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 | 526,854,745 | 6,732 | foxit337 | 1. 0
2. no | 2019-08-31T18:24:41 | 2019-08-31T18:24:41 | {} | NONE |
ClickHouse | ClickHouse | 526,918,370 | 6,765 | den-crane | Все как я и написал
Сейчас (починено https://github.com/yandex/ClickHouse/pull/6249 ):
```
SELECT *
FROM tbl_1
LEFT JOIN tbl_2 ON tbl_1.int_col = tbl_2.int_col
┌─int_col─┬─tbl_2.int_col─┐
│ 1 │ ᴺᵁᴸᴸ │
│ 2 │ ᴺᵁᴸᴸ │
│ 3 │ 3 │
│ 4 │ 4 │
│ ... | 2019-09-01T13:16:38 | 2019-09-01T13:32:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,243,361 | 525 | inv2004 | https://github.com/yandex/ClickHouse/issues/4860 | 2019-09-02T21:09:51 | 2019-09-02T21:09:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,355,988 | 6,789 | blinkov | хорошо б их ещё ссылками на страницы про движки сделать | 2019-09-03T08:18:02 | 2019-09-03T08:18:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,893,536 | 6,707 | filimonov | Http status is the first line of http responce. If you sent at least one header that means you already sent a status.
Progress headers is sent in streaming mode. | 2019-09-01T07:12:01 | 2019-09-01T07:14:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,096,730 | 6,746 | simPod | @alesapin just mailed the dropbox link to the email address given. | 2019-09-02T10:31:06 | 2019-09-02T10:31:06 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,221,630 | 6,780 | alexey-milovidov | Your own test has failed. | 2019-09-02T18:41:05 | 2019-09-02T18:41:05 | {} | MEMBER |
ClickHouse | ClickHouse | 527,224,768 | 1,522 | alexey-milovidov | https://github.com/yandex/ClickHouse/issues/5891 | 2019-09-02T18:59:08 | 2019-09-02T18:59:08 | {} | MEMBER |
ClickHouse | ClickHouse | 527,226,649 | 2,282 | alexey-milovidov | The issue is no longer relevant. I will add a test.
But please note that materialized view with JOIN will be updated only on insert into left table. | 2019-09-02T19:10:28 | 2019-09-02T19:10:28 | {} | MEMBER |
ClickHouse | ClickHouse | 527,232,604 | 6,783 | alexey-milovidov | It isn't fixed. | 2019-09-02T19:47:28 | 2019-09-02T19:47:28 | {} | MEMBER |
ClickHouse | ClickHouse | 527,234,520 | 6,629 | alexey-milovidov | Could you please merge with master once again? | 2019-09-02T20:01:21 | 2019-09-02T20:01:21 | {} | MEMBER |
ClickHouse | ClickHouse | 527,244,080 | 6,732 | filimonov | `free`? | 2019-09-02T21:16:06 | 2019-09-02T21:16:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,623,174 | 6,698 | inv2004 | @alexey-milovidov
There is also another problem: clickhouse-client and clickhouse performance-test have different default-values. | 2019-08-30T14:29:15 | 2019-08-30T14:30:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,654,284 | 6,731 | igor-sh8 | Yes. All column names are identical.
We checked simple MV and MV to Table - result the same.
If we run select query (select part from creating MV query) from table with Kafka Engine - all works. | 2019-08-30T15:55:14 | 2019-08-30T15:55:14 | {} | NONE |
ClickHouse | ClickHouse | 526,828,928 | 6,759 | alexey-milovidov | Run `./stress` and
```
watch -n1 -x clickhouse-client --query "SELECT elapsed, ProfileEvents.Values[indexOf(ProfileEvents.Names, 'UserTimeMicroseconds')] / 1000 AS user_ms, arraySum(arrayFilter(x, y -> y LIKE 'RWLock%WaitMilliseconds', ProfileEvents.Values, ProfileEvents.Names)) AS lock_ms, substring(query, 1, 200) F... | 2019-08-31T13:01:17 | 2019-08-31T13:01:31 | {} | MEMBER |
ClickHouse | ClickHouse | 526,873,971 | 6,764 | alexey-milovidov | ```
┌────────elapsed─┬─os_thread_ids─────────────┬─query───────────────────────────────────────────────────────────────────┐
│ 1417.381492847 │ [22720] │ ALTER TABLE a MODIFY COLUMN `x` Nullable(UInt8) │
│ 1417.3811985 │ [22721,22759] │ SELECT x FROM a WHERE x ... | 2019-08-31T23:59:01 | 2019-08-31T23:59:01 | {} | MEMBER |
ClickHouse | ClickHouse | 526,925,931 | 6,758 | Tasselmi | > HashJoin
Thanks for your reply.
If I use 20 shards and 1 replica.
Is this plan realistic?
We do not have much money to buy machines. | 2019-09-01T14:57:07 | 2019-09-01T14:57:07 | {} | NONE |
ClickHouse | ClickHouse | 527,119,844 | 6,746 | alesapin | @simPod, does segfault reproduced only once or it's constant error? | 2019-09-02T11:53:36 | 2019-09-02T11:53:36 | {} | MEMBER |
ClickHouse | ClickHouse | 527,234,722 | 6,205 | alexey-milovidov | It is slightly incorrectly merged:
```
2019.08.29 08:46:45.606915 [ 174 ] {} <Fatal> BaseDaemon: ########################################
2019.08.29 08:46:45.607293 [ 174 ] {} <Fatal> BaseDaemon: (version 19.14.1.1207) (from thread 36) Received signal Segmentation fault (11).
2019.08.29 08:46:45.607567 [ 174 ] {}... | 2019-09-02T20:02:58 | 2019-09-02T20:02:58 | {} | MEMBER |
ClickHouse | ClickHouse | 526,597,306 | 6,729 | akuzm | `SELECT number % 10000000 AS k, count() FROM system.numbers GROUP BY k` is failing with x0.88 slowdown.
OTOH, another query shows x1.26 speedup: `SELECT URL, count() AS PageViews FROM hits_100m_single WHERE CounterID = 34 AND EventDate >= '2013-07-01' AND EventDate <= '2013-07-31' AND NOT Refresh AND IsLink AND NOT ... | 2019-08-30T13:14:52 | 2019-08-30T13:14:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,631,267 | 6,746 | simPod | @alesapin what exactly do you mean by that? | 2019-08-30T14:50:22 | 2019-08-30T14:50:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,633,815 | 6,588 | thyn | > @thyn There is a way to move table to a separate device:
>
> 1. Stop `clickhouse-server`.
> 2. Move table data directory to desired location.
> 3. Create symlink in a place of previous directory that points to new place. Make sure it is accessible to `clickhouse` user.
> 4. Start `clickhouse-server`.
I need ... | 2019-08-30T14:57:19 | 2019-08-30T14:57:19 | {} | NONE |
ClickHouse | ClickHouse | 526,846,917 | 6,758 | Tasselmi | > > 2TB logs data one day
>
> Raw or compressed? In average CH able to compress 5-10 times.
> How many days are you going to store?
>
> > how many cores and ram volume do I need in total ?
>
> It's impossible to evaluate because the answer depends of your requirements to retention / latency / accuracy (CH has... | 2019-08-31T16:54:38 | 2019-08-31T16:54:38 | {} | NONE |
ClickHouse | ClickHouse | 526,955,368 | 6,656 | alexey-milovidov | I have to temporarily disable all LIVE VIEW tests because we need to be 100% green to discover real issues, but I expect it is possible to fix ASAP. | 2019-09-01T21:45:20 | 2019-09-01T21:45:20 | {} | MEMBER |
ClickHouse | ClickHouse | 527,146,339 | 6,773 | alexey-milovidov | Ok. But test is missing.
Could you please make a simple functional test
(it can involve `remote('127.0.0.{2,3}', ...` to emulate distributed queries, see examples around; it should contain `shard` in it's name as a tag that it performs distributed query) | 2019-09-02T13:21:23 | 2019-09-02T13:21:23 | {} | MEMBER |
ClickHouse | ClickHouse | 526,586,194 | 6,549 | thyn | I'm using freeze partitions to backup, every day
> > -r-xr-xr-x 1 clickhouse clickhouse 48 Aug 29 09:58 /data/clickhouse/data/actions/actions/20190708_786_786_0_263702/quality.mrk
>
> The issue might be somewhat related (but should not) to FREEZE partitions.
| 2019-08-30T12:40:18 | 2019-08-30T13:34:12 | {} | NONE |
ClickHouse | ClickHouse | 526,657,595 | 6,707 | jcdang | If any of the steps fail we regard the entire run as a failure and just do the cleanup. | 2019-08-30T16:05:07 | 2019-08-30T16:05:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,913,215 | 6,765 | den-crane | Пример? Версия КХ?
Была проблема с ключевыми полями -- приходилось алиасить, починено месяц назад.
```
create table t(a Int64, b Int64) engine = TinyLog;
insert into t values(1,1);
create table s(a Int64, b Int64) engine = TinyLog;
set join_use_nulls = 1
select t.a, t.b, s.a, s.alias_a, s.b from t left join ... | 2019-09-01T12:04:42 | 2019-09-01T12:08:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,946,293 | 2,480 | alexey-milovidov | I'm definitely sure it is fixed at least in version 19.14-testing. | 2019-09-01T19:26:03 | 2019-09-01T19:26:03 | {} | MEMBER |
ClickHouse | ClickHouse | 526,954,789 | 6,656 | alexey-milovidov | There are still some issues with LIVE VIEWs:
```
2019.09.01 23:28:44.803583 [ 2476 ] {} <Fatal> BaseDaemon: ########################################
2019.09.01 23:28:44.803678 [ 2476 ] {} <Fatal> BaseDaemon: (version 19.14.1.1232) (from thread 2072) Received signal Segmentation fault (11).
2019.09.01 23:28:44.803... | 2019-09-01T21:35:17 | 2019-09-01T21:35:17 | {} | MEMBER |
ClickHouse | ClickHouse | 527,065,695 | 6,747 | lmikhailov | Thanks for clarification! | 2019-09-02T09:00:18 | 2019-09-02T09:00:18 | {} | NONE |
ClickHouse | ClickHouse | 526,599,851 | 2,353 | 4ertus2 | I don't think we should do anything here. It's possible to call any column as a function in common DB systems.
```
mysql> create table test.t(`count()` INT);
Query OK, 0 rows affected (0.08 sec)
mysql> insert into test.t (`count()`) values(1);
Query OK, 1 row affected (0.01 sec)
mysql> select * from test.t;
... | 2019-08-30T13:22:50 | 2019-08-30T13:22:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,606,322 | 6,746 | alexey-milovidov | More details:
```
$ addr2line -afiCe clickhouse 0x6c96f1b 0x753ac60
0x0000000006c96f1b
DB::DataTypeNullable::enumerateStreams(std::function<void (std::vector<DB::IDataType::Substream, std::allocator<DB::IDataType::Substream> > const&)> const&, std::vector<DB::IDataType::Substream, std::allocator<DB::IDataType::Su... | 2019-08-30T13:42:23 | 2019-08-30T13:42:23 | {} | MEMBER |
ClickHouse | ClickHouse | 526,608,436 | 6,549 | thyn | > > But I see that I have more then 20 processes of clickhouse server.
>
> Please recheck. It should be one process (and lot of threads).
Yes you are right. Turned on htop tree view .
Can't reproduce for now. Everything is updated ok. Backups already finished and shadow files were deleted. Will try tomorrow | 2019-08-30T13:48:48 | 2019-08-30T13:48:48 | {} | NONE |
ClickHouse | ClickHouse | 526,916,173 | 6,765 | PlatonB | Дополнил. | 2019-09-01T12:47:19 | 2019-09-01T12:47:19 | {} | NONE |
ClickHouse | ClickHouse | 527,094,874 | 6,772 | yingfeng | Hi, @filimonov
Let me tell how Converthouse comes. The work of Converthouse is based on the previously encapsulated library [Prophet](https://github.com/infinivision/prophet), which was initiated due to the requirements to deliver a cloud-native distributed no-sql store [Elasticell](https://github.com/deepfabric/elas... | 2019-09-02T10:24:52 | 2019-09-02T10:24:52 | {} | NONE |
ClickHouse | ClickHouse | 527,191,016 | 6,667 | akuzm | Looking at the performance test results, many DISTINCT queries show an improvement of 5-10%, while some GROUP BY queries regress. The most regressing one is `SELECT URL, count() AS PageViews FROM hits_100m_single WHERE CounterID = 34 AND EventDate >= '2013-07-01' AND EventDate <= '2013-07-31' AND NOT Refresh AND IsLink... | 2019-09-02T15:50:17 | 2019-09-02T15:50:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,630,837 | 6,746 | alesapin | @simPod Can you provide table partitions? | 2019-08-30T14:49:13 | 2019-08-30T14:49:13 | {} | MEMBER |
ClickHouse | ClickHouse | 526,920,981 | 6,765 | den-crane | >А как это у Вас с моим вариантом инструкции вывелись два NULL,
>а не два нуля? У меня стабильно нули получаются.
Эммм. Я написал что это уже починено месяц назад. Логично что я проверяю на починеной версии? | 2019-09-01T13:52:51 | 2019-09-01T13:52:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,089,266 | 6,745 | filimonov | Yes. | 2019-09-02T10:07:16 | 2019-09-02T10:07:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,164,881 | 6,746 | alesapin | reproduced on my machine. | 2019-09-02T14:20:13 | 2019-09-02T14:20:13 | {} | MEMBER |
ClickHouse | ClickHouse | 527,326,615 | 6,335 | filimonov | In other words - hostnames of servers were not unique? | 2019-09-03T06:43:28 | 2019-09-03T06:43:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,643,627 | 3,140 | 4ertus2 | I've tried to implement a simple solution and it's not working. To much places to change.
It's probably better to fix it with #4219: attach a hidden column with alias into select list and remove it from output. | 2019-08-30T15:25:20 | 2019-08-30T15:27:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,838,902 | 6,758 | den-crane | >2TB logs data one day
Raw or compressed? In average CH able to compress 5-10 times.
How many days are you going to store?
>how many cores and ram volume do I need in total ?
It's impossible to evaluate because the answer depends of your requirements to retention / latency / accuracy (CH has sampling and able... | 2019-08-31T15:21:32 | 2019-08-31T15:30:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,955,271 | 6,656 | alexey-milovidov | I have no idea what happens inside LIVE VIEW (why we need double shared_ptr, etc.)
```
{
/// add user to the blocks_ptr
std::shared_ptr<BlocksPtr> stream_blocks_ptr = blocks_ptr;
{
std::lock_guard lock(mutex);
if (!(*blocks_ptr))
{
if (getNewBlocks())
... | 2019-09-01T21:43:58 | 2019-09-01T21:43:58 | {} | MEMBER |
ClickHouse | ClickHouse | 527,018,680 | 6,745 | xuxudede | > format_csv_delimiter
--format_csv_delimiter="|" | 2019-09-02T06:26:29 | 2019-09-02T06:26:29 | {} | NONE |
ClickHouse | ClickHouse | 527,125,096 | 6,772 | zhangxu19830126 | Hi, @filimonov
The design of the Converthouse is very similar to the Cloud-Table. Is the solution for the Cloud-Table implementation determined? This blog was post a long time ago.
When we first considered Converthouse, we considered that the design changes were relatively small. Can you tell me which parts are ca... | 2019-09-02T12:12:24 | 2019-09-02T12:12:24 | {} | NONE |
ClickHouse | ClickHouse | 527,202,224 | 6,335 | astudnev | I found probable reason for this behavior. The deal is that i use short names for servers ( like ch1..chN)
which are resolved to different IPs on servers. In some configuration it may be a normal situation.
Look like this was the reason, as after i changed the short names to unique, the issue dissapeared | 2019-09-02T16:42:01 | 2019-09-02T16:42:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,207,181 | 6,758 | den-crane | Start with POC - 8 serves (4 shards *2 replica) with RAID10 on 12 HDD disks. | 2019-09-02T17:09:10 | 2019-09-02T17:09:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,607,781 | 6,741 | hczhcz | `min()` is not a math operation as we usually understand. It is a selection operation which is similar to `any()`. It applies to any comparable value including strings, tuples, etc.
Thus, `min` value on an empty list is not `nan` but something "undefined". The default value `0` is used to comply with the SQL standar... | 2019-08-30T13:46:50 | 2019-08-30T13:46:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,677,216 | 6,746 | simPod | @alesapin do you have any simple guide, how to do that? Is SQL insert enought? Any place where to send it, I'd prefer not to upload it to github. | 2019-08-30T17:07:28 | 2019-08-30T17:07:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,836,104 | 6,707 | filimonov | Comment one line above says "if the data was not sent to client".
Check docs / code regarding `send_progress_in_http_headers`, `buffer_size` and `wait_end_of_query` | 2019-08-31T14:44:52 | 2019-08-31T14:44:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,872,190 | 6,707 | jcdang | `send_progress_in_http_headers` This is headers and doesn't send a body.
`wait_end_of_query ` This is if exceeding the buffer_size will store to file.
`buffer_size` This doesn't trigger a flush to out outputstream if exceeded.
Once you send the response body, you're not allowed to send headers anymore. But what y... | 2019-08-31T23:15:08 | 2019-08-31T23:15:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,163,332 | 6,746 | simPod | I managed to reproduce, sending new email with data. | 2019-09-02T14:15:10 | 2019-09-02T14:15:10 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,610,092 | 6,731 | filimonov | And do column names also match (inserted stream = Kafka table definition = MV definition) ? | 2019-08-30T13:53:10 | 2019-08-30T13:53:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,653,558 | 6,746 | alesapin | @simPod is it possible? | 2019-08-30T15:53:07 | 2019-08-30T15:53:07 | {} | MEMBER |
ClickHouse | ClickHouse | 526,920,000 | 6,765 | PlatonB | > Сейчас (починено #6249 ):
>
> ```
> SELECT *
> FROM tbl_1
> LEFT JOIN tbl_2 ON tbl_1.int_col = tbl_2.int_col
>
> ┌─int_col─┬─tbl_2.int_col─┐
> │ 1 │ ᴺᵁᴸᴸ │
> │ 2 │ ᴺᵁᴸᴸ │
> │ 3 │ 3 │
> │ 4 │ 4 │
> │ 5 │ 5 │
> └─────────... | 2019-09-01T13:38:48 | 2019-09-01T13:43:08 | {} | NONE |
ClickHouse | ClickHouse | 526,946,408 | 2,335 | alexey-milovidov | @vitlibar Please confirm that this issue is outdated and fixed. | 2019-09-01T19:27:44 | 2019-09-01T19:27:44 | {} | MEMBER |
ClickHouse | ClickHouse | 527,122,554 | 6,746 | alesapin | I cannot reproduce it locally :( Maybe my table scheme is incorrect? You send only partitions without `.sql` file, but they contain other columns. Also, maybe I'm doing another alter. Can I paste here deduced table scheme and `ALTER` query with real column names, so you can validate them? | 2019-09-02T12:03:08 | 2019-09-02T12:03:08 | {} | MEMBER |
ClickHouse | ClickHouse | 527,292,580 | 6,773 | yuzhichang | @alexey-milovidov I don't understand how `remote('127.0.0.{2,3}', ... to emulate distributed queries` work. The log indicates that the standalone node receives the query which aim to 127.0.0.2, without listening on 127.0.0.2. Could you provide document or code link? | 2019-09-03T03:45:10 | 2019-09-03T05:13:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,326,925 | 6,335 | astudnev | yes | 2019-09-03T06:44:42 | 2019-09-03T06:44:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,371,047 | 6,707 | filimonov | To consider merging such a feature in current form - we at least need to check / to test / to ensure that it works with other logic which already exists.
At least those scenarios should be tested:
1) multiquery + send_progress_in_http_headers
2) multiquery + query which will start streaming (i.e. result bigger tha... | 2019-09-03T09:01:09 | 2019-09-03T09:37:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,638,087 | 6,750 | thyn | Fixed by specifying database. Wrong default database. | 2019-08-30T15:09:18 | 2019-08-30T15:09:18 | {} | NONE |
ClickHouse | ClickHouse | 526,663,150 | 6,747 | den-crane | I think toDate function is not monotonic (because it's argument can be string) so probably partition elimination does not work.
CH does not store a result of `toDate(test_date)` in case of `PARTITION BY toDate(test_date)`
CH always stores MAX_MIN of column (test_date) as part's attribute (as DateTime as you created... | 2019-08-30T16:22:10 | 2019-08-30T16:52:40 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,846,399 | 6,758 | Tasselmi | Thanks for your reply.
I want to store 90 days of logs raw data( 2TB per day ) .
I am planing to buy 20 machines ( each with 32cores and 192GB RAM and 4TB ssd ) .
The cluster will be 10 shards and each 2 replicas.
We expect to run query or function such as `left join` (at most join 2 times in one query on 2 key... | 2019-08-31T16:49:02 | 2019-08-31T16:49:02 | {} | NONE |
ClickHouse | ClickHouse | 526,965,268 | 6,764 | alexey-milovidov | Another extremely rare deadlock found:
```
Run clickhouse-server with ASan.
./stress 16 3000
watch -n0.1 -x clickhouse-client --query "SELECT elapsed, query_id, os_thread_ids, ProfileEvents.Values[indexOf(ProfileEvents.Names, 'UserTimeMicroseconds')] / 1000 AS user_ms, arraySum(arrayFilter(x, y -> y LIKE 'RWL... | 2019-09-02T00:13:24 | 2019-09-02T00:13:24 | {} | MEMBER |
ClickHouse | ClickHouse | 527,123,202 | 6,746 | simPod | @alesapin
```sql
CREATE TABLE default.counters_bytes (`fetched_at` DateTime, `device_id` UUID, `port_id` UUID, `contract_id` UUID, `customer_id` UUID, `server_id` UUID, `counter_in` UInt64, `counter_out` UInt64, `bps_in` Nullable(UInt64), `bps_out` Nullable(UInt64), `server_port_link_type` UInt8, `peering_cz_traff... | 2019-09-02T12:05:28 | 2019-09-02T12:05:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,197,366 | 6,667 | akuzm | ```
select count(*), count(distinct URL) FROM hits_100m_single WHERE CounterID = 34 AND EventDate >= '2013-07-01' AND EventDate <= '2013-07-31' AND NOT Refresh AND IsLink AND NOT IsDownload;
┌─count()─┬─uniqExact(URL)─┐
│ 47740 │ 15628 │
└─────────┴────────────────┘
```
Just in case, checked that t... | 2019-09-02T16:17:39 | 2019-09-02T16:17:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,273,271 | 6,357 | yingfeng | Our proposal has been raised and tracked in issue [6772](https://github.com/yandex/ClickHouse/issues/6772). | 2019-09-03T01:47:09 | 2019-09-03T01:47:09 | {} | NONE |
ClickHouse | ClickHouse | 527,297,567 | 6,168 | YiuRULE | Ok, normally review ready :) | 2019-09-03T04:17:08 | 2019-09-03T04:17:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,647,648 | 6,735 | volfco | ``` <remote_servers>
<slomo_aus>
<shard>
<internal_replication>true</internal_replication>
<replica>
<host>server1.site.net</host> <!--shard 1-->
<port>9000</port>
</replica>... | 2019-08-30T15:36:27 | 2019-08-30T15:36:27 | {} | NONE |
ClickHouse | ClickHouse | 526,818,408 | 6,707 | jcdang | Is there a certain header you're passing to clickhouse? When I test, the body is always flushed at the end.
https://github.com/yandex/ClickHouse/blob/efd9bf2d3fac03df66c8d8cca337665d4024728a/dbms/programs/server/HTTPHandler.cpp#L669 | 2019-08-31T10:07:07 | 2019-08-31T10:07:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,920,177 | 6,758 | den-crane | > I want to store 90 days of logs raw data( 2TB per day ) .
> I am planing to buy 20 machines ( each with 32cores and 192GB RAM and 4TB ssd ) .
90*2/10 = 18TB data per node (`4TB ssd`), you need to evaluate compression rate.
>We expect to run query or function such as left join (at most join 2 times in one query... | 2019-09-01T13:41:25 | 2019-09-01T13:41:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,923,473 | 6,765 | alexey-milovidov | Да, это ещё не вошло в релиз. Есть в 19.14 testing. | 2019-09-01T14:25:00 | 2019-09-01T14:25:00 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 527,122,889 | 2,335 | vitlibar | > @vitlibar Please confirm that this issue is outdated and fixed.
Yes, the dictionaries don't behave the way they used to. If some dictionary couldn't be loaded the server starts anyway. If you want to check if all the dictionaries are actually loaded just execute
```
SELECT name, status from system.dictionaries
... | 2019-09-02T12:04:19 | 2019-09-02T12:11:02 | {} | MEMBER |
ClickHouse | ClickHouse | 527,392,007 | 6,783 | blinkov | @tavplubix значит надо чинить в китайской версии.
@BayoNet можешь записать в список это всё поправить, чтоб не потерялось? | 2019-09-03T10:00:54 | 2019-09-03T10:00:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,632,315 | 6,746 | alesapin | If it's synthetic than generating query/script will be OK. | 2019-08-30T14:53:12 | 2019-08-30T14:53:12 | {} | MEMBER |
ClickHouse | ClickHouse | 526,804,748 | 6,707 | filimonov | The question is: how will you understand that some step failed if you ignore http output? | 2019-08-31T06:19:41 | 2019-08-31T06:20:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,827,725 | 5,417 | amosbird | > I'm not sure what the results mean
It might related to the load factor being too small which triggers edge resize. https://github.com/yandex/ClickHouse/blob/master/dbms/src/Common/HashTable/HashTable.h#L159
tweaking it to 80%-95% probably gives more reliable results. | 2019-08-31T12:44:05 | 2019-08-31T12:44:05 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 527,053,836 | 6,746 | alesapin | @simPod any updates? | 2019-09-02T08:26:52 | 2019-09-02T08:26:52 | {} | MEMBER |
ClickHouse | ClickHouse | 527,120,794 | 6,746 | simPod | It's constant, just reproduced on my instance. @alesapin | 2019-09-02T11:56:50 | 2019-09-02T11:56:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,680,930 | 6,746 | alexey-milovidov | It's better to put table data from filesystem into `.tar` and send it in any preferred form to `clickhouse-feedback@yandex-team.com` (it is private e-mail address) or upload to Dropbox or Google Drive and send a link. | 2019-08-30T17:19:26 | 2019-08-30T17:19:26 | {} | MEMBER |
ClickHouse | ClickHouse | 527,131,746 | 6,746 | simPod | @alesapin avg cpu load is around 20 %
some mutations are happening from time to but all are is_done=1
Also there's nothing stuck in merges table, all is merged within few seconds max.
Not sure what might be the cause, maybe there are some partitions that cause problems but it's not the one I sent to you? | 2019-09-02T12:34:34 | 2019-09-02T12:34:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,594,915 | 6,733 | alexey-milovidov | > more than one table
But it is not stated in setting name or setting description. | 2019-08-30T13:07:12 | 2019-08-30T13:07:12 | {} | MEMBER |
ClickHouse | ClickHouse | 526,608,869 | 6,741 | alexey-milovidov | > A possible thing to do is to add -OrNull and -OrZero as aggregate function combinators which fill the default value.
That's what I have in my long list of appreciated features. | 2019-08-30T13:49:54 | 2019-08-30T13:49:54 | {} | MEMBER |
ClickHouse | ClickHouse | 526,805,721 | 6,707 | jcdang | The HTTP protocol will send a non 200 code. The status code could vary but it's usually a 404 or a 500. For example, if you want curl to report a failure on the exit code you can add `-f`. | 2019-08-31T06:37:03 | 2019-08-31T06:37:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,894,411 | 6,707 | jcdang | So back to your question but modified slightly... What happens half way through the progress something fails and we can't modify the status which was already sent? Are you telling me I should grep for an error in the response body? | 2019-09-01T07:25:40 | 2019-09-01T07:25:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,956,397 | 6,656 | alexey-milovidov | Also please look at the PVS-Studio report: https://clickhouse-test-reports.s3.yandex.net/0/4155771106f9313610875e84cfdc2233cef10115/pvs_studio_report/pvs-studio-html-report/sources/LiveViewEventsBlockInputStream.h_31.html#ln217 | 2019-09-01T21:57:17 | 2019-09-01T21:57:17 | {} | MEMBER |
ClickHouse | ClickHouse | 527,135,630 | 6,746 | alesapin | Yes, maybe you sent other partitions. Also, maybe not all columns exist since table creation, especially `server_port_link_type`? Maybe some of them are added with `ALTER TABLE ADD COLUMN` query? If so try to find partition with the smallest number of columns.
Can you try to reproduce it on your machine with sent pa... | 2019-09-02T12:46:55 | 2019-09-02T12:46:55 | {} | MEMBER |
ClickHouse | ClickHouse | 527,388,128 | 6,783 | tavplubix | > не-не, так нельзя, надо разбираться что с относительной ссылкой не так
В китайской версии нет `settings.md`, поэтому там используется английский `settings.md`, который ссылается (относительно) на template в английском `formats.md`. На китайском языке есть свой `formats.md`, в котором нет template. Поэтому относите... | 2019-09-03T09:49:46 | 2019-09-03T09:49:46 | {} | MEMBER |
ClickHouse | ClickHouse | 526,608,516 | 6,745 | filimonov | CHeck `format_csv_delimiter` | 2019-08-30T13:48:59 | 2019-08-30T13:48:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,632,122 | 6,746 | alesapin | Your table data which allows to reproduce the issue. | 2019-08-30T14:52:37 | 2019-08-30T14:52:37 | {} | MEMBER |
ClickHouse | ClickHouse | 526,817,064 | 6,707 | filimonov | And now you will have a lot of queries. First 100 was successful, and output was started, "200 ok" header was already sent and some responce body too. And query number 101 fails (and we can't modify the status which was already sent)? How can we let you know about the problem? | 2019-08-31T09:45:22 | 2019-08-31T09:45:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 526,842,593 | 6,761 | alexey-milovidov | CC @yuzhichang | 2019-08-31T16:09:10 | 2019-08-31T16:09:10 | {} | MEMBER |
ClickHouse | ClickHouse | 526,935,068 | 6,768 | alexey-milovidov | It also does not work for formats. | 2019-09-01T16:43:02 | 2019-09-01T16:43:02 | {} | MEMBER |
ClickHouse | ClickHouse | 527,075,195 | 6,772 | filimonov | It's a bit hard to follow the description, because you redefine some commonly used (also in ClickHouse) terms (shard / replcation).
As i understand Converthouse works on the top of ClickHouse and it manages redistibution / resharding of the data dynamically, by rewriting 'external' queries to 'internal' queries acc... | 2019-09-02T09:27:19 | 2019-09-02T09:29:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,081,447 | 3,750 | filimonov | Is it still actual? Usually merges shouldn't take so much RAM, and executed in streaming manner.
Can you show the create table statement? | 2019-09-02T09:44:49 | 2019-09-02T09:44:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,122,560 | 6,767 | 4ertus2 | It's something between multiple JOINS, PredicateExpressionsOptimizer and selecting from VIEW. As workaround it's possible to use JOINs with ```enable_optimize_predicate_expression = 0```. There's another issue with troubles between multiple JOINs and predicate push down #4731 | 2019-09-02T12:03:09 | 2019-09-02T12:03:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 527,124,475 | 6,746 | alesapin | Still cannot reproduce:
```
click.sas.yp-c.yandex.net :) select count(*) from counters_bytes;
SELECT count(*)
FROM counters_bytes
┌──cou... | 2019-09-02T12:10:06 | 2019-09-02T12:10:06 | {} | MEMBER |
ClickHouse | ClickHouse | 548,507,846 | 7,207 | jameshartig | We are having a similar issue running 19.11.7.40. RSS continues to grow until it's killed by the kernel for using more than the cgroup's memory limit. Here you can see the last 7 days of usage:
Our servers h... | 2019-10-31T18:24:53 | 2019-10-31T18:24:53 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.