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 | 669,198,467 | 13,386 | nikitamikhaylov | @akuzm It is a part of previous PR to backport it to stable branches. | 2020-08-05T13:37:54 | 2020-08-05T13:37:54 | {} | MEMBER |
ClickHouse | ClickHouse | 669,229,048 | 12,817 | vitlibar | Can't resolve the conflict easily. | 2020-08-05T14:33:14 | 2020-08-05T14:33:14 | {} | MEMBER |
ClickHouse | ClickHouse | 670,050,126 | 8,049 | BooBSD | Привет. Есть ли прогресс в перемножении Decimal? | 2020-08-06T16:54:24 | 2020-08-06T16:54:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 670,143,225 | 7,666 | SaltTan | The grouping issue:
```
SELECT
version() AS h,
count()
FROM remote('localhost,remotehost', system, one)
GROUP BY h
┌─h─────────┬─count()─┐
│ 20.5.4.40 │ 2 │
└───────────┴─────────┘
SELECT
materialize(version()) AS h,
count()
FROM remote('localhost,remotehost', system, one)
GROU... | 2020-08-06T19:15:10 | 2020-08-06T19:15:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,193,800 | 13,285 | alesapin | > Failure in functional and performance tests is unrelated (no changes in code in that PR).
>
> BTW - the number of test differs:
>
> ```
> Integration tests (asan) — fail: 0, passed: 765, error: 0
> Integration tests (release) — fail: 0, passed: 764, error: 0
> ```
>
> Asan run misses the test
> `test_con... | 2020-08-05T13:28:59 | 2020-08-05T13:28:59 | {} | MEMBER |
ClickHouse | ClickHouse | 669,215,331 | 13,147 | akuzm | Still slower even w/o `insert` for different types: https://clickhouse-test-reports.s3.yandex.net/13147/0d947b64bdccf69854cbfd7b939a89ae65ec0792/performance_comparison/report.html#fail1 | 2020-08-05T14:08:57 | 2020-08-05T14:08:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,222,640 | 12,255 | vitlibar | Backported by hand. | 2020-08-05T14:22:06 | 2020-08-05T14:22:06 | {} | MEMBER |
ClickHouse | ClickHouse | 669,516,074 | 13,327 | den-crane | OK. And how to reproduce it?
@templarzq
```
create table T1 Engine=MergeTree order by tuple() as select intDiv(number,1000) tid from numbers(1000000);
create table T2 Engine=Join(any, left,tid) as select number%1000+5 id, intDiv(number,1000)+100 tid from numbers(1000000);
select joinGet('db.T2','id',tid) as... | 2020-08-05T21:18:21 | 2020-08-05T21:18:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,602,615 | 13,147 | akuzm | Just with static asserts it's OK: https://clickhouse-test-reports.s3.yandex.net/13147/ead62bc9d7c0572054b20f54f3a3a47ac0399f6b/performance_comparison/report.html#fail1
Maybe the problem was in ALWAYS_INLINE for resize.
Should we commit this one as a cleanup PR? The assertion is good to have. | 2020-08-05T23:49:45 | 2020-08-05T23:49:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,668,613 | 13,327 | templarzq | create table xxx on cluster bench_cluster(
xxx
)
ENGINE =Join(ANY, LEFT, Id) | 2020-08-06T03:56:48 | 2020-08-06T03:56:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 668,561,289 | 9,624 | nikitamikhaylov | Human readable stacktrace from stateful tests : https://gist.github.com/nikitamikhaylov/13797e63d09737c215e04b548396d084
Also stateful tests are Ok. locally.
UPD: Fixed. | 2020-08-04T12:17:44 | 2020-08-06T13:54:08 | {} | MEMBER |
ClickHouse | ClickHouse | 669,236,273 | 13,387 | keliss | Oh, actually, I can explain why the disk usage looks like this. It's due to data cleanup that we perform for the old data. Closing this issue. | 2020-08-05T14:46:27 | 2020-08-05T14:46:27 | {} | NONE |
ClickHouse | ClickHouse | 669,298,392 | 13,390 | vitlibar | If you want to allow `admin` to change access rights of anyone except `admin` himself, the current implementation of `GRANT/REVOKE` commands doesn't support that. | 2020-08-05T16:33:10 | 2020-08-05T16:33:45 | {} | MEMBER |
ClickHouse | ClickHouse | 669,599,091 | 13,360 | den-crane | check this comment https://github.com/ClickHouse/ClickHouse/issues/3174#issuecomment-423435071
1. How many columns in your tables?
2. How many merges do you see simultaneously usually ? `select count() from system.merges` ?
3. Do you use insert into Distributed table? | 2020-08-05T23:37:14 | 2020-08-05T23:38:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,742,736 | 13,327 | templarzq | bench_cluster have more than 1 node. | 2020-08-06T06:55:30 | 2020-08-06T06:55:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,774,920 | 12,390 | etah000 | @alexey-milovidov I added two setting variables in "Settings.h", but I cannot find a suitable way to get their values in the method " ParserAlterQuery::parseImpl" from "ParserAlterQuery.cpp". Passing those two parameters all the way down is a little clumsy and tedious. Do we have a global variable for "context" or "s... | 2020-08-06T07:58:32 | 2020-08-06T07:58:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,046,678 | 13,285 | filimonov | Failure in functional and performance tests is unrelated (no changes in code in that PR).
BTW - the number of test differs:
```
Integration tests (asan) — fail: 0, passed: 765, error: 0
Integration tests (release) — fail: 0, passed: 764, error: 0
```
Release run misses the test
```test_config_corresponding_... | 2020-08-05T08:03:55 | 2020-08-05T08:07:27 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,076,410 | 3,214 | alexey-milovidov | Released in 20.5. | 2020-08-05T09:05:33 | 2020-08-05T09:05:33 | {} | MEMBER |
ClickHouse | ClickHouse | 669,229,479 | 12,218 | vitlibar | Backported by hand. | 2020-08-05T14:34:04 | 2020-08-05T14:34:04 | {} | MEMBER |
ClickHouse | ClickHouse | 669,665,846 | 13,360 | mjodeh | I stopped the insertions for 3 hours, and once I started it again it worked fine for 45 minutes then the exceptions were there again and the exceptions are not for all dbs for some of them, I tried to connect them to the column's count and had no connection, the db which had 837 columns did not throw the exception | 2020-08-06T03:45:26 | 2020-08-06T03:45:26 | {} | NONE |
ClickHouse | ClickHouse | 669,931,266 | 13,248 | akuzm | ```
SELECT
toUInt32(x),
y,
z
FROM pk
WHERE (x >= toDateTime(1023)) AND (x <= toDateTime(7))
```
https://clickhouse-test-reports.s3.yandex.net/13283/45b57f2df588154b6f766efe28bdfd19985eb6f3/fuzzer/fuzzer.log#fail1 | 2020-08-06T13:37:04 | 2020-08-06T13:37:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 668,812,892 | 13,160 | adevyatova | It is not build, errors:
file: \ClickHouse\docs\tools\output.md
```
yaml.scanner.ScannerError: mapping values are not allowed here
in "<unicode string>", line 6, column 62:
... MS, data is stored in this order:
^
```
file: \ClickHouse\docs\ru\index.md
```
yaml.s... | 2020-08-04T20:37:20 | 2020-08-05T06:01:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,187,708 | 13,385 | den-crane |
it seems now it's kafka_poll_max_batch_size
kafka_poll_max_batch_size, 0, "Maximum amount of messages to be polled in a single Kafka poll.
kafka_max_block_size, 0, "Number of row collected by poll(s) for flushing data from Kafka.
https://github.com/ClickHouse/ClickHouse/pull/11388 | 2020-08-05T13:17:24 | 2020-08-05T13:19:06 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,554,920 | 13,198 | CurtizJ | Queries, that lead to assertion:
```
CREATE TABLE pk_func (`d` DateTime, `ui` UInt32 ) ENGINE = MergeTree ORDER BY toDate(d);
INSERT INTO pk_func SELECT '2020-05-05 01:00:00', number FROM numbers(1000000);
SELECT NULL, * FROM pk_func ORDER BY toDate(d) ASC, ui ASC LIMIT 1;
``` | 2020-08-05T22:32:57 | 2020-08-05T22:32:57 | {} | MEMBER |
ClickHouse | ClickHouse | 669,637,800 | 13,388 | alexey-milovidov | 8fd4038234 | 2020-08-06T01:59:38 | 2020-08-06T01:59:38 | {} | MEMBER |
ClickHouse | ClickHouse | 669,832,301 | 13,407 | alesapin | https://gist.github.com/alesapin/f74a02045a240c055c340afd2d1ff7f0 | 2020-08-06T09:53:11 | 2020-08-06T09:53:11 | {} | MEMBER |
ClickHouse | ClickHouse | 669,973,687 | 13,407 | alesapin | Stress tests memory:
```
2020.08.06 17:01:11.563033 [ 308095 ] {} <Information> DatabaseOrdinary (test_01056_dict_data): Starting up tables.
2020.08.06 17:01:11.564429 [ 308095 ] {} <Information> DatabaseOrdinary (test_01083): Total 1 tables and 0 dictionaries.
2020.08.06 17:01:11.602007 [ 308095 ] {} <Error> Appli... | 2020-08-06T14:46:15 | 2020-08-06T14:47:06 | {} | MEMBER |
ClickHouse | ClickHouse | 670,012,122 | 13,097 | 4ertus2 | style_check.out.log
```
/place/sandbox-data/tasks/3/3/741970133/ClickHouse/src/Common/FieldVisitors.h:169: else if constexpr(sizeof(U) == 16)
xargs: xmllint: No such file or directory
``` | 2020-08-06T15:49:04 | 2020-08-06T15:49:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 670,138,147 | 7,666 | SaltTan | I cannot reproduce the sorting issue anymore (`DB::Exception: Block structure mismatch in MergingSorted stream: different values of constants`).
I tried 20.3.11, 20.3.13, 20.4.7, 20.5.2, 20.5.4 | 2020-08-06T19:03:52 | 2020-08-06T19:03:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 670,188,247 | 13,430 | alexey-milovidov | I found that it was picked just as a random color scheme from the internet:
https://www.vscolors.com/theme/dc453211cc656843a1dbe2671b88b3b2
| 2020-08-06T20:57:49 | 2020-08-06T20:57:49 | {} | MEMBER |
ClickHouse | ClickHouse | 668,879,839 | 13,343 | nikitamikhaylov | > @nikitamikhaylov FYI
>
> You'll probably have to change all this when adding parallel formatting, so it's good that we have more test cases now.
Thanks! I will think. | 2020-08-04T23:33:35 | 2020-08-04T23:33:35 | {} | MEMBER |
ClickHouse | ClickHouse | 669,107,973 | 11,524 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/13368 | 2020-08-05T10:12:03 | 2020-08-05T10:12:03 | {} | MEMBER |
ClickHouse | ClickHouse | 669,195,902 | 12,508 | qoega | https://clickhouse-test-reports.s3.yandex.net/0/73c349fd7fcc5bb0c66df2b7a0f0f575fc5a47e2/integration_tests_(asan).html
Timeouts in
test_storage_rabbitmq/test.py::test_rabbitmq_tsv_with_delimiter
test_storage_rabbitmq/test.py::test_rabbitmq_many_materialized_views
| 2020-08-05T13:32:59 | 2020-08-05T13:32:59 | {} | MEMBER |
ClickHouse | ClickHouse | 669,291,120 | 13,390 | vitlibar | To prohibit `admin` from changing access rights of anybody use
```
REVOKE GRANT OPTION FOR ALL ON *.* FROM admin
```
Or just grant access rights to admin without grant option. | 2020-08-05T16:20:08 | 2020-08-05T16:32:08 | {} | MEMBER |
ClickHouse | ClickHouse | 669,040,024 | 13,360 | qoega | Which version do you use? | 2020-08-05T07:49:56 | 2020-08-05T07:49:56 | {} | MEMBER |
ClickHouse | ClickHouse | 669,076,025 | 3,214 | alexey-milovidov | Fixed in #11661. | 2020-08-05T09:04:46 | 2020-08-05T09:04:46 | {} | MEMBER |
ClickHouse | ClickHouse | 669,234,202 | 13,384 | alesapin | fast test ok
| 2020-08-05T14:42:45 | 2020-08-05T14:42:45 | {} | MEMBER |
ClickHouse | ClickHouse | 669,237,169 | 12,901 | vitlibar | Can't resolve the conflict easily. | 2020-08-05T14:48:07 | 2020-08-05T14:48:07 | {} | MEMBER |
ClickHouse | ClickHouse | 669,603,377 | 13,147 | akuzm | Stress test failure is 'server failed to restart after test', but there are no errors in server log and it wasn't even restarted. I think we fixed something like this recently. | 2020-08-05T23:52:36 | 2020-08-05T23:52:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,129,605 | 13,369 | CurtizJ | Duplicate of #12562 | 2020-08-05T11:04:13 | 2020-08-05T11:04:13 | {} | MEMBER |
ClickHouse | ClickHouse | 669,222,537 | 12,250 | vitlibar | Backported by hand. | 2020-08-05T14:21:55 | 2020-08-05T14:21:55 | {} | MEMBER |
ClickHouse | ClickHouse | 669,722,854 | 12,188 | compasses | I thought the mutation operation should only modify the parts constrained by the where condition in mutation SQL. If it not work like that, the mutation performance will be very poor for large tables. Am I right? | 2020-08-06T06:08:49 | 2020-08-06T06:08:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,799,571 | 13,285 | alesapin | Failures are not relevant to changes, let's merge | 2020-08-06T08:48:11 | 2020-08-06T08:48:11 | {} | MEMBER |
ClickHouse | ClickHouse | 669,934,430 | 13,327 | den-crane | It's because T2 (Join) is empty on shards (it's not a replicated engine). And joinGet works on shards against empty table.
You can fill T2 at all nodes with the same data or perform joinGet at the initiator using `from()`.
```SQL
create table T1 on cluster segmented (tid UInt64) Engine=MergeTree order by tuple();
... | 2020-08-06T13:42:50 | 2020-08-06T13:47:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,973,937 | 13,407 | alesapin | Functional tests ok. | 2020-08-06T14:46:40 | 2020-08-06T14:46:40 | {} | MEMBER |
ClickHouse | ClickHouse | 668,957,504 | 6,063 | zhang2014 | maybe we can close this issue ? | 2020-08-05T03:23:31 | 2020-08-05T03:23:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,236,711 | 13,382 | alesapin | Fast test ok. | 2020-08-05T14:47:15 | 2020-08-05T14:47:15 | {} | MEMBER |
ClickHouse | ClickHouse | 669,515,460 | 13,333 | alesapin | Actually our code works badly with impossible KeyConditions. In this case, we have condition:
```
(column 0 in [4294967280, +inf)), (column 0 in (-inf, 0]), and
>>>>>first<<<<< >>>>>second<<<<<
```
Which is obviously always false for any interval. But for example, for interval `[0, +in... | 2020-08-05T21:16:58 | 2020-08-05T21:16:58 | {} | MEMBER |
ClickHouse | ClickHouse | 669,668,754 | 13,327 | templarzq | maybe the option "on cluster xxx" lead to this result? | 2020-08-06T03:57:31 | 2020-08-06T03:57:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,967,820 | 13,418 | den-crane | You can create own macros in any combination and use them as you want. layer / shard / replica -- just a recommendation.
You don't need layer because it's for huge clusters like 100+ nodes.
so for example create two shard macros **shard_double** and **shard_triple**
```
s1 shard_double=1 shard_triple=1 repli... | 2020-08-06T14:36:42 | 2020-08-06T14:48:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,136,750 | 12,562 | alexey-milovidov | And it's all about formatting of progress messages. | 2020-08-05T11:20:56 | 2020-08-05T11:20:56 | {} | MEMBER |
ClickHouse | ClickHouse | 669,173,316 | 13,383 | alexey-milovidov | CC @Enmk | 2020-08-05T12:48:20 | 2020-08-05T12:48:20 | {} | MEMBER |
ClickHouse | ClickHouse | 669,180,733 | 13,384 | akuzm | Maybe just `ln -s /var/log/clickhouse-server /test_output`? This way you'll always have the log, even if the script fails in some unpredictable way. | 2020-08-05T13:04:00 | 2020-08-05T13:04:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,181,610 | 13,384 | akuzm | > Maybe just `ln -s /var/log/clickhouse-server /test_output`? This way you'll always have the log, even if the script fails in some unpredictable way.
I confused the argument order again... There is a good mnemonic -- this command is similar `cp`, so the arguments are in the same order, e.g. `cp <src> <dest>`, but i... | 2020-08-05T13:05:34 | 2020-08-05T13:05:34 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,233,275 | 13,018 | vitlibar | Can't resolve the conflict easily. | 2020-08-05T14:41:02 | 2020-08-05T14:41:02 | {} | MEMBER |
ClickHouse | ClickHouse | 669,235,617 | 12,816 | vitlibar | Can't resolve the conflict easily. | 2020-08-05T14:45:17 | 2020-08-05T14:45:17 | {} | MEMBER |
ClickHouse | ClickHouse | 669,755,223 | 12,188 | compasses | OK thanks | 2020-08-06T07:21:12 | 2020-08-06T07:21:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,800,681 | 13,379 | alesapin | Failures are not relevant to changes. | 2020-08-06T08:50:35 | 2020-08-06T08:50:35 | {} | MEMBER |
ClickHouse | ClickHouse | 669,216,336 | 12,361 | vitlibar | Already in `20.5`. | 2020-08-05T14:10:45 | 2020-08-05T14:10:45 | {} | MEMBER |
ClickHouse | ClickHouse | 669,232,824 | 13,019 | vitlibar | Can't resolve the conflict easily. | 2020-08-05T14:40:12 | 2020-08-05T14:40:12 | {} | MEMBER |
ClickHouse | ClickHouse | 668,961,759 | 13,327 | templarzq | it works before version 20.5 (include 20.5.1.1) | 2020-08-05T03:41:05 | 2020-08-05T03:41:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,092,151 | 12,508 | qoega | ASAN heap-use-after-free
https://clickhouse-test-reports.s3.yandex.net/0/73c349fd7fcc5bb0c66df2b7a0f0f575fc5a47e2/integration_tests_(asan).html
```
2020.08.01 05:22:49.730743 [ 109 ] {} <Fatal> BaseDaemon: ########################################
2020.08.01 05:22:49.730837 [ 109 ] {} <Fatal> BaseDaemon: (version 20... | 2020-08-05T09:37:58 | 2020-08-05T09:37:58 | {} | MEMBER |
ClickHouse | ClickHouse | 669,136,566 | 12,562 | alexey-milovidov | It's an issue with formatting in `clickhouse-client` in interactive mode.
This command is producing correct output:
`clickhouse-client --query "select 1 format JSON" | less` | 2020-08-05T11:20:25 | 2020-08-05T11:20:25 | {} | MEMBER |
ClickHouse | ClickHouse | 669,580,594 | 13,394 | azat | Although not sure about bug fix label, but too late anyway | 2020-08-05T22:49:41 | 2020-08-05T22:49:41 | {} | MEMBER |
ClickHouse | ClickHouse | 669,754,062 | 12,188 | filimonov | > If it not work like that, the mutation performance will be very poor for large tables.0 Am I right?
Non-affected parts are not processed just renamed (renames are cheap) . That FR is about avoiding renames when it is not neccessary.
Anyway ALTER rename/delete is quite heavy operation (that's why it called ALTER... | 2020-08-06T07:18:15 | 2020-08-06T07:18:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,877,836 | 13,391 | BayoNet | 1. Add a link from the setting description to the JSON format description.
2. Add a link from the JSON format description to the setting description.
Add examples of output with the setting turned off and on. | 2020-08-06T11:43:25 | 2020-08-06T11:43:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,927,966 | 12,550 | Akazz | JFYI: `test_storage_rabbitmq/test.py::test_rabbitmq_many_materialized_views` from integration tests "flaps" | 2020-08-06T13:30:41 | 2020-08-06T13:30:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,932,711 | 13,283 | akuzm | * ast fuzzer failure is https://github.com/ClickHouse/ClickHouse/issues/13248
* perf test is broken in master
* 00933_test_fix_extra_seek_on_compressed_cache is in CI Fest #7 https://github.com/ClickHouse/ClickHouse/issues/12508 | 2020-08-06T13:39:45 | 2020-08-06T13:39:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 668,884,503 | 13,203 | akuzm | Functional stateless test -- 00834_kill_mutation, known problem
Performance test -- long synthetic_hardware_benchmark test, also known problem | 2020-08-04T23:47:28 | 2020-08-04T23:47:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,106,465 | 12,508 | azat | >ASAN heap-use-after-free
@qoega 73c349fd7f has been reverted in 4fa0b16eaa2073374050fc823f697223f9816639 | 2020-08-05T10:08:39 | 2020-08-05T10:08:39 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 669,107,440 | 11,524 | alexey-milovidov | `arrayIntersect` has been speed up due to lower initial hash table size. | 2020-08-05T10:10:51 | 2020-08-05T10:10:51 | {} | MEMBER |
ClickHouse | ClickHouse | 669,234,815 | 12,843 | vitlibar | Can't resolve the conflict easily. | 2020-08-05T14:43:51 | 2020-08-05T14:43:51 | {} | MEMBER |
ClickHouse | ClickHouse | 669,282,839 | 13,237 | vdimir | I see this error, I'll fix it
```
2020-08-05 15:30:37 In file included from ../src/Functions/multiFuzzyMatchAnyIndex.cpp:3:
2020-08-05 15:30:37 /ClickHouse/src/Functions/MultiMatchAnyImpl.h:123:114: error: too few arguments to function call, expected 5, have 4
2020-08-05 15:30:37 MatchImpl<false, false>... | 2020-08-05T16:05:55 | 2020-08-05T16:05:55 | {} | MEMBER |
ClickHouse | ClickHouse | 669,324,802 | 10,851 | tavplubix | Some queries (such as `SHOW DATABASES`, `SELECT ... FROM system.tables`, `SYSTEM RESTART REPLICAS`, `SYSTEM STOP/START MERGES`, etc) fail when MySQL master is not available:
```
:) show databases
Received exception from server (version 20.7.1):
Code: 1002. DB::Exception: Received from localhost:9000. DB::Exception:... | 2020-08-05T17:26:36 | 2020-08-05T17:26:36 | {} | MEMBER |
ClickHouse | ClickHouse | 669,670,414 | 13,327 | den-crane | > maybe the option "on cluster xxx" lead to this result?
I don't see how.
Please provide reproducible example. | 2020-08-06T04:04:22 | 2020-08-06T04:04:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,742,502 | 13,327 | templarzq | create table T1 on cluster bench_cluster Engine=MergeTree order by tuple() as select intDiv(number,1000) tid from numbers(1000000);
create table T2 on cluster bench_cluster(
id UInt32,
tid UInt64
) Engine=Join(any, left,tid);
insert into T2 select number%1000+5 id, intDiv(number,1000)+100 tid from numbers(1000... | 2020-08-06T06:54:54 | 2020-08-06T06:54:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,920,518 | 13,348 | akuzm | Ran the perf test under address sanitizer once, no errors. It was very slow, some queries take 10 or 20 minutes to complete.
`ch/docker/test/performance-comparison/perf.py --runs 1 ch/tests/performance/linear_regression.xml` | 2020-08-06T13:16:46 | 2020-08-06T13:16:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,937,840 | 13,343 | akuzm | I'll take a look at the perf test -- probably it depends on the old broken behavior and doesn't expect the header somewhere. | 2020-08-06T13:48:50 | 2020-08-06T13:48:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 668,959,904 | 10,851 | zhang2014 | `Performance — 1 errors, 1 too long, 10 slower, 250 unstable`
is not related to this patch
@tavplubix Could you tell me the `Yandex synchronization check (only for Yandex employees)` and `Yandex third-party checks (only for Yandex employees)` failure reason? this will help me to fix it : )
| 2020-08-05T03:33:29 | 2020-08-05T03:34:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,046,916 | 13,360 | mjodeh | 20.3.8.53 | 2020-08-05T08:04:22 | 2020-08-05T08:04:22 | {} | NONE |
ClickHouse | ClickHouse | 669,127,756 | 10,851 | tavplubix | ```
src/Databases/MySQL/MaterializeMySQLSyncThread.h:3:10: fatal error: 'config_core.h' file not found
```
Can be fixed with
```
#if !defined(ARCADIA_BUILD)
# include "config_core.h"
#endif
```
And
```
no such file or directory: '$(SOURCE_ROOT)/clickhouse/src/Databases/DatabaseMySQL.cpp'
```
Can be fix... | 2020-08-05T10:59:23 | 2020-08-05T10:59:23 | {} | MEMBER |
ClickHouse | ClickHouse | 669,236,783 | 12,922 | vitlibar | Can't resolve the conflict easily. | 2020-08-05T14:47:23 | 2020-08-05T14:47:23 | {} | MEMBER |
ClickHouse | ClickHouse | 669,238,710 | 12,938 | vitlibar | Can't resolve the conflict easily. | 2020-08-05T14:50:54 | 2020-08-05T14:50:54 | {} | MEMBER |
ClickHouse | ClickHouse | 669,641,712 | 7,690 | keyboardblue | Still reproduce on 20.5.3.27, any updates there? | 2020-08-06T02:13:58 | 2020-08-06T02:13:58 | {} | NONE |
ClickHouse | ClickHouse | 669,663,342 | 13,360 | mjodeh | thanks @den-crane
the columns number varies from 20 column to 837 columns and only 3 columns in every table are not nullables the rest are nullables.
`select count() from system.merges` is always `16` **this count does not change after executing `optimize table` with or without `final` keyword also the progress do... | 2020-08-06T03:35:07 | 2020-08-06T03:35:07 | {} | NONE |
ClickHouse | ClickHouse | 669,796,277 | 13,190 | Slach | @blinkov could you look this PR again? | 2020-08-06T08:41:24 | 2020-08-06T08:41:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,907,673 | 13,409 | akuzm | The server failed to start and there is no log, maybe the yesterday's fix will help. The script still tried to run tests even when the server is down, looks like a bug. | 2020-08-06T12:51:54 | 2020-08-06T12:51:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,033,751 | 6,556 | zhegao9 | > Workaround: create table with Engine=MySQL where that field will be listed as String instead of FixedString.
This works for me.
MySQL:
```sql
create table t (
`uuid` binary(16),
...
)
```
Clickhouse:
```sql
create table t (
`uuid` String,
...
) ENGINE=MYSQL(...);
```
| 2020-08-05T07:35:43 | 2020-08-05T07:36:32 | {} | NONE |
ClickHouse | ClickHouse | 669,148,668 | 1,661 | compasses | @MeteHanC Hi, I just wonder your test case run against on no replica merge tree? do you have checked the ReplicatedMergeTree, and QPS benchmark? Maybe no problem for one time update, but how about many updates in a short time, hybrid with many selects.
Just want to know the realtime performance of CH, @blinkov do you ... | 2020-08-05T11:51:22 | 2020-08-05T11:51:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,934,188 | 13,283 | akuzm | So what I did eventually is:
* Removed a potentially unsafe transformation from query fuzzer (replacing random child ASTLiteral with another AST node). See the comment on why it is not safe.
* Fixed some problems in dictionary ASTs including broken clone() methods.
I'll merge this now. | 2020-08-06T13:42:26 | 2020-08-06T13:42:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 670,145,133 | 7,666 | KochetovNicolai | @alexey-milovidov the problem is the same as I showed for `randConstant` and `now64`.
Maybe just remove `isDeterministicInScopeOfQuery` for it? | 2020-08-06T19:19:21 | 2020-08-06T19:19:21 | {} | MEMBER |
ClickHouse | ClickHouse | 670,186,056 | 13,430 | alexey-milovidov | The documentation was using questionable choice of color scheme. See the screenshot:
Green on gray and orange on gray does not have enough contrast. | 2020-08-06T20:52:18 | 2020-08-06T20:52:30 | {} | MEMBER |
ClickHouse | ClickHouse | 669,165,327 | 13,379 | alexey-milovidov | The notorious usability of git: https://gist.github.com/myusuf3/7f645819ded92bda6677 | 2020-08-05T12:30:58 | 2020-08-05T12:30:58 | {} | MEMBER |
ClickHouse | ClickHouse | 669,255,853 | 12,508 | qoega | 00620_optimize_on_nonleader_replica_zookeeper
https://clickhouse-test-reports.s3.yandex.net/0/73c349fd7fcc5bb0c66df2b7a0f0f575fc5a47e2/functional_stateless_tests_(unbundled).html | 2020-08-05T15:20:14 | 2020-08-05T15:20:14 | {} | MEMBER |
ClickHouse | ClickHouse | 669,314,713 | 13,340 | MyroTk | Testflows check is green and there's no conflicts, should be ready for merge. | 2020-08-05T17:05:49 | 2020-08-05T17:05:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,619,960 | 13,147 | alexey-milovidov | > Maybe the problem was in ALWAYS_INLINE for resize.
It has crucial effect for performance of clang build (that we does not test for performance).
Maybe it will allow us to replace gcc with clang in near future. | 2020-08-06T00:53:15 | 2020-08-06T00:53:27 | {} | MEMBER |
ClickHouse | ClickHouse | 669,649,364 | 7,690 | filimonov | See https://github.com/ClickHouse/ClickHouse/pull/11512 | 2020-08-06T02:41:14 | 2020-08-06T02:41:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 669,847,112 | 13,385 | andrei-karpliuk | > it seems now it's kafka_poll_max_batch_size
>
> kafka_poll_max_batch_size, 0, "Maximum amount of messages to be polled in a single Kafka poll.
> kafka_max_block_size, 0, "Number of row collected by poll(s) for flushing data from Kafka.
>
> #11388
thanks, it works | 2020-08-06T10:27:07 | 2020-08-06T10:27:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 658,847,027 | 11,576 | jondeandres | Any update on this? | 2020-07-15T15:51:48 | 2020-07-15T15:51:48 | {} | NONE |
ClickHouse | ClickHouse | 658,848,852 | 12,525 | gane5hvarma | @den-crane got it, but querying by event_received_at will decrease the query performance right if its not a primary key. | 2020-07-15T15:55:46 | 2020-07-15T15:55:46 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.