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 | 341,268,389 | 1,430 | alexey-milovidov | Ok. | 2017-11-01T23:00:22 | 2017-11-01T23:00:22 | {} | MEMBER |
ClickHouse | ClickHouse | 341,949,945 | 1,433 | alexey-milovidov | https://github.com/yandex/ClickHouse/commit/8e4253ac73b12fd6a8c16a0c1ef78fdd46b8ea01 | 2017-11-05T05:19:04 | 2017-11-05T05:19:04 | {} | MEMBER |
ClickHouse | ClickHouse | 342,908,097 | 1,467 | valyala | @inkrement , try minimizing and then mangling the dataset, so the error is still reproducible. Then post reproducible case here. This should increase chances for fixing the issue. | 2017-11-08T18:19:33 | 2017-11-08T18:19:33 | {} | NONE |
ClickHouse | ClickHouse | 343,312,760 | 1,360 | alexey-milovidov | Looks like stack traces in FreeBSD build are cut in logs.
To investigate this issue, we need full stack trace.
How to get it? There are two variants:
1. Look at core dump with gdb, type `t apply all bt full`.
2. Run clickhouse-server under gdb, like this:
```
sudo service clickhouse-server stop
sudo -u cli... | 2017-11-09T22:28:40 | 2017-11-09T22:29:13 | {} | MEMBER |
ClickHouse | ClickHouse | 341,432,859 | 1,434 | ztlpn | Yes, it is a limitation of CASE expression WHEN operator (which is implemented via the transform() function). Expressions in THEN clauses must be constants. | 2017-11-02T14:10:21 | 2017-11-02T14:10:21 | {
"-1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,669,457 | 1,435 | ztlpn | @wusihang A cause of "resource temporarily unavailable" errors is that some system limit is exceeded. As you have created a lot of tables, the limit in question is likely either the limit on max number of open files or the limit on max number of threads (Replicated tables are quite heavy in this regard, each creating ... | 2017-11-03T10:39:49 | 2017-11-03T10:39:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,993,518 | 1,436 | ztlpn | Thanks for a simple and reproducible test case @pdrtrifork ! I've been able to pinpoint the cause of the issue - when materialized view is populated from a MergeTree table, the optimizer moves WHERE condition to PREWHERE and thus garbles the original query.
While the fix is waiting to be released, here is a simple w... | 2017-11-05T18:19:23 | 2017-11-05T18:19:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 342,352,338 | 1,402 | vavrusa | I wonder where should this configuration go, using SETTINGS is not going to be pretty (especially when you create multiple tables all using the same broker), so either:
* an additional ENGINE parameter with a tuple of SSL settings
* an additional ENGINE parameter with a string of librdkafka (or JSON) settings
* an a... | 2017-11-07T02:07:04 | 2017-11-07T02:07:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 342,825,662 | 1,453 | StarWix | It would be useful to have a setting to autoconvert -0 to 0.
Otherwise you have to wrap expression with multiply or division operators to `if` function, `-0` - `0` converting. It looks ugly:
```
SELECT
if(x = 0, 0, x) AS y,
count()
FROM
(
SELECT 0 * -1. AS x
UNION ALL
SELECT 0.
... | 2017-11-08T14:00:48 | 2017-11-08T14:00:48 | {} | NONE |
ClickHouse | ClickHouse | 342,966,368 | 1,463 | suspend0 | note that `quantilesTDigestMerge` has great performance and can be a substitute for our use case.
Thank you! | 2017-11-08T21:30:29 | 2017-11-08T21:30:29 | {
"hooray": 2
} | NONE |
ClickHouse | ClickHouse | 343,327,573 | 1,471 | bobrik | I removed problematic queue item and restarted clickhouse, no more errors in logs. | 2017-11-09T23:37:16 | 2017-11-09T23:37:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,332,504 | 1,471 | alexey-milovidov | This is rare issue, when wrong file size or checksum is written in ZooKeeper.
We have the following scenario:
- one replica is doing merge and write merged part to disk;
- simultaneously, checksum and file sizes are calculated and written to ZooKeeper;
- somehow, the data written to disk actually have different c... | 2017-11-10T00:04:35 | 2017-11-10T00:04:35 | {} | MEMBER |
ClickHouse | ClickHouse | 343,706,186 | 1,472 | alexey-milovidov | Perfect, thanks! | 2017-11-12T01:07:17 | 2017-11-12T01:07:17 | {} | MEMBER |
ClickHouse | ClickHouse | 341,040,084 | 1,281 | silviucpp | simple way to replicate is to run `SELECT floor(NULL);` using tabix.io | 2017-11-01T08:59:54 | 2017-11-01T08:59:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,695,647 | 1,435 | ztlpn | @silviucpp It applies to other tables too (replicated tables are just the heaviest ones). Most table engines will store their data in quite a lot of files (e.g. MergeTree tables store each sorted part in a separate directory that contains 2 files per column + some more). This puts a considerable strain on the file syst... | 2017-11-03T12:55:43 | 2017-11-03T12:56:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 342,981,914 | 1,468 | alexey-milovidov | @vavrusa Memory allocation is tracked directly in `HyperLogLogWithSmallSetOptimization`.
It doesn't look related to this issue. | 2017-11-08T22:26:33 | 2017-11-08T22:26:33 | {} | MEMBER |
ClickHouse | ClickHouse | 343,625,341 | 1,478 | silviucpp | @alexey-milovidov from what I see this threads wake up time to time at regular intervals even if there is no activity in that tables. Most of them are using tryWait(duration); which means that the threads is in idle until an event comes or the time expired. When you have such a lot of threads I think the context switch... | 2017-11-11T00:52:48 | 2017-11-11T00:52:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,855,676 | 1,486 | shernupup | I set the <max_memory_usage>80000000000</max_memory_usage> in 'users.xml' | 2017-11-13T09:07:14 | 2017-11-13T09:07:14 | {} | NONE |
ClickHouse | ClickHouse | 340,812,689 | 208 | Dimon70007 | @ztlpn Спасибо за помощь. Добавил `<send_progress_in_http_headers>true</send_progress_in_http_headers>`
в config.xml - присылает обычные заголовки до получения ответа.
А настройка `<output_format_json_quote_64bit_integers>false</output_format_json_quote_64bit_integers>`
может как то влиять на отправку heade... | 2017-10-31T16:07:20 | 2017-10-31T20:02:40 | {} | NONE |
ClickHouse | ClickHouse | 341,005,310 | 827 | jaykelin | Which nodes in Zookeeper could be clear is difficult to select.
is there prossible to delete nodes in Zookeeper during droping partition? | 2017-11-01T06:53:34 | 2017-11-01T06:53:34 | {} | NONE |
ClickHouse | ClickHouse | 341,056,144 | 827 | ztlpn | Yes, I plan to fix it soon. | 2017-11-01T09:49:00 | 2017-11-01T09:49:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,343,422 | 1,434 | filimonov | For the record: syntax without CASE variable works:
```
:) SELECT
:-] number,
:-] CASE
:-] WHEN number=0 THEN number+1
:-] WHEN number=1 THEN number*10
:-] ELSE number
:-] END
:-] FROM system.numbers
:-] WHERE number < 10
:-] LIMIT 10;
SELECT
number,
caseWithoutExpression(numbe... | 2017-11-02T07:55:17 | 2017-11-02T07:55:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,417,623 | 1,404 | silviucpp | Great let me know if you need other tests !
Silviu | 2017-11-02T13:17:27 | 2017-11-02T13:17:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,777,702 | 1,443 | ztlpn | Ok, so you are not using Replicated tables (not a recommended way but understandable if you don't want the additional operational burden of a ZooKeeper cluster).
Do you insert data via the Distributed table? If yes, then you should check if there is a lot of files in the `/var/lib/clickhouse/data/shark/request/<some... | 2017-11-03T17:46:16 | 2017-11-03T17:46:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,518,297 | 827 | Foxsa | Any updates? | 2017-11-10T16:22:37 | 2017-11-10T16:22:37 | {} | NONE |
ClickHouse | ClickHouse | 343,618,887 | 1,480 | alexey-milovidov | We can provide `clickhouse format` mode of operation for clickhouse binary. | 2017-11-11T00:03:46 | 2017-11-11T00:05:26 | {} | MEMBER |
ClickHouse | ClickHouse | 340,888,965 | 1,404 | silviucpp | Hello @ludv1x ,
Have you tested using internal_replication = true ?
Silviu | 2017-10-31T19:59:40 | 2017-10-31T19:59:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,063,890 | 208 | ztlpn | С system.numbers прогресс не работает (попробуйте задать запрос в clickhouse-client, прогрессбара тоже не будет, только скорость выполнения). Должно работать с нормальной таблицей, для которой можно оценить число строк, которые нужно прочитать и обработать (в вашем запросе можно было бы оценить из значения LIMIT, но эт... | 2017-11-01T10:18:32 | 2017-11-01T10:18:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,675,581 | 1,435 | silviucpp | @ztlpn this problem with "Creating lots of Replicated tables is usually not a good idea" apply only to replicated tables ? or there are other tables that follows the same approach ?
Silviu | 2017-11-03T11:10:30 | 2017-11-03T11:10:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 342,750,015 | 1,461 | filimonov | yap, it looks like after switching to submodules zip releases are not too usable, because they are not self-contained and can't be used autonomously.
To make them usable again probably some custom packing and releasing scenarios should be used. For example something like that: https://github.com/meitar/git-archive-a... | 2017-11-08T08:47:27 | 2017-11-08T09:16:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 342,966,492 | 1,468 | vavrusa | HyperLogLogWithSmallSetOptimization isn't using arena or https://github.com/yandex/ClickHouse/blob/master/dbms/src/Common/Allocator.h, so it's not trivial to limit memory requirement for queries (it's still tracked).
https://github.com/yandex/ClickHouse/blob/b370551e676ddfc8739e4f9190293ffeb14edb1b/dbms/src/Common/H... | 2017-11-08T21:30:54 | 2017-11-08T21:30:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 342,978,665 | 1,466 | alexey-milovidov | Tell me, what patches you have applied to Poco,
and what version of Poco and libressl do you use.
Please, insert debug output in the following places in your code:
```
poco/Crypto/src/OpenSSLInitializer.cpp, OPENSSL_config(NULL);
libressl/crypto/conf/conf_sap.c, OPENSSL_config(const char *config_name)
libress... | 2017-11-08T22:13:55 | 2017-11-08T22:13:55 | {} | MEMBER |
ClickHouse | ClickHouse | 343,119,842 | 1,457 | sunsingerus | Helped - but partially - for Fedora26 only. As far as I see, ClickHouse requires openssl-1.1.0f and Fedora26 is the only who has it bundled (among Fedora & CentOS distros)
As to other distros (Fedora25, CentOS), still unable to build ClickHouse, even after openssl-1.1.0f installed. The problem may be because openssl i... | 2017-11-09T10:56:30 | 2017-11-09T10:56:30 | {} | NONE |
ClickHouse | ClickHouse | 343,367,741 | 1,368 | jaykelin | @ludv1x The case described above does not print error logs in clickhouse-server.log
But in the daily operation and maintenance, print the following ERROR log,then clickhouse-server crash
> <Error> BaseDaemon: ########################################
> <Error> BaseDaemon: (from thread 1974) Received signal Segmentati... | 2017-11-10T03:52:28 | 2017-11-10T03:52:28 | {} | NONE |
ClickHouse | ClickHouse | 343,651,949 | 1,467 | inkrement | The data is somehow public, it's from a Kaggle competition. But I am not sure if I am allowed to upload it here. Anyway, you can download the files on kaggle: https://www.kaggle.com/c/kkbox-churn-prediction-challenge/data. The first data set was user_logs.csv.7z and then I tried to add user_logs_v2.csv.7z to the same t... | 2017-11-11T09:23:48 | 2017-11-11T09:23:48 | {} | NONE |
ClickHouse | ClickHouse | 340,888,639 | 1,298 | silviucpp | @alexey-milovidov having support for offsets is not so important as @filimonov pointed there are simple alternatives for this like:
```sql
select count(*) as count, toRelativeDayNum(date+3*3600) - toRelativeDayNum(X+3*3600) from table
when date >= X and date =< Y
```
Probably we can close this bug
| 2017-10-31T19:58:30 | 2017-10-31T19:58:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,750,568 | 1,378 | proller | looks like its your /usr/bin/cmake problem, try to reinstall or update. | 2017-11-03T16:12:04 | 2017-11-03T16:12:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,317,723 | 1,360 | slw | # gdb /usr/local/bin/clickhouse 72296
GNU gdb 6.1.1 [FreeBSD]
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is... | 2017-11-09T22:49:39 | 2017-11-09T22:49:39 | {} | NONE |
ClickHouse | ClickHouse | 343,705,328 | 1,476 | alexey-milovidov | https://github.com/yandex/ClickHouse/commit/e611e1cc6659526bf519e344f3b81c7066779ca9 | 2017-11-12T00:44:16 | 2017-11-12T00:44:16 | {} | MEMBER |
ClickHouse | ClickHouse | 343,851,887 | 1,486 | jackpgao | How much memory do you allocated to CK? | 2017-11-13T08:52:06 | 2017-11-13T08:52:06 | {} | NONE |
ClickHouse | ClickHouse | 340,889,986 | 1,028 | silviucpp | @alexey-milovidov any news or ETA about this bug ? Seems pretty important feature | 2017-10-31T20:03:29 | 2017-10-31T20:03:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 340,917,986 | 208 | Dimon70007 | ``` $ echo "
> SELECT
> id,
> count() as count
> FROM
> (
> SELECT
> number AS id,
> toDateTime('2017-10-30 00:02:03') + (id * 1000) AS begin,
> toDateTime('2017-10-30 00:02:03') + (id * 2000) AS end
> FROM system.numbers
> ... | 2017-10-31T21:49:07 | 2017-11-01T08:22:49 | {} | NONE |
ClickHouse | ClickHouse | 341,928,750 | 1,453 | alexey-milovidov | This is documented behaviour.
We can consider implementing some functions or settings to make arithmetic operations filter out special values like infinities, nans and -0. (This will have some runtime cost.) | 2017-11-04T20:48:19 | 2017-11-04T20:48:19 | {} | MEMBER |
ClickHouse | ClickHouse | 342,034,647 | 1,443 | qietingfengling | hi,if I set the internal_relication ture,clickhouse can use replicated tables?
yes ,I insert data via the Distributed table.
there have lot of files in the /var/lib/clickhouse/data/shark/request/<something>/ directory
the error log is as follow,if there any way to fix it
![image](https://user-images.githubuse... | 2017-11-06T03:03:41 | 2017-11-06T03:03:41 | {} | NONE |
ClickHouse | ClickHouse | 343,128,429 | 1,443 | ztlpn | Can you show the stack trace that appears in the error log? The lines after `Stack trace:`. | 2017-11-09T11:34:54 | 2017-11-09T11:37:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,624,093 | 1,478 | alexey-milovidov | > Code is the same so it's not point to ducplicate it.
Great, thanks!
> Btw not sure why Stat * stat_ has _ at the end .. ?
I think it is copy paste from functions, that have another `stat` variable inside.
For more clarity, you can rename `stat_` to `out_stat` and all other out-parameters in the same way.
... | 2017-11-11T00:42:24 | 2017-11-11T00:42:24 | {} | MEMBER |
ClickHouse | ClickHouse | 341,422,591 | 1,404 | ludv1x | More generally, it is reproduced if all replicas of the shard are local replicas. | 2017-11-02T13:36:25 | 2017-11-02T13:36:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,422,729 | 1,404 | ludv1x | Thank you! We will fix it in the next release. | 2017-11-02T13:36:54 | 2017-11-02T13:36:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,671,338 | 1,443 | ztlpn | Hi @qietingfengling !
The likely cause if the problem is that replicas of some shard do not contain the same data. So when the Distributed table chooses one or the other replica for the execution of the query the results are different. To find which shard is the problematic one, execute the query manually on every r... | 2017-11-03T10:48:40 | 2017-11-03T10:49:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,831,643 | 1,342 | zhkvia | I think I'll work on this as my thesis. Can't set the exact deadline, though | 2017-11-03T21:27:33 | 2017-11-03T21:27:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,950,143 | 1,420 | alexey-milovidov | Filimonov, thank you!
Slach:
If you have exported this data from Yandex.Metrika Logs API, please file an issue to them. | 2017-11-05T05:25:36 | 2017-11-05T05:25:36 | {} | MEMBER |
ClickHouse | ClickHouse | 343,527,060 | 1,458 | bobrik | We downgraded from 3.5.3-beta to 3.4.10 and it looks a lot better. | 2017-11-10T16:54:22 | 2017-11-10T16:54:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,626,853 | 1,480 | alexey-milovidov | https://github.com/yandex/ClickHouse/commit/726301572f2e0de5335cdd2c071e0a912c282ec4
```
$ ./clickhouse-format --help
Usage: ./clickhouse-format [options] < query
Allowed options:
-h [ --help ] produce help message
--hilite add syntax highlight with ANSI terminal escape
... | 2017-11-11T01:06:54 | 2017-11-11T01:07:34 | {
"heart": 2
} | MEMBER |
ClickHouse | ClickHouse | 343,643,488 | 1,458 | velom | > It means - 3.5.1 with patches.
I will try to figure out, what patches do we use.
There are no patches. We have built 3.5.1-alfa from source, because it was last version at the moment. We need a [Dynamic Reconfiguration](http://zookeeper.apache.org/doc/r3.5.1-alpha/zookeeperReconfig.html), because on our huge zk c... | 2017-11-11T06:03:32 | 2017-11-11T06:03:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,707,930 | 1,435 | silviucpp | Number of files opened limit I'm not sure will bring too much performance penalty. You can increase the limit. But for number of threads even if you raise that limit the system will slow down a lot as also you mentioned early.
I saw some cloud solutions for clickhouse and I was curious how they can work as time cli... | 2017-11-03T13:48:14 | 2017-11-03T13:48:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,711,793 | 1,444 | ztlpn | Confirmed, thanks for a simple and reproducible test case! | 2017-11-03T14:03:06 | 2017-11-03T14:03:06 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,822,421 | 1,449 | alexey-milovidov | What exactly does it mean? If it will include with `-isystem` compiler argument, then it will suppress all warnings. But we don't want to blindly suppress all warnings. | 2017-11-03T20:45:52 | 2017-11-03T20:45:52 | {} | MEMBER |
ClickHouse | ClickHouse | 341,993,683 | 1,454 | ztlpn | Yes, because they weren't created with a CREATE TABLE query :)
You can get information about the structure of the system tables with a DESCRIBE TABLE query. | 2017-11-05T18:21:56 | 2017-11-05T18:21:56 | {
"confused": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 342,450,013 | 1,457 | sunsingerus | Just tested this issue on Fedora 26, CentOS 6.9 & 7.4 - the same result.
So, it does not depend on OS - whether it is Fedora or CentOS
| 2017-11-07T11:08:06 | 2017-11-07T11:29:37 | {} | NONE |
ClickHouse | ClickHouse | 342,909,675 | 1,403 | valyala | @filimonov , try using [chproxy](https://github.com/Vertamedia/chproxy). It must kill the query immediately after the client closes http connection. | 2017-11-08T18:23:01 | 2017-11-08T18:23:01 | {} | NONE |
ClickHouse | ClickHouse | 342,912,985 | 1,394 | valyala | FYI, we successfully used `ClickHouse` on top of [google cloud storage](https://cloud.google.com/storage/) via [gcsfuse](https://github.com/GoogleCloudPlatform/gcsfuse). | 2017-11-08T18:29:32 | 2017-11-08T18:29:32 | {} | NONE |
ClickHouse | ClickHouse | 342,979,108 | 1,466 | alexey-milovidov | You can also give me the output of `bt full` in gdb (ensure that you build all code, including libressl with -g option). | 2017-11-08T22:15:41 | 2017-11-08T22:15:41 | {} | MEMBER |
ClickHouse | ClickHouse | 343,632,390 | 1,480 | bobrik | Awesome, thanks! | 2017-11-11T02:06:29 | 2017-11-11T02:06:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,738,270 | 1,067 | filimonov | There is also Buffer table engine, which can be used for solving that
problem (with certain limitations, check documentation:
https://clickhouse.yandex/docs/en/table_engines/buffer.html ).
Also there is an experimental Kafka table engine (can be used as a source for materialized view), and Kafka->Clickhouse bridge... | 2017-11-12T13:43:58 | 2017-11-12T13:48:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,574,745 | 1,442 | sunsingerus | By the way, just saw the following line in the above log:
```
Using internal compiler: headers=/usr/share/clickhouse/headers : /usr/share/clickhouse/bin/clang
```
Have no `/usr/share/clickhouse` at all, may be this is related to the issue
| 2017-11-02T22:17:22 | 2017-11-02T22:17:22 | {} | NONE |
ClickHouse | ClickHouse | 341,648,344 | 1,435 | wusihang | Because of network traffic restriction , clickhouse server's ip was banned by ISP. | 2017-11-03T08:52:38 | 2017-11-03T08:52:38 | {} | NONE |
ClickHouse | ClickHouse | 342,608,095 | 1,458 | vavrusa | I've seen @proller adding vendored ZK 3.4.10 recently, so I guess that should work? | 2017-11-07T20:14:46 | 2017-11-07T20:14:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 342,769,251 | 1,461 | sunsingerus | Ok, thanks for clarification on release file.
As to attempts to build directly from git sources, I've tried, and, unfortunately, also failed due to cmake
error. As you can see, I need ClickHouse on RedHat-based platform (Fedora & CentOS) and building v1.1.54310-stable from sources stuck into [this issue](https://... | 2017-11-08T10:03:25 | 2017-11-08T10:03:25 | {} | NONE |
ClickHouse | ClickHouse | 342,885,983 | 687 | maxistua | Any updates on this? | 2017-11-08T17:08:43 | 2017-11-08T17:08:43 | {} | NONE |
ClickHouse | ClickHouse | 343,692,243 | 1,067 | zero-master | @shuttie , Are you aware of any such existing batch writer for clickhouse? | 2017-11-11T20:32:25 | 2017-11-11T20:32:25 | {} | NONE |
ClickHouse | ClickHouse | 343,771,413 | 1,118 | telemmaite | Any new information about this, please? | 2017-11-12T21:57:42 | 2017-11-12T21:57:42 | {
"+1": 3
} | NONE |
ClickHouse | ClickHouse | 341,412,021 | 1,404 | ludv1x | Yes, it is reproduced if you use `127.0.0.1` and `<internal_replication>true</internal_replication>`! | 2017-11-02T12:54:54 | 2017-11-02T12:54:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,624,571 | 1,435 | wusihang | Does clickhouse has any restriction on number of tables with ReplicatedMerge engine?
I created approximately 300 ReplicatedMerge tables , then I couldn't create more tables with ReplicatedMerge because of "Resource temporarily unavailable" error. A lot of deamon processes was found on my clickhouse server. | 2017-11-03T05:20:29 | 2017-11-03T05:21:46 | {} | NONE |
ClickHouse | ClickHouse | 341,717,559 | 1,442 | proller | Please show full cmake output.
seems your error hides in ... cut ... | 2017-11-03T14:23:43 | 2017-11-03T14:23:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,864,270 | 1,436 | pdrtrifork | A bit more information: If I use the Memory table engine for the events table, the materialized view behaves as expected. But that is not an option in our case :-) | 2017-11-04T02:03:07 | 2017-11-04T02:03:07 | {} | NONE |
ClickHouse | ClickHouse | 341,911,776 | 1,385 | alexey-milovidov | https://github.com/yandex/ClickHouse/pull/1451 | 2017-11-04T16:48:39 | 2017-11-04T16:48:39 | {} | MEMBER |
ClickHouse | ClickHouse | 343,526,221 | 1,458 | rpfennin | We've been running on Zookeeper version: 3.4.8 for some time with no issue. | 2017-11-10T16:51:17 | 2017-11-10T16:51:17 | {} | NONE |
ClickHouse | ClickHouse | 343,622,265 | 1,458 | alexey-milovidov | We use
```
Zookeeper version: 3.5.1-metrika--1
```
It means - 3.5.1 with patches.
I will try to figure out, what patches do we use. | 2017-11-11T00:28:39 | 2017-11-11T00:28:39 | {} | MEMBER |
ClickHouse | ClickHouse | 343,650,733 | 1,478 | silviucpp | @alexey-milovidov I'll first move the ReplicatedMergeTreeAlterThread and ReplicatedMergeTreeCleanupThread which are the simplest tasks. I'll test and if all is ok I'll push on a separate branch and we can discuss on this.
Only place where a task can spend unpredictable time from what I see is zookeeper. Is the bug y... | 2017-11-11T08:59:15 | 2017-11-11T08:59:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,733,404 | 1,067 | shuttie | @zero-master https://github.com/nikepan/clickhouse-bulk , but I've never tried this thing in action. | 2017-11-12T12:22:51 | 2017-11-12T12:22:51 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 341,704,279 | 1,443 | qietingfengling | Hi @ztlpn
All of the system.replication_queue is the same result as folllow:
my definitions of the table as follow:
![2](https://user-images.githubusercontent.com/9961712/32376046-9a9ed19e-c0de-11e7-8f92... | 2017-11-03T13:33:34 | 2017-11-03T13:33:34 | {} | NONE |
ClickHouse | ClickHouse | 342,013,332 | 1,434 | filimonov | Not important as there is a workaround, but also looks like it shouldn't be too hard to make them work for non constant case the same way as caseWithoutExpression work, instead of declaring another out-of-standard behavior. | 2017-11-05T23:02:27 | 2017-11-05T23:02:27 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,622,314 | 1,458 | alexey-milovidov | > I've seen @proller adding vendored ZK 3.4.10 recently, so I guess that should work?
This is relevant only for client library. | 2017-11-11T00:28:59 | 2017-11-11T00:28:59 | {} | MEMBER |
ClickHouse | ClickHouse | 342,074,349 | 1,442 | sunsingerus | Have the following news up to the moment:
release archive (say v1.1.54310-stable.tar.gz) does not contain full sources any more (requires to clone submodules ) and also does not contain git info, thus running
```
git submodule update --init --recursive
```
fails. Am I doing something wrong? Is there any way to bui... | 2017-11-06T08:10:20 | 2017-11-06T08:10:20 | {} | NONE |
ClickHouse | ClickHouse | 342,454,662 | 1,461 | sunsingerus | Tested this on Fedora26, CentOS 6.9 CentsOS 7.4, result is all the same | 2017-11-07T11:28:39 | 2017-11-07T11:28:39 | {} | NONE |
ClickHouse | ClickHouse | 343,353,178 | 1,471 | bobrik | Thanks, that makes sense. I'll update the ticket if we ever see this again. | 2017-11-10T02:05:55 | 2017-11-10T02:05:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,628,017 | 1,478 | alexey-milovidov | > If my assumption is ok if zookeper or disk io it's exhausted dedicated threads per table wont help.
That's correct. What I'm afraid is if all threads in threadpool will lock or wait, but not yield to another tasks. The refactoring should be implemented very carefully.
Also I'm afraid of long waiting due to ZooK... | 2017-11-11T01:18:27 | 2017-11-11T01:18:27 | {} | MEMBER |
ClickHouse | ClickHouse | 340,960,945 | 1,298 | alexey-milovidov | Ok. | 2017-11-01T02:31:53 | 2017-11-01T02:31:53 | {} | MEMBER |
ClickHouse | ClickHouse | 341,670,637 | 1,342 | adhamelia | +1 for support | 2017-11-03T10:45:29 | 2017-11-03T10:45:29 | {} | NONE |
ClickHouse | ClickHouse | 341,818,762 | 1,436 | pdrtrifork | I made a simplified test case of the above without using replicated or distributed tables with the exact same result:
CREATE DATABASE testmv3;
USE testmv3;
CREATE TABLE testmv3.events (
EventDate Date,
SessionUuid FixedString(16),
EventType UInt8,
EventText String
) ENGINE = MergeTree(
Even... | 2017-11-03T20:31:07 | 2017-11-03T20:31:07 | {} | NONE |
ClickHouse | ClickHouse | 342,906,180 | 1,460 | valyala | Just use [symbolic links](https://en.wikipedia.org/wiki/Symbolic_link).
Suppose you have `<path>/opt/clickhouse/</path>`. Then create symbolic links in `/opt/clickhouse/data/` for every db you need:
```
ln -s /mnt/ssd/fastdb /opt/clickhouse/data/fastdb
ln -s /mnt/hdd/slowdb /opt/clickhouse/data/slowdb
```
The a... | 2017-11-08T18:14:20 | 2017-11-08T18:14:20 | {} | NONE |
ClickHouse | ClickHouse | 343,677,689 | 1,482 | ztlpn | It is possible right now! (starting from the 54310 release)
Documentation is under development now, but here are a few examples in tests:
https://github.com/yandex/ClickHouse/blob/master/dbms/tests/queries/0_stateless/00502_custom_partitioning_local.sql
https://github.com/yandex/ClickHouse/blob/master/dbms/tests/q... | 2017-11-11T16:47:22 | 2017-11-11T16:47:22 | {
"hooray": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,336,183 | 1,360 | alexey-milovidov | The relevant part of stack trace looks incorrect.
```
Thread 33 (Thread 31d1600 (LWP 100801/clickhouse)):
#0 0x0000000809ae4a4a in thr_kill () from /lib/libc.so.7
No symbol table info available.
#1 0x0000000809ae4a1b in raise () from /lib/libc.so.7
No symbol table info available.
#2 0x0000000809ae4989 in abort... | 2017-11-10T00:26:48 | 2017-11-10T00:29:52 | {} | MEMBER |
ClickHouse | ClickHouse | 343,622,934 | 827 | alexey-milovidov | We are doing this task right now. | 2017-11-11T00:33:32 | 2017-11-11T00:33:32 | {} | MEMBER |
ClickHouse | ClickHouse | 343,625,901 | 1,476 | alexey-milovidov | This is by design.
For float, the user could expect one of two possibilities:
1. Format with sign and point, example: `-AB.CDE`. That's unusual, but consistent with hex behaviour for unsigned integers.
2. Format native representation of number as bytes in memory in hex. That's easy, but will be inconsistent, b... | 2017-11-11T00:57:42 | 2017-11-11T00:57:42 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 343,737,797 | 1,484 | filimonov | Good news. :) Probably compile instructions should also be refreshed a bit. | 2017-11-12T13:36:36 | 2017-11-12T13:36:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 341,826,380 | 1,436 | alex-zaitsev | We had intermittent issues with MV consistency as well, and it may be perfectly related to not respecting WHERE condition sometimes. | 2017-11-03T21:02:54 | 2017-11-03T21:02:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 342,802,998 | 1,457 | filimonov | Please, try that:
`yum install openssl-static` | 2017-11-08T12:26:50 | 2017-11-08T12:26:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 343,095,130 | 1,443 | qietingfengling | hi @ztlpn if there have a lot of files in the /var/lib/clickhouse/data/shark/request/<something>/ directory,what can I do | 2017-11-09T09:21:25 | 2017-11-09T09:21:25 | {} | NONE |
ClickHouse | ClickHouse | 343,159,496 | 1,360 | slw | Current time 2017-11-09 16:44 MSK
Current requset:
SELECT (intDiv(toUInt32(time), 5) * 5) * 1000 as t, avg(nic_speed) as nic FROM default.net WHERE srv='edge13' AND date >= toDate(1510224251) AND time >= toDateTime(1510224251) GROUP BY t ORDER BY t
(generated by grafana as 'Last 3 hours')
clickhouse-1.1.54304_... | 2017-11-09T13:49:21 | 2017-11-09T13:49:21 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.