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 | 719,945,409 | 16,560 | vzakaznikov | > TestFlows failed due to RBAC tests, need to ask @vzakaznikov to proceed.
Timeout during executing `ALTER TABLE ... DROP COLUMN`. We have seen a similar issue with `ALTER TABLE ... CLEAR COLUMN`. Not sure why `ALTER` is timing out. Right now we are working to add instrumentation and dumping of `clickhouse-server.lo... | 2020-10-31T15:01:12 | 2020-10-31T15:02:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,021,066 | 16,564 | l1t1 | even use a sql instead of a datafile
```
# sqlite3 -noheader -separator , :memory: "select 'jerry' name, 9 age union select 'ox',1"
jerry,9
ox,1
# sqlite3 -noheader -separator , :memory: "select 'jerry' name, 9 age union select 'ox',1"| clickhouse-client --query="INSERT INTO some_table FORMAT CSV"
# clickhouse... | 2020-11-01T03:30:17 | 2020-11-01T03:30:17 | {} | NONE |
ClickHouse | ClickHouse | 720,123,938 | 16,581 | den-crane | я не прикалываюсь :
Interesting question what these functions should return in case of dates before 1970.
Negative number of seconds? And what it should be in case of toUInt256.
| 2020-11-01T17:35:00 | 2020-11-01T17:35:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,167,158 | 16,587 | den-crane | ```
20.11.1.5027.
cat test_data.csv.gz|gunzip |clickhouse-client -q 'insert into my_table format TSV'
select sumMap(sm) from (select sumMap([1],[1]) as sm from my_table group by col1, col2)
settings max_bytes_before_external_group_by = 100000;
┌─sumMap(sm)──────┐
│ ([1],[1785408]) │
└─────────────────┘
┌... | 2020-11-01T23:15:01 | 2020-11-01T23:34:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,268,516 | 16,588 | jjtjiang | @den-crane This is the result. thanks
desc test.hive_locks2
|name|type|default_type|default_expression|comment|codec_expression|ttl_expression|
|----|----|------------|------------------|-------|----------------|--------------|
|HL_LOCK_EXT_ID|Int64||||||
|HL_LOCK_INT_ID|Int64||||||
|HL_TXNID|Nullable(Int64)|||||... | 2020-11-02T06:34:48 | 2020-11-02T06:34:48 | {} | NONE |
ClickHouse | ClickHouse | 720,298,191 | 16,395 | macEar | It works without proxy confuguration, but when I specify proxy, it starts to log errors like before or so:
```
Error message: Poco::Exception. Code: 1000, e.code() = 0, e.displayText() = SSL Exception: error:1400010B:SSL routines::wrong version number (version 20.9.3.45 (official build))
```
I use the following... | 2020-11-02T07:43:30 | 2020-11-02T07:43:30 | {} | NONE |
ClickHouse | ClickHouse | 720,377,952 | 5,803 | alexey-milovidov | You can specify username and password in `credentials` as implemented here: https://github.com/ClickHouse/ClickHouse/pull/7092 | 2020-11-02T10:18:08 | 2020-11-02T10:18:08 | {} | MEMBER |
ClickHouse | ClickHouse | 719,562,453 | 6,472 | stephanieee777 | > @sifan0212 you should upgrade all nodes to 20.9.2.20
All nodes start from same docker images and have the same version(20.9.2.20). | 2020-10-30T13:47:48 | 2020-10-30T13:47:48 | {} | NONE |
ClickHouse | ClickHouse | 719,923,016 | 16,504 | tavplubix | Functional stateless tests - 01534_lambda_array_join - known issue
Timeouts - known issue
Waiting for Style check | 2020-10-31T11:47:36 | 2020-10-31T11:47:36 | {} | MEMBER |
ClickHouse | ClickHouse | 719,972,123 | 16,485 | socketpair | https://cmake.org/cmake/help/v3.0/prop_tgt/INTERPROCEDURAL_OPTIMIZATION.html
Cmake knows how to enable LTO both for clang and gcc. Is not it ?
So why not set it to true + check if it is supported. If not supported -- abort ? | 2020-10-31T18:42:35 | 2020-10-31T18:42:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,009,237 | 16,564 | l1t1 | sqlite have a command line to convert sql output to a csv, use this as the input
```
sqlite3 -noheader -separator ',' mydb.sqlite "select * from mytable limit 10"
0','0','???','g000:daccb7','True
0','4713','???','g001:1f24fa','False
101.4','4713.001014','FAIL','g001:e8aeab','False
1592.8','4713.015928','FAIL','g... | 2020-11-01T00:51:43 | 2020-11-01T02:53:50 | {} | NONE |
ClickHouse | ClickHouse | 720,150,865 | 16,585 | den-crane | Or you can configure TTL for this tables and remove old records automatically. | 2020-11-01T21:01:00 | 2020-11-01T21:01:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,312,322 | 16,586 | alexey-milovidov | @alesapin You mean clickhouse-format? | 2020-11-02T08:13:19 | 2020-11-02T08:13:19 | {} | MEMBER |
ClickHouse | ClickHouse | 719,684,317 | 12,085 | crakjie | Was this issue backported in version 20.3.20.6 ?
The changelog seems to say that was backported but I see no trace in the above log unlike v20.5.
| 2020-10-30T17:17:24 | 2020-10-30T17:17:24 | {} | NONE |
ClickHouse | ClickHouse | 719,736,227 | 10,293 | azat | Refs: #16121 | 2020-10-30T18:53:39 | 2020-10-30T18:53:39 | {} | MEMBER |
ClickHouse | ClickHouse | 719,828,960 | 16,560 | azat | >Functional stateless tests (release, DatabaseOrdinary) — fail: 1, passed: 2393, skipped: 4
Should be fixed by #16561 | 2020-10-30T22:32:21 | 2020-10-30T22:32:21 | {} | MEMBER |
ClickHouse | ClickHouse | 719,877,394 | 16,560 | alexey-milovidov | TestFlows failed due to RBAC tests, need to ask @vzakaznikov to proceed. | 2020-10-31T03:37:28 | 2020-10-31T03:37:28 | {} | MEMBER |
ClickHouse | ClickHouse | 720,122,310 | 16,581 | zvonand | > @zvonand you mean?
>
> select toInt128(now64())
>
> SELECT toUInt256(now64())
Yes
But this issue solely concerns PR #9404 | 2020-11-01T17:22:37 | 2020-11-01T17:22:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,310,105 | 15,951 | TCeason | The pr has been merged, long time waiting, bro. @liangli1213 If anything goes wrong with your test, @me. | 2020-11-02T08:08:31 | 2020-11-02T08:08:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,542,856 | 6,472 | den-crane | @sifan0212 you should upgrade all nodes to 20.9.2.20 | 2020-10-30T13:10:32 | 2020-10-30T13:10:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,547,964 | 2,301 | KochetovNicolai | Now I see 3 messages `Reading approx` in logs. Looks like now we execute 1 subquery for in instead of 2. It is probably syntax-level optimization.
I don;t see how new pipeline can help with this problem for now. | 2020-10-30T13:20:39 | 2020-10-30T13:20:39 | {} | MEMBER |
ClickHouse | ClickHouse | 719,589,261 | 16,555 | tavplubix | > But I don't understand why test fix helps. We will still have a replica that will dropped last?
All replicas will become inactive, so do not wait for the last one to execute some queue entry, because the last replica will be dropped soon too.
| 2020-10-30T14:35:44 | 2020-10-30T14:35:44 | {} | MEMBER |
ClickHouse | ClickHouse | 719,902,534 | 16,529 | azat | >@azat SELECT 1 query can took arbitrarly long time in Linux
@alexey-milovidov I do understand that this can be in theory, but is this something that will happen in practice?
>, (tens of minutes is totally realistic and easy reproducible).
I doubt that is can happen w/o:
- lowering cpu frequency
- CPU defici... | 2020-10-31T08:23:47 | 2020-10-31T08:23:47 | {} | MEMBER |
ClickHouse | ClickHouse | 720,019,601 | 16,564 | l1t1 | full steps
```sql
# sqlite3 test.sqlite
SQLite version 3.7.17 2013-05-20 00:56:22
Enter ".help" for instructions
Enter SQL statements terminated with a ";"
sqlite> create table a as select 'tom' name, 10 age union select 'apple',11;
sqlite> .exit
# clickhouse-client --query="create table some_table (a String... | 2020-11-01T03:10:25 | 2020-11-01T03:10:25 | {} | NONE |
ClickHouse | ClickHouse | 720,338,403 | 16,577 | tavplubix | Fixed in #16584 | 2020-11-02T09:04:10 | 2020-11-02T09:04:10 | {} | MEMBER |
ClickHouse | ClickHouse | 720,339,293 | 16,428 | filimonov | From an operational perspective `select * from system.jobs` to understand what is happening right now sounds very cool.
There are already quite a lot of different types of background operations (dictionary reload, moves, fetches, sending parts, executing distributed ddl, rabbit, kafka, mutations, merges, 'special' t... | 2020-11-02T09:05:44 | 2020-11-02T09:19:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,407,383 | 16,564 | alexey-milovidov | @siradjev Do you want to have support for data export as SQLite database with a single table and data import of single table from SQLite database? | 2020-11-02T11:14:13 | 2020-11-02T11:14:13 | {} | MEMBER |
ClickHouse | ClickHouse | 719,539,663 | 16,509 | alexey-milovidov | > 01133_max_result_rows
\- flaky, @KochetovNicolai | 2020-10-30T13:03:34 | 2020-10-30T13:03:42 | {} | MEMBER |
ClickHouse | ClickHouse | 719,565,694 | 11,387 | alesapin | Seems to be done? | 2020-10-30T13:53:54 | 2020-10-30T13:53:54 | {} | MEMBER |
ClickHouse | ClickHouse | 719,567,101 | 16,550 | alesapin | > Ok but I don't understand why it was wrong.
Because it's not a logical error, it's a user error. | 2020-10-30T13:56:32 | 2020-10-30T13:56:32 | {} | MEMBER |
ClickHouse | ClickHouse | 719,949,487 | 16,529 | azat | >Actually i know some people which were experimenting with that using query_log with engine = Null and materialized view on that.
Using Null engine as default does not looks sane.
>BTW - maybe MV on query_log created before the test can help collecting test-related query_log?
I tried, but creating MV each test... | 2020-10-31T15:32:34 | 2020-10-31T15:32:34 | {} | MEMBER |
ClickHouse | ClickHouse | 720,018,451 | 16,564 | l1t1 | and a sqlite datafile may include many tables, you cannot convert it to a table, but a database | 2020-11-01T02:55:33 | 2020-11-01T02:55:33 | {} | NONE |
ClickHouse | ClickHouse | 720,118,415 | 16,581 | den-crane | @zvonand you mean?
select toInt128(now64())
SELECT toUInt256(now64())
| 2020-11-01T16:55:46 | 2020-11-01T16:55:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,223,944 | 12,409 | amosbird | It's fixed in https://github.com/ClickHouse/ClickHouse/pull/13015 | 2020-11-02T04:02:45 | 2020-11-02T04:02:45 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 720,265,944 | 16,590 | alexey-milovidov | We don't need Tornado at all, because we build a static website.
I will merge this PR to avoid insulting @dependabot-preview and the people who enabled it. | 2020-11-02T06:28:31 | 2020-11-02T06:29:11 | {} | MEMBER |
ClickHouse | ClickHouse | 720,331,636 | 16,428 | filimonov | Related https://github.com/ClickHouse/ClickHouse/pull/7859 | 2020-11-02T08:51:51 | 2020-11-02T08:51:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,683,069 | 5,225 | tvsanthoshkumar | I am also facing the same problem.
@lzy305 Can you help us how did you solve this? | 2020-10-30T17:15:30 | 2020-10-30T17:15:30 | {} | NONE |
ClickHouse | ClickHouse | 719,727,010 | 16,252 | abyss7 | Looks like the query in your test hangs for more than 10 minutes for some reason | 2020-10-30T18:34:55 | 2020-10-30T18:34:55 | {
"confused": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,859,409 | 15,977 | TCeason | @tavplubix Hi, this pr's integration tests are passed. | 2020-10-31T00:49:43 | 2020-10-31T00:49:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,709,889 | 16,529 | azat | >Good question, not till c7ba289 (just added)
So here are some issues that pops up:
- `query_thread_log.query_execution_ms` is broken (the reason is the same as in c0e15ba348dae4030bd592ca1635fb3b09043afc), and this is a bugfix that should be done separately (and this is not only about logical split, but also to av... | 2020-10-30T18:00:50 | 2020-10-31T08:17:52 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 719,919,382 | 16,565 | UnamedRus | > Additionally, it would help to have DDL commands to allow combining partitions into one, and turning partitions into compact part via explicit command.
related
https://github.com/ClickHouse/ClickHouse/issues/13826
https://github.com/ClickHouse/ClickHouse/issues/8089 | 2020-10-31T11:13:54 | 2020-10-31T11:13:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,973,144 | 16,485 | alexey-milovidov | It's Ok, see https://clickhouse-test-reports.s3.yandex.net/16485/885bd847201befc61f0063a113ab2eb8ad575741/performance_comparison/report.html | 2020-10-31T18:51:56 | 2020-10-31T18:51:56 | {} | MEMBER |
ClickHouse | ClickHouse | 720,209,403 | 15,977 | TCeason | @tavplubix
00840_long_concurrent_select_and_drop_deadlock and 01534_lambda_array_join failed.
Functional stateless tests - 01534_lambda_array_join - known issue.
Is 00840 failed normal?
And I can't open `tress test (undefined) — Timeout :(`'s URL.
| 2020-11-02T02:56:57 | 2020-11-02T02:56:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,613,139 | 16,555 | tavplubix | ```
Functional stateless tests (release, DatabaseOrdinary) - 01079_parallel_alter_add_drop_column_zookeeper - Timeout
Functional stateless tests (release, wide parts enabled) - 01079_parallel_alter_modify_zookeeper - Timeout
```
:(
Have no ideas of what is happening | 2020-10-30T15:15:36 | 2020-10-30T15:15:36 | {} | MEMBER |
ClickHouse | ClickHouse | 720,257,523 | 11,596 | chang0391 | @compasses
If this function can be implemented, it will be very exciting for big data join | 2020-11-02T06:05:29 | 2020-11-02T06:05:29 | {} | NONE |
ClickHouse | ClickHouse | 720,265,951 | 16,590 | dependabot-preview[bot] | :wave: If you need additional help with Dependabot, please fill out GitHub's [Support form](https://support.github.com/contact?form%5Bsubject%5D=Re%3A+Dependabot+Preview) and your request will be routed to the right team at GitHub.
Be sure to include the details of any troubleshooting steps you've tried so far. | 2020-11-02T06:28:32 | 2020-11-02T06:28:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,293,462 | 16,428 | alesapin | > Maybe we should have a unified table for background tasks? For now we can fill it with merges and fetches, later we can add moves, directory monitor, Kafka and others? If we will create one table per task type we may get too many of them...
Clear idea, it's more ClickHouse's way to store everything in a single tab... | 2020-11-02T07:33:39 | 2020-11-02T07:33:39 | {} | MEMBER |
ClickHouse | ClickHouse | 719,552,096 | 1,576 | KochetovNicolai | > Scalar subqueries are still not integrated into query pipeline.
They cannot be cancelled and the progress is not shown.
Yes. The may difficulty using scalar subquery result in primary key condition.
Now we execute scalar subquery first, then analyze PK condition, then build a pipeline.
We may try to move `analy... | 2020-10-30T13:28:19 | 2020-10-30T13:28:50 | {} | MEMBER |
ClickHouse | ClickHouse | 719,728,663 | 14,414 | azat | - 00840_long_concurrent_select_and_drop_deadlock - https://clickhouse-test-reports.s3.yandex.net/16438/3661769e1b4fef9f89be6b562c2bdd4595301ff8/functional_stateless_tests_(thread).html#fail1
- test_input_format_parallel_parsing_memory_tracking/test.py::test_memory_tracking_total - https://clickhouse-test-reports.s3.ya... | 2020-10-30T18:38:18 | 2020-10-30T19:06:20 | {} | MEMBER |
ClickHouse | ClickHouse | 719,907,422 | 16,533 | l1t1 | ```
VM_0_52_centos :) explain pipeline SELECT uniqExact(key) FROM ( SELECT toString(number % 10000) AS key FROM numbers(100000000) ) ;
EXPLAIN PIPELINE
SELECT uniqExact(key)
FROM
(
SELECT toString(number % 10000) AS key
FROM numbers(100000000)
)
┌─explain───────────────────────┐
│ (Express... | 2020-10-31T09:14:14 | 2020-10-31T09:14:14 | {} | NONE |
ClickHouse | ClickHouse | 719,921,983 | 16,555 | tavplubix | Functional stateless tests - 00840_long_concurrent_select_and_drop_deadlock - known issue | 2020-10-31T11:38:08 | 2020-10-31T11:38:08 | {} | MEMBER |
ClickHouse | ClickHouse | 719,948,037 | 16,529 | filimonov | Actually i know some people which were experimenting with that using query_log with engine = Null and materialized view on that.
See https://github.com/ClickHouse/ClickHouse/blob/master/tests/queries/0_stateless/01360_materialized_view_with_join_on_query_log.sql
BTW - maybe MV on query_log created before the tes... | 2020-10-31T15:21:38 | 2020-10-31T15:21:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,972,308 | 16,485 | socketpair | Moreover, I did not check, but seems this PR turns off LTO even for clang | 2020-10-31T18:44:32 | 2020-10-31T18:44:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,215,827 | 1,540 | amosbird | It can be implemented after projection. We can have builtin/virtual projections with this information and automatically expands query pipelines. And it might be also possible to cooperate with minmax skip indices. | 2020-11-02T03:25:13 | 2020-11-02T03:30:03 | {
"+1": 1
} | COLLABORATOR |
ClickHouse | ClickHouse | 720,389,179 | 5,810 | alexey-milovidov | @kreuzerkrieg `-- --tcp_port=XXXX` is pushed into config and available in runtime through config object. | 2020-11-02T10:39:32 | 2020-11-02T10:39:32 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 719,711,209 | 16,557 | abyss7 | Anyway script doesn't seem to work. | 2020-10-30T18:03:24 | 2020-10-30T18:03:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,877,315 | 16,540 | l1t1 | my plan
```sql
EXPLAIN
SELECT
action1,
uniqMerge(uv) AS uv,
uniqMerge(session) AS session
FROM test_merge
GROUP BY action1
ORDER BY uv DESC
LIMIT 10
┌─explain──────────────────────────────────────────────────────────┐
│ Expression (Projection) │
│ Li... | 2020-10-31T03:36:24 | 2020-10-31T03:36:24 | {} | NONE |
ClickHouse | ClickHouse | 719,877,802 | 16,485 | alexey-milovidov | But we build with `clang` and it has different option to enable link-time optimization.
If developers will enable the old option, they will get misleading result. | 2020-10-31T03:41:24 | 2020-10-31T03:41:24 | {} | MEMBER |
ClickHouse | ClickHouse | 719,880,577 | 16,533 | l1t1 | because sort is slower than hash group?
```sql
localhost :) select count(distinct code) from sfz2;
SELECT countDistinct(code)
FROM sfz2
┌─uniqExact(code)─┐
│ 50000000 │
└─────────────────┘
1 rows in set. Elapsed: 6.185 sec. Processed 50.00 million rows, 1.29 GB (8.08 million rows/s., 209.37 MB/s.)... | 2020-10-31T04:11:35 | 2020-10-31T04:11:35 | {} | NONE |
ClickHouse | ClickHouse | 719,881,138 | 16,533 | l1t1 | the time differenct of your example in my host is more
```
VM_0_52_centos :) set max_threads=1;
SET max_threads = 1
Ok.
0 rows in set. Elapsed: 0.018 sec.
VM_0_52_centos :) SELECT uniqExact(key)
:-] FROM
:-] (
:-] SELECT toString(number % 10000) AS key
:-] FROM numbers(100000000)
:-] );
SE... | 2020-10-31T04:17:22 | 2020-10-31T04:17:22 | {} | NONE |
ClickHouse | ClickHouse | 719,905,722 | 16,537 | xiashaohua | > 20.7.1.1 -- is obsolete and not stable. You need to upgrade at least to 20.7.4.11. Or better to 20.8.5.45.
thanks,i will update to a stable version | 2020-10-31T08:56:44 | 2020-10-31T08:56:44 | {} | NONE |
ClickHouse | ClickHouse | 720,225,526 | 16,588 | den-crane | @jjtjiang can you show an output of `desc test.hive_locks2` and `select * from system.settings where changed` | 2020-11-02T04:10:28 | 2020-11-02T04:10:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,307,326 | 16,586 | alesapin | Maybe you can add the `.sh` test? Just count whitespaces in clickhouse-client output? | 2020-11-02T08:02:40 | 2020-11-02T08:02:40 | {} | MEMBER |
ClickHouse | ClickHouse | 720,311,194 | 16,586 | kitaisreal | @alesapin okay, sure. | 2020-11-02T08:10:56 | 2020-11-02T08:10:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,313,965 | 16,586 | alesapin | > @alesapin You mean clickhouse-format?
Yes | 2020-11-02T08:16:44 | 2020-11-02T08:16:44 | {} | MEMBER |
ClickHouse | ClickHouse | 720,370,201 | 16,395 | macEar | I see response code 400 with such messages (took 2 messages for example) on the proxy:
```html
<p>The following error was encountered while trying to retrieve the URL: <a href="L-j£^7æ&# 31;&# 6;&# 5;»½•Ì">L-j£^7æ&# 31;&# 6;&# 5;»½•Ì&... | 2020-11-02T10:03:39 | 2020-11-02T10:03:39 | {} | NONE |
ClickHouse | ClickHouse | 720,377,497 | 16,548 | alesapin | Actually your materialized view table key quite strange. AggregateFunction states don't have a comparison operator and don't work in the primary key (ORDER BY). Just rewrite your matview schema to:
```
CREATE MATERIALIZED VIEW foobar
ENGINE = AggregatingMergeTree() ORDER BY (
id
)
POPULATE
AS SELECT
id,... | 2020-11-02T10:17:16 | 2020-11-02T10:17:32 | {} | MEMBER |
ClickHouse | ClickHouse | 719,581,584 | 1,576 | alexey-milovidov | This all was one of the motivation to adding the new query pipeline :)
To allow dynamic pipeline rebuild after execution of scalar subqueries.
Imagine you can build a pipeline that will do only scalar subqueries first - and after they are done, build the main query pipeline. | 2020-10-30T14:22:27 | 2020-10-30T14:22:48 | {
"eyes": 1
} | MEMBER |
ClickHouse | ClickHouse | 719,724,294 | 16,559 | abyss7 | Do not merge without review. | 2020-10-30T18:29:24 | 2020-10-30T18:29:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,828,404 | 16,558 | azat | >Functional stateless tests (release) — fail: 2, passed: 2397
Some tests failed due to memory limit exceeded, because the query with the `max_memory_limit_for_user`, from `01541_max_memory_usage_for_user`, is still not killed - should be fixed by #16561 | 2020-10-30T22:30:31 | 2020-10-30T22:30:50 | {} | MEMBER |
ClickHouse | ClickHouse | 719,845,024 | 16,529 | alexey-milovidov | @azat `SELECT 1` query can took arbitrarly long time in Linux, (tens of minutes is totally realistic and easy reproducible). | 2020-10-30T23:36:01 | 2020-10-30T23:36:01 | {} | MEMBER |
ClickHouse | ClickHouse | 720,293,878 | 16,428 | alesapin | I'll merge it, if we will found some strong ideas against separate table then I'll change it before release. | 2020-11-02T07:34:34 | 2020-11-02T07:34:34 | {} | MEMBER |
ClickHouse | ClickHouse | 720,380,184 | 16,548 | alesapin | Actually it doesn't work even if you simple detach/attach the inner table for matview or restart server. | 2020-11-02T10:22:31 | 2020-11-02T10:22:31 | {} | MEMBER |
ClickHouse | ClickHouse | 719,490,706 | 16,542 | Fanduzi | I upgraded all nodes, and now alter table delete can be executed normally | 2020-10-30T11:09:29 | 2020-10-30T11:09:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,579,772 | 11,387 | alexey-milovidov | :tada: | 2020-10-30T14:19:19 | 2020-10-30T14:19:19 | {} | MEMBER |
ClickHouse | ClickHouse | 719,806,976 | 16,527 | kitaisreal | @alexey-milovidov thanks for review. I introduces getPageSize function and unified usages of getpagesize and sysconf(_SC_PAGESIZE).
I also replaced most usages of constant 4096 with getPageSize, in places where it probably was used in assumption of memory page size. Probably in some places I should not replace it, pl... | 2020-10-30T21:25:25 | 2020-10-30T21:25:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,988,485 | 16,563 | azat | >Integration tests (release) Pending — Started
>Integration tests (thread) Pending — Started
Hanged? In the queue? | 2020-10-31T21:07:17 | 2020-10-31T21:07:17 | {} | MEMBER |
ClickHouse | ClickHouse | 720,040,399 | 16,578 | azat | >Where can I find tests for ZlibInflatingReadBuffer.h?I grepped tests/ folder and found nothing
```
src/IO/tests/zlib_buffers.cpp:#include <IO/ZlibInflatingReadBuffer.h>
src/IO/tests/zlib_buffers.cpp: DB::ZlibInflatingReadBuffer inflating_buf(std::move(buf), DB::CompressionMethod::Gzip);
``` | 2020-11-01T06:18:46 | 2020-11-01T06:18:46 | {} | MEMBER |
ClickHouse | ClickHouse | 720,126,076 | 16,581 | zvonand | > Interesting question what these functions should return in case of dates before 1970.
I guess in case of `Int` it must be a negative number of seconds, like it already is with `toUnixTimestamp64Milli`:
`SELECT toUnixTimestamp64Milli(toDateTime64('1969-12-31 23:59:00', 0, 'UTC'))` gives us `-60000`
Moreover, ... | 2020-11-01T17:51:58 | 2020-11-01T17:53:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,536,839 | 16,537 | den-crane | It seems this issue will be fixed in the next stable release 20.11 | 2020-10-30T12:57:48 | 2020-10-30T12:57:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,549,557 | 16,273 | alexey-milovidov | `string_sort` tests has slowed down 5% and `collations` has slowed down 2%. But it is below the radars.
| 2020-10-30T13:23:32 | 2020-10-30T13:23:32 | {} | MEMBER |
ClickHouse | ClickHouse | 719,907,223 | 16,533 | l1t1 | ```
VM_0_52_centos :) explain header=1
:-] SELECT uniqExact(key)
:-] FROM
:-] (
:-] SELECT toString(number % 10000) AS key
:-] FROM numbers(100000000)
:-] )
:-] ;
EXPLAIN header = 1
SELECT uniqExact(key)
FROM
(
SELECT toString(number % 10000) AS key
FROM numbers(100000000)
)
┌─explai... | 2020-10-31T09:12:15 | 2020-10-31T09:12:15 | {} | NONE |
ClickHouse | ClickHouse | 719,921,673 | 14,221 | azat | > 01455_shard_leaf_max_rows_bytes_to_read https://clickhouse-test-reports.s3.yandex.net/0/651518566a527f8a0c225f1d8a130712c8b810ee/functional_stateless_tests_(address).html
The test from this PR reported to #14414, and there is how I understand the problem:
This happened because the following happens (in the orde... | 2020-10-31T11:35:28 | 2020-10-31T11:35:28 | {} | MEMBER |
ClickHouse | ClickHouse | 720,149,799 | 16,585 | den-crane | `truncate table system.asynchronous_metric_log;`
And you can disable this logging by
```
cat /etc/clickhouse-server/conf.d/asynchronous_metric_log_disable.xml
<?xml version="1.0"?>
<yandex>
<asynchronous_metric_log remove="1">
</asynchronous_metric_log>
</yandex>
``` | 2020-11-01T20:51:53 | 2020-11-01T20:52:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,149,980 | 16,585 | den-crane | And you can disables other loggers
metric_log
query_log
query_thread_log
trace_log
| 2020-11-01T20:53:32 | 2020-11-01T20:53:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 720,374,109 | 16,564 | siradjev | thanks @l1t1 for examples!
So the main idea is to do it simpler, similar to CSV, TSV, Parquet, ORC, (they all have similar command line tools, but having it working as part of clickhouse binary provides much better efficiency). | 2020-11-02T10:10:51 | 2020-11-02T10:10:51 | {} | NONE |
ClickHouse | ClickHouse | 719,877,475 | 16,561 | alexey-milovidov | Fuzzer has found existing issue: #13281 | 2020-10-31T03:38:19 | 2020-10-31T03:38:19 | {} | MEMBER |
ClickHouse | ClickHouse | 713,181,139 | 12,741 | azat | >Double Distributed JOINs doesn't work correctly with row policy and allow_insecure_prewhere=1:
This is not strictly related to this PR, this is due it does not work even w/o row-level security with PREWHERE in 20.3:
```sql
-- example w/ allow_insecure_prewhere=0
-- and w/o row-level security
SELECT *
FROM ... | 2020-10-20T22:41:50 | 2020-10-31T12:32:26 | {} | MEMBER |
ClickHouse | ClickHouse | 720,407,944 | 16,564 | alexey-milovidov | We will collect "likes" to this task to estimate the demand. If you need this feature, add "thumbs up" reaction to the first message. | 2020-11-02T11:15:13 | 2020-11-02T11:15:13 | {} | MEMBER |
ClickHouse | ClickHouse | 719,584,064 | 2,301 | alexey-milovidov | A test:
1. `SELECT toUInt64(1) IN (SELECT * FROM numbers(10000000))`
2. `SELECT toUInt64(1) IN (SELECT * FROM numbers(10000000)), toUInt64(2) IN (SELECT * FROM numbers(10000000))`
3. `SELECT toUInt64(1) IN (SELECT * FROM numbers(10000000)), toUInt64(2) IN (SELECT * FROM numbers(10000001))` | 2020-10-30T14:26:41 | 2020-10-30T14:26:41 | {} | MEMBER |
ClickHouse | ClickHouse | 720,229,095 | 9,887 | ucasfl | > Interval syntax like '+ 30 days' is not supported
This task has been done in [15978](https://github.com/ClickHouse/ClickHouse/pull/15978) @alex-zaitsev | 2020-11-02T04:27:39 | 2020-11-02T04:27:39 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 719,550,134 | 16,273 | alexey-milovidov | The worst is `SELECT PageCharset FROM hits_100m_single ORDER BY PageCharset, CounterID LIMIT 10 format Null`.
It is slowed down 15%. | 2020-10-30T13:24:41 | 2020-10-30T13:24:41 | {} | MEMBER |
ClickHouse | ClickHouse | 719,579,373 | 16,550 | alexey-milovidov | I thought it should be catched earlier in the control flow. | 2020-10-30T14:18:35 | 2020-10-30T14:18:46 | {} | MEMBER |
ClickHouse | ClickHouse | 719,725,033 | 16,283 | abyss7 | Looks like you broke some other tests | 2020-10-30T18:30:54 | 2020-10-30T18:30:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,726,331 | 16,438 | azat | >Integration tests (release) Pending — Started
Hanged?
>Functional stateless tests (thread) — fail: 1, passed: 2384, skipped: 14
`00840_long_concurrent_select_and_drop_deadlock` does not looks related
>Integration tests (thread) — fail: 1, passed: 924, error: 0
`test_input_format_parallel_parsing_memory_... | 2020-10-30T18:33:32 | 2020-10-30T19:06:04 | {} | MEMBER |
ClickHouse | ClickHouse | 719,948,378 | 16,574 | zlobober | A few more examples (note that error differs in following cases):
```
max42-dev.sas.yp-c.yandex.net :) select sumIf(42, toNullable(1)) from DT
SELECT sumIf(42, toNullable(1))
FROM DT
↓ Progress: 5.00 rows, 45.00 B (47.39 rows/s., 426.47 B/s.)
Received exception from server (version 20.8.1):
Code: 42. DB::... | 2020-10-31T15:24:30 | 2020-10-31T15:24:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 719,949,102 | 16,428 | filimonov | Maybe we should have a unified table for background tasks? For now we can fill it with merges and fetches, later we can add moves, directory monitor, Kafka and others? If we will create one table per task type we may get too many of them... | 2020-10-31T15:29:50 | 2020-10-31T15:29:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 714,164,358 | 16,251 | keliss | Also, the message telling me to "increase (or set to zero) max_[table/partition]_size_to_drop in server config and restart ClickHouse" seems to be misleading - I don't have to restart CH for these settings to apply. Or it's just some CH magic that made the partitions smaller so I was able to drop them normally (I've al... | 2020-10-22T01:41:17 | 2020-10-22T01:43:01 | {} | NONE |
ClickHouse | ClickHouse | 714,206,112 | 14,205 | zhang2014 | The MaterializeMySQL database engine uses the MySQL protocol to access MySQL historical data and consume MySQL new data(so, we need to ensure that ClickHouse has access to the given MySQL port and the given MySQL user must have the required privileges. required privileges: `SELECT`, `RELOAD`, `SUPER`, `REPLICATION SLAV... | 2020-10-22T03:59:09 | 2020-10-22T03:59:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 714,306,437 | 14,731 | azat | >No news so far.
@alexey-milovidov so what is the fate of this PR?
Due to it changes the interface and some internal code uses them and breaks it cannot be applied? | 2020-10-22T07:55:50 | 2020-10-22T07:55:59 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.