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 | 813,920,080 | 22,669 | azat | > Fatal reported #22689
Should be fixed by #22592 | 2021-04-06T08:07:11 | 2021-04-06T08:07:23 | {} | MEMBER |
ClickHouse | ClickHouse | 813,972,817 | 21,509 | alesapin | All failures not related to changes. | 2021-04-06T09:25:49 | 2021-04-06T09:25:49 | {} | MEMBER |
ClickHouse | ClickHouse | 813,992,481 | 22,690 | akuzm | Doesn't reproduce for me either. Let's keep it open for a while, maybe it reproduces again. | 2021-04-06T09:56:25 | 2021-04-06T09:56:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 814,055,304 | 22,646 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/22703 | 2021-04-06T11:46:28 | 2021-04-06T11:46:28 | {} | MEMBER |
ClickHouse | ClickHouse | 813,025,867 | 22,571 | alexey-milovidov | > Functional stateless tests (release) — fail: 1, passed: 2941
Interference in parallel test run, @vitlibar | 2021-04-04T12:36:42 | 2021-04-04T12:36:42 | {} | MEMBER |
ClickHouse | ClickHouse | 813,026,045 | 22,565 | alexey-milovidov | #22587 | 2021-04-04T12:38:05 | 2021-04-04T12:38:05 | {} | MEMBER |
ClickHouse | ClickHouse | 813,128,441 | 22,628 | alexey-milovidov | > 01730_distributed_group_by_no_merge_order_by_long
@vitlibar | 2021-04-05T00:49:08 | 2021-04-05T00:49:08 | {} | MEMBER |
ClickHouse | ClickHouse | 813,177,403 | 22,596 | alexey-milovidov | > https://clickhouse-test-reports.s3.yandex.net/22596/08f129c0d3a3fa9d990c2fd9b2cb1c770661c596/functional_stateless_tests_(address)/test_run.txt.out.log
@tavplubix Should we add "Retry later" to the list of retryable errors in clickhouse-test? | 2021-04-05T04:40:18 | 2021-04-05T04:40:18 | {} | MEMBER |
ClickHouse | ClickHouse | 813,281,751 | 14,298 | mastertheknife | > https://presentations.clickhouse.tech/cpp_russia_2020/index.html
P.S
I spotted this yesterday, i can correlate with some of the things mentioned in this article:
https://blogs.gentoo.org/mgorny/2021/02/19/the-modern-packagers-security-nightmare/
Especially the ClickHouse-Extras/* repositories that aren't forked... | 2021-04-05T08:49:56 | 2021-04-05T08:49:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,349,866 | 22,217 | tavplubix | Functional stateful tests (release, DatabaseReplicated) - `All connection tries failed while connecting to ZooKeeper. nodes: [::1]:29181, [::1]:9181, [::1]:19181`
Functional stateless tests (release, wide parts enabled) - `Session expired`
Performance - unrelated
Yandex synchronization check - was broken in master
... | 2021-04-05T11:32:13 | 2021-04-05T11:32:13 | {} | MEMBER |
ClickHouse | ClickHouse | 814,004,695 | 22,696 | alexey-milovidov | So, you need first to convert your strings to a sequence of bytes in UTF-8 encoding and then pass to CityHash64 function. | 2021-04-06T10:16:27 | 2021-04-06T10:16:27 | {} | MEMBER |
ClickHouse | ClickHouse | 814,048,980 | 19,766 | ghost | @vdimir
> We got with and NULL. But why not wish for id = 1, like
The exceptional case in flow analysis is that:
The number of people decreases from `home` -> `giftbox` but, the number of people increase at `wish`.
```
home (2 people) -> home (1 person)
-> giftbox (1 person) -> wish (2 pe... | 2021-04-06T11:35:04 | 2021-04-06T11:48:48 | {} | NONE |
ClickHouse | ClickHouse | 813,201,589 | 22,583 | azat | >[AST fuzzer (debug) — Logical error: 'Logical error: no information about file column%2Ename.size0 in StorageLog'.](https://clickhouse-test-reports.s3.yandex.net/22583/69296876005c0fa171c755f8b224e4d58192c402/fuzzer_debug/report.html#fail1)
#22654 | 2021-04-05T05:50:44 | 2021-04-05T05:50:44 | {} | MEMBER |
ClickHouse | ClickHouse | 813,650,107 | 22,610 | lehasm | I will revert edits soon... | 2021-04-05T21:06:30 | 2021-04-05T21:06:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,911,632 | 22,647 | azat | >I'll check if it works when we set project name. Or just kill/rm containers and filter by project manually.
There is already `--project-name` (even default, based on dirname, is ok) - https://github.com/ClickHouse/ClickHouse/blob/ae883829efacd726916fad851fd178f1b32e0a02/tests/integration/helpers/cluster.py#L130 | 2021-04-06T07:54:46 | 2021-04-06T07:54:46 | {} | MEMBER |
ClickHouse | ClickHouse | 813,149,601 | 16,518 | alexey-milovidov | Test 1:
```
perl -e 'print "x"x100000'; echo -ne '\033[0;0H'; clickhouse-client
```
Current behaviour:

| 2021-04-05T02:36:10 | 2021-04-05T02:36:10 | {} | MEMBER |
ClickHouse | ClickHouse | 813,982,871 | 22,514 | alesapin | need to support alter setting | 2021-04-06T09:41:18 | 2021-04-06T09:41:18 | {} | MEMBER |
ClickHouse | ClickHouse | 814,019,342 | 19,766 | vdimir | @achimbab
If we consider exception case from https://github.com/ClickHouse/ClickHouse/pull/19766#issuecomment-795175943
```
CREATE TABLE foo (timestamp DateTime, id Int32, page String) ENGINE=TinyLog;
INSERT INTO foo VALUES (1, 1, 'home'), (2, 1, 'home'), (3, 1, 'giftbox'), (4, 1, 'wish'), (1, 2, 'home'),... | 2021-04-06T10:42:33 | 2021-04-06T10:51:36 | {} | MEMBER |
ClickHouse | ClickHouse | 813,249,536 | 22,414 | ilejn | Does it make sense?
Feedback is appreciated. | 2021-04-05T07:39:10 | 2021-04-05T07:39:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,292,587 | 22,514 | alesapin | Ok, everything looks expected. Integration s3 merge tree test depends on the order. Func tests just insert equal data. It's especially not obvious for `01197_summing_enum` test. | 2021-04-05T09:10:24 | 2021-04-05T09:10:24 | {} | MEMBER |
ClickHouse | ClickHouse | 813,556,604 | 22,239 | MaxJPowers | @kitaisreal Thanks you for explaining the underlying mechanisms of cache dictionaries. | 2021-04-05T18:21:59 | 2021-04-05T18:21:59 | {} | NONE |
ClickHouse | ClickHouse | 813,865,782 | 21,326 | azat | Also with this patch the mutations will hang for non existing parts (since it will return nothing from selectPartsToMutate() and so nothing will be notified), i.e.:
```sql
CREATE TABLE test_mutations_in_partitions
(
`date` Date,
`a` UInt64,
`b` String
)
ENGINE = MergeTree
PARTITION BY a
ORDER BY... | 2021-04-06T06:39:30 | 2021-04-06T06:39:30 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 813,914,128 | 22,669 | alesapin | Fatal reported https://github.com/ClickHouse/ClickHouse/issues/22689 | 2021-04-06T07:58:40 | 2021-04-06T07:58:40 | {} | MEMBER |
ClickHouse | ClickHouse | 813,916,825 | 18,956 | gersmann | @UnamedRus very nice, great update! | 2021-04-06T08:02:31 | 2021-04-06T08:02:31 | {} | NONE |
ClickHouse | ClickHouse | 813,014,613 | 22,594 | excitoon | Related: #21852 | 2021-04-04T11:08:45 | 2021-04-04T11:08:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,918,622 | 22,325 | azat | https://clickhouse-test-reports.s3.yandex.net/22325/7ccd784acb5554b6f39f4fdc72028becf5dfee55/stress_test_(memory).html#fail1 - #22592 | 2021-04-06T08:05:03 | 2021-04-06T08:05:03 | {} | MEMBER |
ClickHouse | ClickHouse | 813,023,646 | 22,596 | alexey-milovidov | `00909_ngram_distance`
`00951_ngram_search`
\- gives different result.
`00947_ml_test`
\- rounding errors? | 2021-04-04T12:19:36 | 2021-04-04T12:19:36 | {} | MEMBER |
ClickHouse | ClickHouse | 813,026,836 | 22,587 | alexey-milovidov | > Functional stateless tests (release) — fail: 1, passed: 2941
@vitlibar | 2021-04-04T12:44:26 | 2021-04-04T12:44:26 | {} | MEMBER |
ClickHouse | ClickHouse | 813,030,102 | 22,373 | alesapin | One more time | 2021-04-04T13:08:06 | 2021-04-04T13:08:06 | {} | MEMBER |
ClickHouse | ClickHouse | 813,243,786 | 22,594 | excitoon | Failed tests are unrelated. | 2021-04-05T07:26:06 | 2021-04-05T07:26:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,341,768 | 22,661 | CurtizJ | `01360_materialized_view_with_join_on_query_log` - `DB::Exception: Timeout exceeded (60 s) while flushing system log 'DB::QueryLog'. ... (in query: SYSTEM FLUSH LOGS)` | 2021-04-05T11:08:07 | 2021-04-05T11:08:22 | {} | MEMBER |
ClickHouse | ClickHouse | 813,409,241 | 22,239 | kitaisreal | CacheDictionary size_in_cells means maximum keys size it can store, for example if size_in_cells is 10000 cache dictionary maximum can store 10000 keys with unlimited number of columns.
-What is the distribution for how the elements in the table are stored in the cached dictionary? I'd assume it's a random hash rang... | 2021-04-05T14:03:20 | 2021-04-05T14:03:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,663,837 | 22,654 | CurtizJ | Fail of `01475_read_subcolumns` is suspicious, but it's related to unstable behaviour of tests, that use query log #22666.
1. ```<Error> void DB::SystemLog<DB::QueryLogElement>::flushImpl(const std::vector<LogElement> &, uint64_t) [LogElement = DB::QueryLogElement]: Code: 60, e.displayText() = DB::Exception: Table t... | 2021-04-05T21:33:44 | 2021-04-05T21:34:02 | {} | MEMBER |
ClickHouse | ClickHouse | 813,888,937 | 22,141 | hatrena | The average response from the naive solution is `10 rows in set. Elapsed: 15.929 sec. Processed 204.23 million rows, 20.28 GB (12.82 million rows/s., 1.27 GB/s.)` which is much better than the other solutions somehow. those take minutes. | 2021-04-06T07:19:47 | 2021-04-06T07:19:47 | {} | NONE |
ClickHouse | ClickHouse | 814,005,642 | 22,696 | alexey-milovidov | Looks as simple as `str3.getBytes("UTF-8")`:
https://docs.oracle.com/javase/tutorial/i18n/text/string.html | 2021-04-06T10:18:07 | 2021-04-06T10:18:07 | {} | MEMBER |
ClickHouse | ClickHouse | 813,011,695 | 21,639 | traceon | Great job! Maybe it is worth just disabling jemalloc and rocksdb support in Apple silicon target, and make this PR ready? Those then could be addressed in further iterations, I guess. | 2021-04-04T10:46:53 | 2021-04-04T10:46:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,050,302 | 21,889 | CurtizJ | Yandex synchronization check is broken in master. | 2021-04-04T15:16:49 | 2021-04-04T15:16:49 | {} | MEMBER |
ClickHouse | ClickHouse | 813,177,793 | 22,633 | alexey-milovidov | > Functional stateless tests (release, wide parts enabled) — fail: 1, passed: 2946, skipped: 1
@vitlibar | 2021-04-05T04:41:58 | 2021-04-05T04:41:58 | {} | MEMBER |
ClickHouse | ClickHouse | 813,911,239 | 22,669 | alesapin | Mutation hung, but one of the replicas was not able to finish it https://gist.github.com/alesapin/2681a5f09652431ad6379e54cfd974f3. So it's expected. | 2021-04-06T07:54:12 | 2021-04-06T07:54:12 | {} | MEMBER |
ClickHouse | ClickHouse | 813,101,573 | 22,595 | alexey-milovidov | > 01455_shard_leaf_max_rows_bytes_to_read
@KochetovNicolai | 2021-04-04T21:24:00 | 2021-04-04T21:24:00 | {} | MEMBER |
ClickHouse | ClickHouse | 813,255,195 | 22,514 | alesapin | At least sever didn't crash. I'll check failures. | 2021-04-05T07:51:26 | 2021-04-05T07:51:26 | {} | MEMBER |
ClickHouse | ClickHouse | 814,054,898 | 19,766 | vdimir | > The number of people decreases from `home` -> `giftbox` but, the number of people increase at `wish`.
I considered semantic of this function close to `windowFunnel` with `strict_order`, but with additional features like `forward`/`backward` and `head`/`tail` and returned next event, but not number of matches.
| 2021-04-06T11:45:42 | 2021-04-06T11:45:42 | {} | MEMBER |
ClickHouse | ClickHouse | 814,057,701 | 22,705 | songenjie | > Already exists https://clickhouse.tech/docs/en/sql-reference/statements/alter/partition/#alter_fetch-partition?
it just support fetch partition ,not support fetch part | 2021-04-06T11:50:30 | 2021-04-06T11:50:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,084,085 | 22,626 | UnamedRus | Probably should be not a bug, but a feature. | 2021-04-04T19:04:51 | 2021-04-04T19:04:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,326,377 | 22,661 | CurtizJ | `01526_initial_query_id` - probably bug in `SYSTEM FLUSH LOGS` | 2021-04-05T10:26:45 | 2021-04-05T10:26:45 | {} | MEMBER |
ClickHouse | ClickHouse | 813,326,971 | 22,661 | CurtizJ | `01383_log_broken_table` - `Setting max_memory_usage shouldn't be greater than 6000000`. Constraint is setting in `01294_create_settings_profile`. | 2021-04-05T10:28:29 | 2021-04-05T10:28:37 | {} | MEMBER |
ClickHouse | ClickHouse | 813,340,501 | 18,465 | amosbird | You can try implementing this https://github.com/ClickHouse/ClickHouse/issues/20587 . And it will be accurate if data is distributed by some sharding expression, which is a TODO item of @azat in https://github.com/ClickHouse/ClickHouse/pull/20882 | 2021-04-05T11:05:08 | 2021-04-05T11:05:08 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 813,445,998 | 22,669 | alesapin | It's quite hard to provide a test. | 2021-04-05T15:12:31 | 2021-04-05T15:12:31 | {} | MEMBER |
ClickHouse | ClickHouse | 813,489,229 | 22,661 | alesapin | nasty tests https://github.com/ClickHouse/ClickHouse/pull/22670/files | 2021-04-05T16:33:20 | 2021-04-05T16:33:20 | {} | MEMBER |
ClickHouse | ClickHouse | 814,055,066 | 22,703 | alexey-milovidov | `SELECT -0`
`SELECT +0`
are parsed as
`SELECT 0`
and this is normal (not a bug)
Don't be confused with `SELECT -0.` | 2021-04-06T11:46:00 | 2021-04-06T11:46:17 | {} | MEMBER |
ClickHouse | ClickHouse | 814,057,973 | 22,671 | akuzm | > You have, but probably it's TestKeeper or ClickHouse Keeper, not Apache ZooKeeper.
You're right:
```
<zookeeper>
<implementation>testkeeper</implementation>
</zookeeper>
```
Does it reproduce for you with TestKeeper? | 2021-04-06T11:50:59 | 2021-04-06T11:50:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,025,307 | 22,562 | alexey-milovidov | > E Head https://registry-1.docker.io/v2/yandex/clickhouse-s3-proxy/manifests/latest: EOF
@alesapin | 2021-04-04T12:32:16 | 2021-04-04T12:32:16 | {} | MEMBER |
ClickHouse | ClickHouse | 813,110,113 | 945 | alexey-milovidov |
6e93d93 | 2021-04-04T22:40:35 | 2021-04-04T22:40:35 | {} | MEMBER |
ClickHouse | ClickHouse | 813,123,460 | 22,631 | den-crane | If I remove LowCardinality
```
CREATE TABLE testTSV
(
A Date,
S Nullable(String),
X Int32,
S1 Nullable(String),
S2 Array(String)
)
ENGINE = Memory;
printf '2020-01-01\t\N\t32\t\N\n'|clickhouse-client -q 'insert into testTSV format TSV'
ERROR: Line feed found where tab is expected... | 2021-04-05T00:22:39 | 2021-04-05T00:22:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,628,637 | 22,668 | alexey-milovidov | LGTM. Need to merge with master. | 2021-04-05T20:23:50 | 2021-04-05T20:23:50 | {} | MEMBER |
ClickHouse | ClickHouse | 813,696,061 | 21,888 | rf | Hey, thanks so much for looking into this!
We are a manufacturing analytics company and we have timeseries data that tracks a continuously increasing number. It will increase up to some maximum, then be reset to zero. You can think of it like a spool of wire with a rotary encoder, that is going to count up to some v... | 2021-04-05T22:54:27 | 2021-04-05T22:54:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,914,325 | 22,669 | alesapin | Performance not related to changes. | 2021-04-06T07:58:58 | 2021-04-06T07:58:58 | {} | MEMBER |
ClickHouse | ClickHouse | 813,977,320 | 22,671 | akuzm | > What does "without ZooKeeper" mean?
Not sure :) I just don't have any Zookeeper connection configured.
It reproduces each time for me, the cluster `test_cluster_two_shards_different_databases` is from the stateless tests. | 2021-04-06T09:32:44 | 2021-04-06T09:32:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 814,004,155 | 22,696 | alexey-milovidov | The strings in ClickHouse are presumed to be UTF-8 encoded,
while `str3.getBytes()` in Java may use some different encoding (UTF-16 maybe?) | 2021-04-06T10:15:34 | 2021-04-06T10:15:34 | {} | MEMBER |
ClickHouse | ClickHouse | 814,056,038 | 22,705 | nvartolomei | Already exists https://clickhouse.tech/docs/en/sql-reference/statements/alter/partition/#alter_fetch-partition? | 2021-04-06T11:47:48 | 2021-04-06T11:47:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 814,058,171 | 22,592 | alexey-milovidov | Could you please merge with master? | 2021-04-06T11:51:20 | 2021-04-06T11:51:20 | {} | MEMBER |
ClickHouse | ClickHouse | 813,024,639 | 22,589 | alexey-milovidov | > Functional stateless tests (release) — fail: 1, passed: 2941
@vitlibar | 2021-04-04T12:26:55 | 2021-04-04T12:26:55 | {} | MEMBER |
ClickHouse | ClickHouse | 813,045,118 | 22,599 | kitaisreal | 20.8 not supported. | 2021-04-04T14:43:51 | 2021-04-04T14:43:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,097,340 | 22,585 | alexey-milovidov | Yes, it should not make primary key nullable on INSERT SELECT. | 2021-04-04T20:50:30 | 2021-04-04T20:50:46 | {} | MEMBER |
ClickHouse | ClickHouse | 813,275,034 | 22,514 | alesapin | Support of deduplication for MergeTree in old syntax is quite complicated because we need to serialize part name, which depends on storage partition key, which depends on metadata snapshot. So I'll just ban it. | 2021-04-05T08:34:43 | 2021-04-05T08:34:43 | {} | MEMBER |
ClickHouse | ClickHouse | 813,349,233 | 22,239 | MaxJPowers | @kitaisreal Any advice or thoughts?
| 2021-04-05T11:30:18 | 2021-04-05T11:30:18 | {} | NONE |
ClickHouse | ClickHouse | 813,657,511 | 22,325 | azat | >[ClickHouse build check — 0/11 builds are OK](https://clickhouse-builds.s3.yandex.net/22325/ba238709d6a1992ea21d8ad25804c6222a5d0e49/clickhouse_build_check/report.html)
Some CI issues, rebased against upstream (w/ your typo fix), previous HEAD is https://github.com/ClickHouse/ClickHouse/pull/22325/commits/e86469647... | 2021-04-05T21:20:45 | 2021-04-05T21:20:45 | {} | MEMBER |
ClickHouse | ClickHouse | 814,056,752 | 22,671 | tavplubix | > > What does "without ZooKeeper" mean?
>
> Not sure :) I just don't have any Zookeeper connection configured.
>
> It reproduces each time for me, the cluster `test_cluster_two_shards_different_databases` is from the stateless tests.
You have, but probably it's TestKeeper or ClickHouse Keeper, not Apache ZooKe... | 2021-04-06T11:48:56 | 2021-04-06T11:48:56 | {} | MEMBER |
ClickHouse | ClickHouse | 814,060,878 | 22,671 | tavplubix | Yes, it reproduces with TestKeeper if table name is not qualified. | 2021-04-06T11:56:05 | 2021-04-06T11:56:05 | {} | MEMBER |
ClickHouse | ClickHouse | 813,042,533 | 22,551 | kitaisreal | There is no issue in 20.8 because parse option null_as_default is by default false. | 2021-04-04T14:27:16 | 2021-04-04T14:27:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,064,139 | 22,516 | taiwhite | @Jokser
I did try this option. The bug is present in both 21.2 and 21.4, when zero-copy is enabled and disabled. | 2021-04-04T16:52:41 | 2021-04-04T16:52:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,103,270 | 22,595 | alexey-milovidov | > 00816_long_concurrent_alter_column
#22628 | 2021-04-04T21:38:19 | 2021-04-04T21:38:19 | {} | MEMBER |
ClickHouse | ClickHouse | 813,327,177 | 22,325 | vdimir | > > [Yandex synchronization check (only for Yandex employees)](https://a.yandex-team.ru/review/1712540)
> > [Yandex third-party checks (only for Yandex employees)](https://a.yandex-team.ru/review/1712540)
>
> Any clue?
Same problem with master | 2021-04-05T10:29:05 | 2021-04-05T10:29:05 | {} | MEMBER |
ClickHouse | ClickHouse | 813,419,436 | 14,298 | alexey-milovidov | It's not always possible to fix the upstream. Sometimes they just ignore patches.
Examples:
https://bugs.llvm.org/show_bug.cgi?id=48186
https://issues.apache.org/jira/browse/AVRO-2717?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17314187#comment-17314187
(they don't un... | 2021-04-05T14:23:59 | 2021-04-05T14:23:59 | {} | MEMBER |
ClickHouse | ClickHouse | 813,425,048 | 22,141 | alexey-milovidov | Naive solution is:
```
SELECT click_ip, count() AS c FROM table GROUP BY click_ip ORDER BY c DESC LIMIT 10
```
It may not fit in memory, but you can raise `max_memory_usage` and enable `max_bytes_before_external_group_by`.
Then the query will be able to proceed even with multiple billion records. But it can be... | 2021-04-05T14:32:54 | 2021-04-05T14:32:54 | {} | MEMBER |
ClickHouse | ClickHouse | 813,467,484 | 21,831 | alesapin | Hi @jorisgio, thank you for the reports. Can you provide a full trace of all threads and server logs when some query hangs forever? Currently, it's hard to guess. In both cases, it looks like clickhouse just waits for some background operation or query. | 2021-04-05T15:53:35 | 2021-04-05T15:53:51 | {} | MEMBER |
ClickHouse | ClickHouse | 813,519,318 | 22,670 | akuzm | Looks like we didn't notice them in Fast Test because they are "long". | 2021-04-05T17:26:25 | 2021-04-05T17:26:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,573,971 | 22,520 | stergiotis | @kitaisreal The tuple based API looks good, thanks!
I just wanted to highlight that sipHash is optimised for short messages, thus the overhead of unpacking the key matters. | 2021-04-05T18:52:39 | 2021-04-05T18:52:39 | {} | NONE |
ClickHouse | ClickHouse | 813,828,234 | 21,059 | azat | Refs: #21059 | 2021-04-06T05:15:38 | 2021-04-06T05:15:38 | {} | MEMBER |
ClickHouse | ClickHouse | 813,922,236 | 22,654 | azat | >[Stress test (address) Server failed to start](https://clickhouse-test-reports.s3.yandex.net/22654/42d858b7e5843be490c8a6ebe5b17bbd3da975ed/stress_test_(address).html#fail1)
#22592 | 2021-04-06T08:10:30 | 2021-04-06T08:10:30 | {} | MEMBER |
ClickHouse | ClickHouse | 814,063,465 | 19,766 | ghost | If `sequenceNextNode('forward', 'first_match')(timestamp, 1, page, ref_url = 'black_friday')` returns the current value instead of the next value then `sequenceNextNode()` requires another argument indicating the position of return value.
| 2021-04-06T12:00:32 | 2021-04-06T12:00:54 | {} | NONE |
ClickHouse | ClickHouse | 813,064,846 | 22,585 | CurtizJ | Seems that `data_type_default_nullable` doesn't work correctly. | 2021-04-04T16:57:12 | 2021-04-04T16:57:12 | {} | MEMBER |
ClickHouse | ClickHouse | 813,207,295 | 22,647 | qoega | I'll check if it works when we set project name. Or just kill/rm containers and filter by project manually.
Probably it is better to pull all images before all integration tests run.
| 2021-04-05T06:03:57 | 2021-04-05T06:03:57 | {} | MEMBER |
ClickHouse | ClickHouse | 813,426,802 | 22,633 | alexey-milovidov | It is from .sh tests that have subsequent commands (and does not have `-e`). | 2021-04-05T14:36:19 | 2021-04-05T14:36:19 | {} | MEMBER |
ClickHouse | ClickHouse | 813,869,948 | 22,490 | filimonov | > other replicas only fetch the part
Why? Normally all replicas do merges, not only leader. Also since 20.5 there are no leaders. | 2021-04-06T06:47:20 | 2021-04-06T06:47:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,913,717 | 22,689 | alesapin | Report https://clickhouse-test-reports.s3.yandex.net/22669/c830397317fabbea56bc329b3346f3bcdc3f986f/stress_test_(memory).html#fail1 | 2021-04-06T07:58:01 | 2021-04-06T07:58:01 | {} | MEMBER |
ClickHouse | ClickHouse | 813,974,370 | 22,141 | alexey-milovidov | Great!
By tuning the constants of the optimized variants (with subqueries) you can get better response time. | 2021-04-06T09:28:10 | 2021-04-06T09:28:10 | {} | MEMBER |
ClickHouse | ClickHouse | 814,058,738 | 19,766 | ghost | @vdimir
> Actually, maybe we can consider base as part of chain? Could you provide explicit example why we distinct it?
I will explain why `base_cond` is required for our cases by an example.
**Prepare Data**
```SQL
CREATE TABLE bar (timestamp DateTime, id Int32, page String, ref_url String) ENGINE=TinyLog;... | 2021-04-06T11:52:20 | 2021-04-06T11:54:22 | {} | NONE |
ClickHouse | ClickHouse | 813,318,480 | 18,978 | gyuton | Internal documentation ticket: DOCSUP-8233. | 2021-04-05T10:06:11 | 2021-04-05T10:06:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,318,587 | 22,611 | gyuton | Internal documentation ticket: DOCSUP-8233. | 2021-04-05T10:06:26 | 2021-04-05T10:06:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,435,390 | 22,667 | ucasfl | > Ok.
>
> But we don't have this error in our CI.
This happened when I built with `gcc (Gentoo 10.2.0-r5 p6) 10.2.0` | 2021-04-05T14:52:15 | 2021-04-05T14:52:15 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 813,466,623 | 14,298 | mastertheknife | Yeah of course, but at least you tried :-)
The distro packagers aren't perfect either. At work we use MariaDB 10.3 from CentOS8 for our OLTP billing system. RHEL rebased from 10.3.17 to 10.3.27 due to security bugs, but 10.3.27 has two nasty bugs, e.g. mysqld crashes during loading of big writes (e.g. loading a dump... | 2021-04-05T15:51:52 | 2021-04-05T15:51:52 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,657,556 | 21,348 | kitaisreal | Answered in scope of https://github.com/ClickHouse/ClickHouse/issues/22239. | 2021-04-05T21:20:51 | 2021-04-05T21:20:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,913,350 | 22,490 | Sallery-X | > > other replicas only fetch the part
>
> Why? Normally all replicas do merges, not only leader. Also since 20.5 there are no leaders.
I know the feature, If i use the old version, what about the merge task? | 2021-04-06T07:57:22 | 2021-04-06T07:57:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 813,064,800 | 22,585 | CurtizJ | It reproduces with enabled setting `data_type_default_nullable`.
```sql
SET data_type_default_nullable = 1;
CREATE TABLE test ENGINE = MergeTree ORDER BY number AS SELECT number, toString(rand()) AS x FROM numbers(10000);
DETACH TABLE test;
ATTACH TABLE test;
```
```
2021.04.04 19:53:16.482843 [ 4935 ]... | 2021-04-04T16:56:51 | 2021-04-04T17:01:55 | {} | MEMBER |
ClickHouse | ClickHouse | 813,925,690 | 22,671 | tavplubix | What does "without ZooKeeper" mean? | 2021-04-06T08:15:51 | 2021-04-06T08:15:51 | {} | MEMBER |
ClickHouse | ClickHouse | 820,623,476 | 23,110 | filimonov | > And, what is the max_data_part_size_bytes default value?
Zero, which means "no limit". | 2021-04-15T17:59:44 | 2021-04-15T17:59:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,715,242 | 23,000 | alexey-milovidov | Arcadia:
> error: no such file or directory: '$(SOURCE_ROOT)/clickhouse/src/Dictionaries/LibraryDictionarySourceExternal.cpp' | 2021-04-15T20:37:33 | 2021-04-15T20:37:33 | {} | MEMBER |
ClickHouse | ClickHouse | 820,730,689 | 23,096 | alexey-milovidov | Thank you! | 2021-04-15T21:07:04 | 2021-04-15T21:07:04 | {} | MEMBER |
ClickHouse | ClickHouse | 821,238,317 | 22,352 | KochetovNicolai | Functional stateless tests (release, wide parts enabled)
```
2021-04-16 15:33:55 - Condition: and((bitAnd(z, 3) not in [1, 1]), (y in [1, +inf)))
2021-04-16 15:33:55 + Condition: and((bitAnd(z, 3) not in [1, 1]), and((y in [1, +inf)), (bitAnd(z, 3) not in [1, 1])))
```
that's unusual | 2021-04-16T15:00:10 | 2021-04-16T15:00:10 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.