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 | 675,049,416 | 13,822 | tavplubix | For unknown reason it fails only in CI environment, I cannot reproduce the failure on my machine too. I've disabled the test in master to suppress annoying failures in other PRs. | 2020-08-17T18:46:18 | 2020-08-17T18:46:18 | {} | MEMBER |
ClickHouse | ClickHouse | 675,156,160 | 13,816 | filimonov | If it's not too costy - may be we can calc query_hash in advance and put it into query_log? | 2020-08-17T23:00:04 | 2020-08-17T23:00:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,443,069 | 13,405 | akuzm | So how exactly does this PR achieve a speedup? I've looked at the code but didn't get an idea what it does. | 2020-08-18T12:16:46 | 2020-08-18T12:16:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,450,340 | 13,855 | nikitamikhaylov | The same as #13839 . But the latter was created from web-interface and I was confused to merge master into that branch, so I created new PR from my fork. | 2020-08-18T12:32:37 | 2020-08-18T12:32:37 | {} | MEMBER |
ClickHouse | ClickHouse | 675,633,726 | 10,373 | azat | @4ertus2 this is ready, can you take a look? | 2020-08-18T18:11:50 | 2020-08-18T18:11:50 | {} | MEMBER |
ClickHouse | ClickHouse | 676,081,680 | 13,878 | filimonov | Import them as JSONAsString, parse them with JSON functions (can be in MV) | 2020-08-19T10:11:13 | 2020-08-19T10:11:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,053,528 | 9,624 | KochetovNicolai | Skip `01428_hash_set_nan_key` because it uses `sumDistinct` function which is not implemented.
Skip `01414_push_predicate_when_contains_with_clause` because `EXPLAIN SYNTAX` is not implemented.
| 2020-08-17T18:55:10 | 2020-08-17T18:55:10 | {} | MEMBER |
ClickHouse | ClickHouse | 675,728,373 | 13,156 | akuzm | > >
>
> @akuzm this PR changes the protocol but does not changes the [revision](https://github.com/ClickHouse/ClickHouse/blob/fc7fb3b10a4a6ae83c30b896f63ba322d233c6c7/cmake/autogenerated_versions.txt#L2) (since it should auto generated as stated by the file), but AFAICS [performance tests uses clickhouse-client fro... | 2020-08-18T21:25:13 | 2020-08-18T21:25:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,991,379 | 11,257 | Alex-Burmak | Implemented in https://github.com/ClickHouse/ClickHouse/pull/12638
| 2020-08-19T09:04:48 | 2020-08-19T09:04:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,834,467 | 12,508 | bohutang | ClickHouse Stress Test (undefined) Server failed to start
https://clickhouse-test-reports.s3.yandex.net/13820/26525a5eb1df2bb7cbcb97ddfd2e674feb801406/stress_test_(undefined).html#fail1 | 2020-08-19T03:45:04 | 2020-08-19T09:52:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 674,903,930 | 13,835 | fastio | @den-crane Thank you for reply. I got it. If the ORDER BY expression modified with existing column, the existing data of table should be re-sorted which will pay a huge cost. | 2020-08-17T14:08:45 | 2020-08-17T14:08:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 674,960,412 | 11,870 | KochetovNicolai | Skip `01428_hash_set_nan_key` because it uses `sumDistinct` function which is not implemented. | 2020-08-17T15:49:45 | 2020-08-17T15:49:45 | {} | MEMBER |
ClickHouse | ClickHouse | 675,494,169 | 13,837 | qoega | Actually there are lots of tests that don't use default config at all. You can wait for my fixes - I'll try to fix what have left ASAP. | 2020-08-18T13:57:11 | 2020-08-18T13:57:11 | {} | MEMBER |
ClickHouse | ClickHouse | 675,663,294 | 13,857 | alexey-milovidov | I can stratify it in the way that nans will be compared for bit equality. | 2020-08-18T19:14:27 | 2020-08-18T19:14:27 | {} | MEMBER |
ClickHouse | ClickHouse | 675,915,375 | 13,665 | hczhcz | The cause and the solution are not certain for now. | 2020-08-19T07:55:25 | 2020-08-19T07:55:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,970,336 | 13,878 | hczhcz | For complex JSON structures, it is good to import them as strings and use JSON functions (JSONExtract-). | 2020-08-19T08:50:23 | 2020-08-19T08:50:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,333,847 | 13,848 | Fanduzi | > After testing, it is found that if the newly added column is of string type, this problem will not occur. If it is of int type, it will appear, with 100% recurrence
hi.
alter table inno_topping_user.active_duration_local on cluster ck_cluster add column V_IP String DEFAULT 1 after V_VERSION;
maybe string defa... | 2020-08-18T08:19:05 | 2020-08-18T08:19:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,373,862 | 13,840 | alexey-milovidov | > AST Fuzzer
Most likely fixed in https://github.com/ClickHouse/ClickHouse/pull/12550 | 2020-08-18T09:35:48 | 2020-08-18T09:35:48 | {} | MEMBER |
ClickHouse | ClickHouse | 675,731,765 | 13,156 | azat | >"right" is the new one, I really should have named them "old" and "new" :)
+1
> Does it mean that the new client doesn't work with the old server, and the test times out?
Yes (although I'm not sure about timeout)
> Looks like a bug, the client from master must work normally with older versions of server. T... | 2020-08-18T21:33:53 | 2020-08-18T21:40:29 | {} | MEMBER |
ClickHouse | ClickHouse | 676,015,866 | 12,444 | bharatnc | @alexey-milovidov, so I took this one and I have #13880 open for this:
```bash
:) select * from system.time_zones limit 10;
SELECT *
FROM system.time_zones
LIMIT 10
┌─time_zone───────────┐
│ America/Thunder_Bay │
│ Africa/Blantyre │
│ Africa/Lusaka │
│ America/Managua │
│ Europe/Chisinau ... | 2020-08-19T09:21:53 | 2020-08-19T09:22:13 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 674,834,679 | 13,826 | filimonov | Also it would be nice to support attaching 'offline' parts / partitions from detached folder the same way.
For REPLACE PARTITION it should not be supported (we can't replace month part with day of data, only add days one by one) | 2020-08-17T11:49:41 | 2020-08-17T11:49:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,320,410 | 13,848 | filimonov | Similar: https://github.com/ClickHouse/ClickHouse/issues/13673 | 2020-08-18T07:53:22 | 2020-08-18T07:53:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,439,377 | 7,782 | abyss7 | Do you have any news? | 2020-08-18T12:08:50 | 2020-08-18T12:08:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,550,194 | 12,031 | vitlibar | > does this PR fix this kind of errors?
>
> > GRANT ALL ON `xxx`.* TO `yyyy`
> > Clickhouse error when executing command: Code: 446, Message: default: Introspection functions are disabled
>
> (with Clickhouse 20.6.3.28)
No, it hasn't fixed yet.
To use the `ALL` keyword while you're granting privileges you sh... | 2020-08-18T15:31:19 | 2020-08-18T15:31:43 | {} | MEMBER |
ClickHouse | ClickHouse | 675,874,099 | 13,646 | filimonov | @bharatnc please check follow up PR https://github.com/ClickHouse/ClickHouse/pull/13648
Timezones are taken from tzdb (see contrib/cct/testdata) and added to clickhouse binary using objcopy | 2020-08-19T06:13:16 | 2020-08-19T06:13:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,998,997 | 13,634 | millin | Same issue in 20.5.4.40 | 2020-08-19T09:10:08 | 2020-08-19T09:10:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 674,895,525 | 13,836 | den-crane | you can use (without stopping ClickHouse)
sql: "detach table hits_replica"
sh: rm -rf /var/lib/clickhouse/data/tutorial/hits_replica ; **rm -rf /var/lib/clickhouse/metadata/tutorial/hits_replica.sql** | 2020-08-17T13:53:15 | 2020-08-17T13:53:54 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,331,657 | 13,673 | filimonov | let's continue in #13848 | 2020-08-18T08:14:50 | 2020-08-18T08:14:50 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,484,235 | 13,402 | Akazz | The failing test `1193_metadata_loading` in "Functional stateless tests (release, DatabaseAtomic)" looks irrelevant to the changes in this PR | 2020-08-18T13:38:47 | 2020-08-18T13:38:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,644,032 | 12,508 | azat | `00956_sensitive_data_masking` is still flacky https://clickhouse-test-reports.s3.yandex.net/10373/348ef1256ea8fb8f61109c33bbdd28daf46bdc8e/functional_stateless_tests_(debug).html#fail1
The problem is that it uses the following pattern:
```sh
clickhouse-client -q ... & # run some query in background
clickhouse-cl... | 2020-08-18T18:32:54 | 2020-08-18T19:05:25 | {} | MEMBER |
ClickHouse | ClickHouse | 675,679,591 | 13,402 | excitoon | Progress on #10650 | 2020-08-18T19:51:16 | 2020-08-18T19:51:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,074,194 | 13,876 | filimonov | It's too avoid "too many parts" errors.
See the setting:
```
SELECT *
FROM system.merge_tree_settings
WHERE name = 'number_of_free_entries_in_pool_to_execute_mutation'
┌─name───────────────────────────────────────────────┬─value─┬─changed─┬─description────────────────────────────────────────────────────────... | 2020-08-19T10:05:46 | 2020-08-19T10:07:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,107,822 | 13,822 | zhang2014 | Ok, I will keep observing after merged(I still don't understand why test failed). | 2020-08-19T10:30:15 | 2020-08-19T10:31:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 674,898,479 | 13,798 | den-crane | >Does the mechanical hard drive also have a big impact? raid 0
mechanical hard drive does not affect on memory usage and the error "Allocator: Cannot malloc 4.00"
mechanical hard drive is OK for ClickHouse from performance perspective.
RAID of 12 HDD drives able to read 1GB/s -- it's close to a single SSD. | 2020-08-17T13:58:35 | 2020-08-17T13:58:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,645,539 | 13,798 | den-crane | I don't see how SSD/HDD can be related to the error "Cannot allocate memory".
`overcommit_memory = 0` - did it help? | 2020-08-18T18:35:59 | 2020-08-18T18:35:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,728,994 | 13,156 | akuzm | ```
2020-08-19 00:02:42 + clickhouse-client --port 9001 --query 'select 1'
2020-08-19 00:02:42 Received exception from server (version 20.8.1):
2020-08-19 00:02:42 Code: 62. DB::Exception: Received from localhost:9001. DB::Exception: Syntax error: failed at position 1 (''): . Unrecognized token.
2020-08-19 00... | 2020-08-18T21:26:39 | 2020-08-18T21:26:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,737,185 | 13,156 | akuzm | > The problem is that client and server from the left and right has the same revision, so there is no way to make the compatible, client with previous revision can communicate with the new server, so as new client can communicate with and old server.
> That said that no stable release will be affected of course (but I... | 2020-08-18T21:48:21 | 2020-08-18T21:48:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,876,387 | 13,646 | bharatnc | > @bharatnc please check follow up PR #13648
>
> Timezones are taken from tzdb (see contrib/cct/testdata) and added to clickhouse binary using objcopy
:+1: thanks, clearer to me now! | 2020-08-19T06:19:56 | 2020-08-19T06:19:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 674,863,079 | 12,736 | vitlibar | I've implemented <user_directories> in the main config (https://github.com/ClickHouse/ClickHouse/pull/13425), it should be useful for your PR. | 2020-08-17T12:52:50 | 2020-08-17T12:52:50 | {} | MEMBER |
ClickHouse | ClickHouse | 674,911,110 | 13,677 | hagen1778 | > May be optimization is fired before replacement of "real" table to an "inserted blocks only" happens (similar to #10603 ) ?
Could be, I'm not sure. Probably statement in `LEFT JOIN` scans "real" data.
```
...
LEFT JOIN (
SELECT
id,
maxMerge(latest_checkout_time) as current_latest_check... | 2020-08-17T14:21:28 | 2020-08-17T14:24:18 | {} | NONE |
ClickHouse | ClickHouse | 675,214,480 | 13,845 | den-crane | 1. How did you get data from Oracle?
2. What query did you use to insert data to CH?
3. What is the table DDL?
4. Did you try to use MergeTree instead of ReplacingMergeTree?
5. Can you provide data sample so I can reproduce your issue?
6. What CH version do you use? | 2020-08-18T02:36:22 | 2020-08-18T02:36:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,232,725 | 13,845 | mbtolou | answer :
1.By C# driver
2.bulk insert
3.
```
CREATE TABLE MDM.BIZ_PUB_DATA_USE_T_S
(
`SERVER_ID` UInt8,
`DATA_ID` String,
`TV` DateTime('Asia/Tehran'),
`DELTA_A` Nullable(Decimal(12, 5)),
`DELTA_R` Nullable(Decimal(12, 5)),
`CT` Nullable(Decimal(12, 5)),
`PT` Nullable(Decimal(... | 2020-08-18T03:48:54 | 2020-08-18T03:48:54 | {} | NONE |
ClickHouse | ClickHouse | 675,233,391 | 13,845 | mbtolou | i found another problem :
this query :
```
SELECT addHours(toDate('2020-03-21'),0) H1,addHours(toDate('2020-03-21'),1) H2
```
return same result
| 2020-08-18T03:51:46 | 2020-08-18T03:52:09 | {} | NONE |
ClickHouse | ClickHouse | 675,395,997 | 13,677 | KochetovNicolai | I think that:
1. We don't need a separates setting for `select` like `max_rows_to_read_estimated` or `estimated_max_rows_to_read` because it would be extremely difficult to explain what it does and why it is needed. `max_rows_to_read` fits perfect here.
2. ~~We shouldn't fail because of `max_rows_to_read` while inse... | 2020-08-18T10:23:18 | 2020-08-18T10:58:09 | {} | MEMBER |
ClickHouse | ClickHouse | 675,448,809 | 13,839 | nikitamikhaylov | I'm confused. | 2020-08-18T12:29:18 | 2020-08-18T12:29:18 | {} | MEMBER |
ClickHouse | ClickHouse | 675,986,413 | 13,874 | alexey-milovidov | It's not a bug.
When index is out of bound in `arr[idx]`, we return default value of type of array elements. If array elements are numbers, we return 0, if strings, we return empty string. If array elements are Nullable of something, we return NULL.
This is implemented that way to avoid producing Nullable type wh... | 2020-08-19T09:01:18 | 2020-08-19T09:01:35 | {} | MEMBER |
ClickHouse | ClickHouse | 675,000,015 | 13,348 | KochetovNicolai | msan and tsan also do not reproduce anything. | 2020-08-17T17:06:41 | 2020-08-17T17:06:41 | {} | MEMBER |
ClickHouse | ClickHouse | 675,206,476 | 13,845 | mbtolou | Data of Oracle
Data Imported in clickhouse
my table engine is **ReplacingMergeTree**
... | 2020-08-18T02:04:45 | 2020-08-18T02:04:45 | {} | NONE |
ClickHouse | ClickHouse | 675,260,589 | 13,845 | mbtolou |
`
SELECT toDateTime('2020-03-21 00:00:00'),toDateTime('2020-03-21 01:00:00'),toDateTime('2020-03-21 23:00:00')
`
| 2020-08-18T05:32:21 | 2020-08-18T05:34:16 | {} | NONE |
ClickHouse | ClickHouse | 675,377,422 | 13,779 | alexey-milovidov | Squashing is Ok. | 2020-08-18T09:42:58 | 2020-08-18T09:42:58 | {} | MEMBER |
ClickHouse | ClickHouse | 675,594,837 | 13,864 | filimonov | 
| 2020-08-18T16:53:47 | 2020-08-18T16:53:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,667,649 | 13,856 | alexey-milovidov | Thank you!
I wonder how it is possible that documentation was so wrong! | 2020-08-18T19:23:58 | 2020-08-18T19:23:58 | {} | MEMBER |
ClickHouse | ClickHouse | 674,887,613 | 13,822 | zhang2014 | @tavplubix
I'm not sure what the reason for this failure. I've run 100 times this test in my env(so far I've run 30 times and not found failure.) I will wait run test result. maybe we could comment the relevant SELECT first? | 2020-08-17T13:39:22 | 2020-08-17T13:39:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,226,351 | 13,630 | hexiaoting | Add new feature: Format LineAsString #13846 | 2020-08-18T03:22:16 | 2020-08-18T03:22:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,227,885 | 13,629 | hexiaoting | Just remove all the codes related to ODBCDriver format? | 2020-08-18T03:28:23 | 2020-08-18T03:28:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,249,144 | 13,845 | mbtolou | I am using this driver :
this lines fetch data from IDataReader and insert into ClickHouse in raw binary format .
https://github.com/DarkWanderer/ClickHouse.Client/blob/c17fb5ca42bcb9d38e270d238fae50524dc1b6d4/ClickHouse.Client/Copy/ClickHouseBulkCopy.cs#L134
this driver work correctly on other days | 2020-08-18T04:57:40 | 2020-08-18T05:45:13 | {} | NONE |
ClickHouse | ClickHouse | 675,337,440 | 13,848 | IVitamin-C | > > After testing, it is found that if the newly added column is of string type, this problem will not occur. If it is of int type, it will appear, with 100% recurrence
>
> hi.
> alter table inno_topping_user.active_duration_local on cluster ck_cluster add column V_IP String DEFAULT 1 after V_VERSION;
>
> maybe ... | 2020-08-18T08:25:42 | 2020-08-18T08:25:42 | {} | NONE |
ClickHouse | ClickHouse | 675,377,544 | 13,779 | nikitamikhaylov | > @nikitamikhaylov TSan report in cache dictionary.
Already fixed in my PR. | 2020-08-18T09:43:16 | 2020-08-18T09:43:16 | {} | MEMBER |
ClickHouse | ClickHouse | 675,854,917 | 13,646 | bharatnc | > BTW list of timezone packaged is not full...
@filimonov I'm curious about:
1. the whereabouts of the list of timezones and where it's located.
2. and how the list is being packaged into the Clickhouse binary.
Would you be able to point me towards 1 & 2? | 2020-08-19T05:10:56 | 2020-08-19T05:10:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,975,561 | 13,868 | alexey-milovidov | BTW, everything related to NaNs falls into "implementation specific behaviour". | 2020-08-19T08:53:54 | 2020-08-19T08:53:54 | {} | MEMBER |
ClickHouse | ClickHouse | 675,974,549 | 13,878 | hexiaoting | @hczhcz
> For complex JSON structures, it is good to import them as strings and use JSON functions (JSONExtract-).
You mean that take the entire json object as a string?
But in the way, it is stored as a single column and we cannot take advantage of column storage in clickhouse? | 2020-08-19T08:53:14 | 2020-08-19T09:00:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,054,832 | 13,872 | filimonov | > c++: fatal error: Killed signal terminated program cc1plus
Memory issue? Check dmesg (oom killer), check if you have enough ram (or swap enabled) | 2020-08-19T09:50:25 | 2020-08-19T09:50:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,091,873 | 13,172 | excitoon | Continued in #13402 | 2020-08-19T10:18:31 | 2020-08-19T10:18:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,223,918 | 7,690 | fktr | Still reproduce on 20.6.3, any updates there? | 2020-08-18T03:13:22 | 2020-08-18T03:13:22 | {} | NONE |
ClickHouse | ClickHouse | 675,330,569 | 13,848 | IVitamin-C | After testing, it is found that if the newly added column is of string type, this problem will not occur. If it is of int type, it will appear, with 100% recurrence | 2020-08-18T08:12:44 | 2020-08-18T08:12:44 | {} | NONE |
ClickHouse | ClickHouse | 675,443,794 | 13,405 | akuzm | The Yandex synchronization failure is because the internal repo uses some old Poco version which lacks some methods:
```
$(SOURCE_ROOT)/clickhouse/src/IO/HTTPCommon.cpp:122:26: error: no member named 'setProxyProtocol' in 'Poco::Net::HTTPClientSession'
session->setProxyProtocol(proxy_scheme);
... | 2020-08-18T12:18:23 | 2020-08-18T12:18:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,752,360 | 13,792 | azat | > Fast test — fail: 1, passed: 1951, skipped: 234. Other tests won't be run before fast test will
Will rebase after #13867 | 2020-08-18T22:32:28 | 2020-08-18T22:32:28 | {} | MEMBER |
ClickHouse | ClickHouse | 676,069,606 | 13,875 | filimonov | Can you check (in the logs /var/log/clickhouse-server )
1) what was happening with that directory before (grep 20200818_6048_6048 )
2) show the stacktrace of error (lines after that error message till next start)
| 2020-08-19T10:02:40 | 2020-08-19T10:02:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 674,891,602 | 10,373 | azat | >@akuzm can you please take a look is it false-positive or not? (or maybe I should just restart the build? by i.e. rebasing against upstream/master)
Never mind, the problem is in the protocol ABI breakage (queries like `select * from system.numbers limit 1` via python clickhouse client works forever, because of the ... | 2020-08-17T13:46:13 | 2020-08-17T13:46:29 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 674,927,362 | 13,835 | den-crane | Yes.
And BTW, `MODIFY ORDER BY` does not change primary index, it changes only rows sorting (for new parts).
SHOW CREATE TABLE default.users_online
CREATE TABLE default.users_online
(
when DateTime,
uid UInt64,
duration UInt64,
version Int32
)
ENGINE = MergeTree()
PARTITION BY toYYYY... | 2020-08-17T14:49:52 | 2020-08-17T14:51:13 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 674,962,349 | 12,195 | akuzm | `clickhouse-local` segfaults on a query that is used to process perftest results, this is why we don't get a good error message. I found this in `compare.log`:
```
right/scripts/compare.sh: line 268: 3768 Aborted (core dumped) clickhouse-local --query "
-- Join the metric names back to the metric ... | 2020-08-17T15:53:09 | 2020-08-17T15:53:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,189,334 | 13,798 | judeshawn | > > Does the mechanical hard drive also have a big impact? raid 0
>
> mechanical hard drive does not affect on memory usage and the error "Allocator: Cannot malloc 4.00"
>
> mechanical hard drive is OK for ClickHouse from performance perspective.
> RAID of 12 HDD drives able to read 1GB/s -- it's close to a sing... | 2020-08-18T00:59:37 | 2020-08-18T00:59:37 | {} | NONE |
ClickHouse | ClickHouse | 675,372,826 | 13,816 | alexey-milovidov | It's in order of 200..300 MB/sec/core - more than enough.
Yes, it's a good idea to put it in query log. | 2020-08-18T09:33:52 | 2020-08-18T09:33:52 | {} | MEMBER |
ClickHouse | ClickHouse | 675,600,150 | 13,864 | filimonov | And single query log from both versions:
```
type: QueryFinish
event_date: 2020-08-18
event_time: 2020-08-18 17:01:12
query_start_time: 2020-08-18 17:01:12
query_duration_ms: 1
read_rows: 128
read_bytes: 2048
written_rows: 0
written_b... | 2020-08-18T17:04:03 | 2020-08-18T17:04:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,020,877 | 13,164 | RonanMorgan | ok so indeed I made a mistake using the ReplacingMergeTree, since the VIEW are working like a trigger, obviously I won't have the same results if I create my view before or after having filled the source table.
I also noted that creating a VIEW on a distributed table work only if the data is saved on the node where ... | 2020-08-19T09:25:22 | 2020-08-19T09:25:53 | {} | NONE |
ClickHouse | ClickHouse | 676,033,019 | 12,508 | alexey-milovidov | move_partition_from_table_zookeeper
https://clickhouse-test-reports.s3.yandex.net/13867/3ba9e9f6ced82b0c458bbce4bd99366baf6d5951/functional_stateless_tests_(release,_databaseatomic).html#fail1 | 2020-08-19T09:33:35 | 2020-08-19T09:33:35 | {} | MEMBER |
ClickHouse | ClickHouse | 675,243,200 | 13,848 | IVitamin-C | If the server restarts, if the table of new fields is not deleted, the server will not be able to start. If you execute "deache table", "DB:: exception: table columns structure in zookeeper is different from local table structure (version 20.3.5.21)" will be returned after "attach table on cluster testcluster" | 2020-08-18T04:33:18 | 2020-08-18T04:33:18 | {} | NONE |
ClickHouse | ClickHouse | 675,498,249 | 13,837 | vitlibar | > Actually there are lots of tests that don't use default config at all. You can wait for my fixes - I'll try to fix what have left ASAP.
Ok, I'll wait. | 2020-08-18T14:04:24 | 2020-08-18T14:04:24 | {} | MEMBER |
ClickHouse | ClickHouse | 676,062,522 | 13,876 | KochetovNicolai | Technically, inserts do not prevent mutations from executing. However, importing data increases the number of background merges. Merges and mutations uses the same pool of threads, so mutations may go slower or be delayed.
Try to increase `background_pool_size` up to 32, 64 or 128 (it will require to restart server)... | 2020-08-19T09:57:45 | 2020-08-19T09:57:45 | {} | MEMBER |
ClickHouse | ClickHouse | 675,313,234 | 13,848 | Fanduzi | I also face a similar situation in our RC environment.
```
bj2-all-clickhouse-test-02 :) select version();
SELECT version()
┌─version()─┐
│ 20.3.4.10 │
└───────────┘
```
we add one column to a table `A` yesterday,
>structure of table `A`
>```
>CREATE TABLE inno_topping_user.active_duration_local (
>`I... | 2020-08-18T07:38:57 | 2020-08-18T08:03:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,514,244 | 7,782 | amosbird | @abyss7 I just took another look and it seems this PR doesn't give much ROI. I'll close it and find another way. | 2020-08-18T14:31:36 | 2020-08-18T14:31:36 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 675,565,258 | 13,854 | KochetovNicolai | Introduced by https://github.com/ClickHouse/ClickHouse/pull/13151
Affected only 20.7 | 2020-08-18T15:58:13 | 2020-08-18T15:58:13 | {} | MEMBER |
ClickHouse | ClickHouse | 675,597,301 | 13,864 | filimonov | Logs:
```
ClickHouse client version 20.1.16.120 (official build).
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 20.1.16 revision 54431.
b8e8f21eaa7a :) SELECT col1, col2 FROM test PREWHERE col1 = '042fecec-ea68-1b26-00a4-a4f3b3987574';
SELECT
col1,
col2
FROM... | 2020-08-18T16:58:32 | 2020-08-18T16:58:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,662,330 | 13,857 | alexey-milovidov | Implementation specific behaviour for nans is generally Ok. | 2020-08-18T19:12:16 | 2020-08-18T19:12:16 | {} | MEMBER |
ClickHouse | ClickHouse | 675,865,158 | 13,873 | mjodeh | We faced this issue and we solved by controlling the insertion rate to be `1 bulk insertion per 2 seconds`
and we faced it only `system database` | 2020-08-19T05:46:34 | 2020-08-19T05:46:34 | {} | NONE |
ClickHouse | ClickHouse | 675,996,322 | 13,878 | hczhcz | @hexiaoting
Simply materialize what you extract from json. | 2020-08-19T09:08:19 | 2020-08-19T09:08:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 674,894,127 | 13,835 | den-crane | Code: 36. DB::Exception: Received from localhost:9000. DB::Exception: Existing column version is used in the expression that was added to the sorting key. **You can add expressions that use only the newly added columns.**
CREATE TABLE default.users_online
(
when DateTime,
uid UInt64,
duration UInt6... | 2020-08-17T13:50:36 | 2020-08-17T13:51:30 | {
"+1": 7
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 674,943,018 | 13,717 | 4ertus2 | Failed test is not related: `DB::Exception: Unknown function sumDistinct` | 2020-08-17T15:17:17 | 2020-08-17T15:17:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,326,558 | 13,844 | filimonov | It looks like warnings that should not affect the compilation.
Some libs which have no mac-related code will be empty (the list contains windows-related object files)
MacOS is not a primary platform for us, so we just check very basic things / do smoke tests - code compiles and can be started.
If you want to i... | 2020-08-18T08:04:43 | 2020-08-18T08:04:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,383,061 | 13,572 | alexey-milovidov | @abyss7 Your changes are inconsistent: 5ac2ed9 | 2020-08-18T09:55:13 | 2020-08-18T09:55:13 | {} | MEMBER |
ClickHouse | ClickHouse | 675,542,610 | 12,031 | bgranvea | does this PR fix this kind of errors?
> GRANT ALL ON `xxx`.* TO `yyyy`
Clickhouse error when executing command: Code: 446, Message: default: Introspection functions are disabled
(with Clickhouse 20.6.3.28) | 2020-08-18T15:18:22 | 2020-08-18T15:18:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,641,904 | 10,373 | azat | >Functional stateless tests (debug) — fail: 1, passed: 2168, skipped: 14
The 00956_sensitive_data_masking is flacky (#13867 ) | 2020-08-18T18:28:46 | 2020-08-18T19:10:57 | {} | MEMBER |
ClickHouse | ClickHouse | 676,056,617 | 13,873 | filimonov | Duplicate of https://github.com/ClickHouse/ClickHouse/issues/9637
Fixed in 20.4 (fix can't be backported)
| 2020-08-19T09:53:25 | 2020-08-19T09:53:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,376,778 | 13,779 | alexey-milovidov | @nikitamikhaylov TSan report in cache dictionary. | 2020-08-18T09:41:36 | 2020-08-18T09:41:36 | {} | MEMBER |
ClickHouse | ClickHouse | 675,633,803 | 13,156 | azat | >Performance — Timeout :(
@akuzm this PR changes the protocol but does not changes the [revision](https://github.com/ClickHouse/ClickHouse/blob/fc7fb3b10a4a6ae83c30b896f63ba322d233c6c7/cmake/autogenerated_versions.txt#L2) (since it should auto generated as stated by the file), but AFAICS [performance tests uses clic... | 2020-08-18T18:12:02 | 2020-08-18T18:12:02 | {} | MEMBER |
ClickHouse | ClickHouse | 675,662,037 | 13,857 | alexey-milovidov | I see, it is falling back to generic implementation if the type is Nullable,
and generic implementation uses the same comparison operator as for sorting, that requires total order. | 2020-08-18T19:11:36 | 2020-08-18T19:11:36 | {} | MEMBER |
ClickHouse | ClickHouse | 675,773,197 | 13,845 | den-crane | > `SELECT toDateTime('2020-03-21 00:00:00'),toDateTime('2020-03-21 01:00:00'),toDateTime('2020-03-21 23:00:00')`
>
> 
Do you mean it's a wrong day for the time shift ? | 2020-08-18T23:46:56 | 2020-08-18T23:46:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 675,828,932 | 13,845 | mbtolou | yes.
There is also another error ,that related to this error :
for example :
```
CREATE TABLE test.USE_T
(
`DATA_ID` String,
`DATET` DateTime('Asia/Tehran'),
`DELTA_A` Nullable(Decimal(12, 5))
)
ENGINE = ReplacingMergeTree(DATET)
PARTITION BY toYYYYMM(DATET)
ORDER BY (DATET, DATA_ID)
SETTIN... | 2020-08-19T03:23:43 | 2020-08-19T03:28:40 | {} | NONE |
ClickHouse | ClickHouse | 679,202,319 | 13,991 | zhang2014 | Maybe you can change the log level to debug? | 2020-08-24T15:37:15 | 2020-08-24T15:37:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 679,294,481 | 13,957 | den-crane | https://stackoverflow.com/questions/63541036/data-type-for-null-and-string-with-in-clickhouse | 2020-08-24T18:33:03 | 2020-08-24T18:33:03 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.