV header: 'network' at position 0
Set the 'input_format_skip_unknown_fields' parameter explicitly to ignore and proceed
```
So you have to rename the attribute `prefix` to `network`.
It's bad that the error isn't actually dis... | 2020-07-08T10:34:32 | 2020-07-08T10:44:16 | {} | MEMBER |
ClickHouse | ClickHouse | 655,461,929 | 2,451 | alexey-milovidov | > The REVOKE SELECT on system.* does not work (maybe some tables are required to be visible).
@vitlibar Why not? | 2020-07-08T11:33:08 | 2020-07-08T11:33:08 | {} | MEMBER |
ClickHouse | ClickHouse | 656,046,350 | 12,076 | filimonov | Sorry, it looks like I will not have time for finishing that this week, and next 2 weeks I'm on vacation, so if someone wants to pick it in the meantime - it's ok for me (i can back to that in 3-4 weeks).
Some general thoughts:
1) realtime - clickhouse will have higher priority for accessing the disk, it's risk-pr... | 2020-07-09T10:26:31 | 2020-07-09T12:33:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 656,143,934 | 12,293 | zhang2014 | https://github.com/ClickHouse/ClickHouse/blob/master/src/Interpreters/PredicateRewriteVisitor.cpp#L88
Is expected, but i thought maybe we could remove it. | 2020-07-09T13:58:08 | 2020-07-09T13:58:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 656,272,140 | 12,329 | den-crane | @hiliuxg No it's against standard sql. CH feature.
pssql
```
select a, t.a, a/2 as a, a from (select 3 a) as t
a | a | a | a
---+---+---+---
3 | 3 | 1 | 3
``` | 2020-07-09T18:04:59 | 2020-07-09T18:04:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,519,533 | 10,510 | akuzm | @Quid37 do you plan to continue working on this?
I think distributed tracing may be useful to debug problems in large cluster, related to communication between different server. This means we should pay most of attention to the events related to inter-server communication. This includes queries made during distribut... | 2020-07-08T13:27:31 | 2020-07-08T13:27:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,799,720 | 12,300 | alexey-milovidov | > @alex-zaitsev is this really an option for log capturing including crashes?
Yes. ClickHouse will terminate in 10 seconds after crash* and system logs are flushed every 7 seconds. Most likely logs before crash will be written successfully.
\* kill -9, stack overflow and kernel-side general protection fault are n... | 2020-07-08T22:57:09 | 2020-07-08T22:57:09 | {} | MEMBER |
ClickHouse | ClickHouse | 655,812,403 | 12,162 | alexey-milovidov | Duplicate of #11362
Fixed in #12021
Cannot reproduce with
```
clickhouse-benchmark --extremes 1 <<< "SELECT if(number % 2, toFixedString(toString(number), 2), toFixedString(toString(-number), 1025)) FROM system.numbers LIMIT 10"
``` | 2020-07-08T23:38:54 | 2020-07-08T23:38:54 | {} | MEMBER |
ClickHouse | ClickHouse | 655,877,848 | 12,292 | alexey-milovidov | > Stress test (memory)
https://github.com/ClickHouse/ClickHouse/issues/11815#issuecomment-655877204 | 2020-07-09T03:31:57 | 2020-07-09T03:31:57 | {} | MEMBER |
ClickHouse | ClickHouse | 656,140,128 | 12,329 | den-crane | Use table specification if you need aliased column
select a, **t.a**, a/2 as a, a from (select 3 a) as t
```
┌───a─┬─a─┬───a─┬───a─┐
│ 1.5 │ 3 │ 1.5 │ 1.5 │
└─────┴───┴─────┴─────┘
``` | 2020-07-09T13:51:25 | 2020-07-09T13:51:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 656,286,589 | 12,322 | wanghaifen | Thanks Denis：
the mean is 1shard with 1replicas，why set this，because the cluster is test cluster
| |
18101350634
|
|
邮箱：18101350634@163.com
|
Signature is customized by Netease Mail Master
On 07/10/2020 02:29, Denis Zhuravlev wrote:
cluster with 3 nodes
Do you mean 1shard with 3 replicas or you mean cross-repli... | 2020-07-09T18:36:39 | 2020-07-09T18:36:39 | {} | NONE |
ClickHouse | ClickHouse | 655,456,625 | 2,451 | sushiMix | in 20.5 I have seen some new improvements on user management but system tables can still be requested by any user. You can potentially use the ROW policy on each table to filter the SELECT result for system tables but need to do it for each table (and need to track new tables on each upgrade). The REVOKE SELECT on syst... | 2020-07-08T11:22:10 | 2020-07-08T11:23:30 | {} | NONE |
ClickHouse | ClickHouse | 655,493,598 | 12,181 | azat | >DB::Exception: Cannot reserve 1.00 MiB
FWIW there is ENOSPC error from rabbitmq itself (which is not passed with bind mount from host, like /var/lib/clickhouse):
>rabbitmq1_1 | 2020-07-08 01:28:46.381 [info] <0.356.0> Free disk space is insufficient. Free bytes: 46157824. Limit: 50000000
>rabbitmq1_1 | Mnesia... | 2020-07-08T12:39:46 | 2020-07-08T12:39:46 | {} | MEMBER |
ClickHouse | ClickHouse | 655,797,170 | 12,178 | alexey-milovidov | > Performance — 1 faster, 4 slower, 190 unstable
Because there was performance improvement in master yesterday in #12196 | 2020-07-08T22:49:01 | 2020-07-08T22:49:01 | {} | MEMBER |
ClickHouse | ClickHouse | 655,841,895 | 12,179 | alexey-milovidov | > Yandex synchronization check (only for Yandex employees)
Will fix after merge. | 2020-07-09T01:29:21 | 2020-07-09T01:29:21 | {} | MEMBER |
ClickHouse | ClickHouse | 656,338,942 | 12,322 | den-crane | BTW I don't understand is it work OK or not?
```
single node drop:
ALTER TABLE market.vb_xue_order_sku_user_lesson_detail_5_minutes_replica DROP
PARTITION (20200709, 168);
``` | 2020-07-09T20:36:59 | 2020-07-09T20:37:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 656,345,264 | 12,302 | filimonov | Here
https://github.com/ClickHouse/ClickHouse/blob/7d28a4cb1558a0b188f77609b417581022a22e4c/src/Storages/Kafka/KafkaBlockInputStream.cpp#L139-L153
or here
https://github.com/ClickHouse/ClickHouse/blob/7d28a4cb1558a0b188f77609b417581022a22e4c/src/Storages/Kafka/KafkaBlockInputStream.cpp#L204-L207
If would be... | 2020-07-09T20:51:54 | 2020-07-09T20:54:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,457,080 | 12,291 | Slach | @CurtizJ could you look at this issue? | 2020-07-08T11:23:06 | 2020-07-08T11:23:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,475,648 | 11,867 | alexey-milovidov | `totals_mode` affects only WITH TOTALS, not WITH ROLLUP / WITH CUBE. | 2020-07-08T12:02:32 | 2020-07-08T12:02:32 | {} | MEMBER |
ClickHouse | ClickHouse | 655,492,900 | 2,451 | vitlibar | I think I'll provide implicit access only to few system tables, not all of them, or maybe implement more `SHOW <something>` privileges. | 2020-07-08T12:38:09 | 2020-07-08T12:43:38 | {} | MEMBER |
ClickHouse | ClickHouse | 655,597,368 | 9,075 | nikitamikhaylov | > Cool, which parameter I should use for that ? And probably subjective, but currently copier works significantly slower that "old one" insert from remote(..)
`<number_of_splits>1</number_of_splits>` -- just add it to the copier's task.
How many copiers do you execute? There is no need to use only one process.
An... | 2020-07-08T15:39:37 | 2020-07-08T15:39:37 | {} | MEMBER |
ClickHouse | ClickHouse | 655,620,604 | 12,300 | alex-zaitsev | @nvartolomei , you can use system.text_log and any ClickHouse output format of your choice, including JSONs. Only a pull adapter is needed | 2020-07-08T16:21:53 | 2020-07-08T16:21:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,633,001 | 12,111 | akuzm | > Please link all the issues found by this fuzzer to let people know how good this tool is.
Added the list of issues to the changelog entry.
The tests are green and I cleaned up the code, so merging now. Next step is to start running it in the CI. | 2020-07-08T16:46:07 | 2020-07-08T16:46:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 655,768,862 | 12,293 | alexey-milovidov | CC @zhang2014 | 2020-07-08T21:31:46 | 2020-07-08T21:31:46 | {} | MEMBER |
ClickHouse | ClickHouse | 655,819,880 | 12,110 | alexey-milovidov | I have different (and correct) exception message:
```
CREATE TABLE delta_codec_synthetic
(
`id` UInt64 CODEC(Delta, ZSTD(1048577))
)
ENGINE = MergeTree()
ORDER BY tuple()
Received exception from server (version 20.6.1):
Code: 433. DB::Exception: Received from localhost:9000. DB::Exception: ZSTD codec c... | 2020-07-09T00:06:41 | 2020-07-09T00:06:41 | {} | MEMBER |
ClickHouse | ClickHouse | 633,495,249 | 11,115 | alesapin | Failures expected (we are trying to fix CI). | 2020-05-25T10:09:52 | 2020-05-25T10:09:52 | {} | MEMBER |
ClickHouse | ClickHouse | 633,726,508 | 11,149 | filimonov | > I would also add details about the performance tests (and notes from the PR description) into the commit itself
And also maybe it worth updating title of the PR so that it includes the possible "performance" benefits.
Changelog entry contains that. Commit messages - discussed on telegram (it's a nice practice, bu... | 2020-05-25T21:56:02 | 2020-05-25T22:58:32 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.