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 | 550,244,400 | 7,647 | alexey-milovidov | CC @yuzhichang | 2019-11-06T10:20:06 | 2019-11-06T10:20:06 | {} | MEMBER |
ClickHouse | ClickHouse | 550,277,315 | 751 | ilyabreev | Same issue here. Resolved when `GROUP BY` clause contains same columns as view sorting key. | 2019-11-06T11:55:32 | 2019-11-06T11:55:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,549,172 | 7,646 | jsding | Hi,
Here is the kafka-consumer-group output, I checked and confirmed all kafka clickhouse table use
kafka_num_consumers = 1 settings
```
dingjianshengdeMacBook-Pro:bin ding$ ./kafka-consumer-groups.sh --bootstrap-server mykafkaserver:9092 --describe --group ck_event_prod_v3 --offsets
TOPIC PARTITIO... | 2019-11-06T23:32:12 | 2019-11-06T23:35:44 | {} | NONE |
ClickHouse | ClickHouse | 550,971,319 | 7,669 | alesapin | It will be released today to 19.17.1 prestable release and in the next monday it will be available in 19.17.2 stable. | 2019-11-07T08:14:11 | 2019-11-07T08:14:11 | {} | MEMBER |
ClickHouse | ClickHouse | 549,888,857 | 7,546 | ariesdevil | @alexey-milovidov use `ninja clickhouse` still cause error happen:
```
[3336/3339] Linking CXX static library dbms/src/Functions/libclickhouse_functionsd.a
/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/ranlib: file: dbms/src/Functions/libclickhouse_functionsd.a(addressToLine.... | 2019-11-05T16:05:17 | 2019-11-05T16:05:17 | {} | NONE |
ClickHouse | ClickHouse | 549,953,233 | 6,642 | CurtizJ | Probably it's not about skip indices #7639. | 2019-11-05T18:22:57 | 2019-11-05T18:22:57 | {} | MEMBER |
ClickHouse | ClickHouse | 549,970,223 | 7,608 | alexey-milovidov | I have checked manually the query that has the most slowdown on performance test:
```
SELECT SearchEngineID AS k, uniqCombined(18)(UserID) FROM hits_100m_single GROUP BY k
```
I have downloaded two binaries, one from master and one from this PR:
```
wget https://clickhouse-builds.s3.yandex.net/0/6a871f579fa96... | 2019-11-05T19:02:09 | 2019-11-05T19:13:48 | {} | MEMBER |
ClickHouse | ClickHouse | 550,068,071 | 7,601 | alexey-milovidov | Proper solution is to use shared_ptr or COW for Context. But as Context already contains shared_ptr inside, it will be too clumsy and we need to rewrite it.
> But I guess it will be better to copy context in each function that uses context reference in the execute?
It's also possible, but there are some other iss... | 2019-11-05T23:20:12 | 2019-11-05T23:22:13 | {} | MEMBER |
ClickHouse | ClickHouse | 550,163,300 | 3,529 | wawanawna | Hello. We started see the same exception
`DB::Exception: Cannot allocate block number in ZooKeeper: Coordination::Exception: Connection loss (version 19.11.11.57 (official build))
at ru.yandex.clickhouse.except.ClickHouseExceptionSpecifier.specify(ClickHouseExceptionSpecifier.java:58) ~[clickhouse-jdbc-0.2.jar:... | 2019-11-06T06:19:40 | 2019-11-06T06:19:40 | {
"eyes": 2
} | NONE |
ClickHouse | ClickHouse | 550,203,848 | 7,643 | alexey-milovidov | But our strings have no information about encoding. The only way to solve is to provide another function `javaHashUTF16` that will calculate javaHash under the assumption that the string is in UTF-16. | 2019-11-06T08:36:50 | 2019-11-06T08:36:50 | {} | MEMBER |
ClickHouse | ClickHouse | 550,338,173 | 7,651 | ghost | @alexey-milovidov
Thank you for your code-review. :)
I fixed code according to your recommendations and some more modifications as the below Java's `StringUTF16.hashCode()`.
```Java
class StringUTF16 {
...
public static int hashCode(byte[] value) {
int h = 0;
int length = value.le... | 2019-11-06T14:38:06 | 2019-11-06T14:38:06 | {} | NONE |
ClickHouse | ClickHouse | 550,380,068 | 7,660 | ishirav | Thank you for your reply.
1 - I don't believe this was due to a long-running query - our queries are usually very short (a few seconds or less). Could there be another reason? Maybe it's related to our use of buffer tables in front of each MergeTree table?
2 - that seems to match what I saw in the log. A fix for ... | 2019-11-06T16:08:45 | 2019-11-06T16:08:45 | {} | NONE |
ClickHouse | ClickHouse | 550,568,911 | 7,666 | SaltTan | > It seems some scalar function is not suitable for scalar optimization. Try setting `enable_scalar_subquery_optimization` to false for the moment.
There is no such setting in 19.16.2 | 2019-11-07T00:50:06 | 2019-11-07T00:50:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,985,320 | 7,664 | alesapin | ` Merge with master — Cannot fetch mergecommit` -- strange, pull request is mergeable... | 2019-11-07T08:57:01 | 2019-11-07T08:57:01 | {} | MEMBER |
ClickHouse | ClickHouse | 549,977,276 | 7,608 | alexey-milovidov | Then I run two servers simultaneously in comparison mode (this is the new feature of clickhouse-benchmark that should show statistical significant results if there any):
```
/usr/bin/clickhouse benchmark --port 9000 --port 9001 <<< "SELECT SearchEngineID AS k, uniqCombined(18)(UserID) FROM hits_100m_single GROUP BY... | 2019-11-05T19:18:54 | 2019-11-05T19:20:21 | {} | MEMBER |
ClickHouse | ClickHouse | 550,020,261 | 6,526 | boqu | @alexey-milovidov Thank you for your reply! One more question, is there any plan to backport the fix to 19.11 or any old version? | 2019-11-05T21:07:39 | 2019-11-05T21:07:39 | {} | NONE |
ClickHouse | ClickHouse | 550,571,284 | 7,611 | RaulGuo | > database
My standard_cluster is defined like this:
```
<distribute_cluster>
<shard>
<weight>1</weight>
<internal_replication>true</internal_replication>
<replica>
<host>ip1</host>
<port>9000</port>
<default_database>circular_replica_1</default_database>
</replica>
<replica>
<host... | 2019-11-07T01:00:25 | 2019-11-07T01:00:25 | {} | NONE |
ClickHouse | ClickHouse | 549,912,641 | 7,628 | Tasselmi | @alesapin
```
wget http://hlzstest.oss-cn-hangzhou.aliyuncs.com/clickhouse/clickhouse-common-static-19.15.3.6-2.x86_64.rpm;
wget http://hlzstest.oss-cn-hangzhou.aliyuncs.com/clickhouse/clickhouse-server-19.15.3.6-2.noarch.rpm;
wget http://hlzstest.oss-cn-hangzhou.aliyuncs.com/clickhouse/clickhouse-client-19.15.3.6-... | 2019-11-05T16:59:23 | 2019-11-05T17:01:42 | {} | NONE |
ClickHouse | ClickHouse | 549,914,013 | 7,628 | Tasselmi | > @Tasselmi why haven't you just followed the official instruction from https://clickhouse.yandex/#quick-start ? (which shows up after clicking on CentOS/RedHat)
It is too too slow. about 10kb/s . | 2019-11-05T17:02:35 | 2019-11-05T17:02:35 | {} | NONE |
ClickHouse | ClickHouse | 549,807,661 | 7,613 | nvartolomei | Problem: If one replica has IPs in config but another one has hostnames they will acquire locks and either report on alter or wait for leader to report in different paths.
https://github.com/ClickHouse/Clickhouse/blob/112fc71276517adf29939baf1353048e414f4877/dbms/src/Interpreters/DDLWorker.cpp#L730
`shard_path` o... | 2019-11-05T12:44:40 | 2019-11-05T19:10:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,126,574 | 7,643 | ghost | I found what `-1 -2` mean.
Those mean BOM(Byte Order Mark) of utf-16.
0xFEFF are used in utf16, ([See more details](https://en.wikipedia.org/wiki/Byte_order_mark))
So the javaHash() function have to recognize the charsets of source string, because the Java language calculates a hashCode value based on a charset.
``... | 2019-11-06T03:29:32 | 2019-11-06T03:32:28 | {} | NONE |
ClickHouse | ClickHouse | 550,197,137 | 7,641 | alexey-milovidov | Fuzz test: `SELECT bitTestAny(-103, 422)` | 2019-11-06T08:16:53 | 2019-11-06T08:16:53 | {} | MEMBER |
ClickHouse | ClickHouse | 550,239,192 | 6,642 | CurtizJ | Fixed in master. | 2019-11-06T10:06:50 | 2019-11-06T10:06:50 | {} | MEMBER |
ClickHouse | ClickHouse | 550,269,150 | 7,610 | huang--yijian | I mean the damaged part was dropped successfully by your solution.
What's the root cause of this damaged part?
Is this problem fixed in later versions of ClickHouse ? | 2019-11-06T11:28:22 | 2019-11-06T11:28:22 | {} | NONE |
ClickHouse | ClickHouse | 550,274,582 | 7,653 | alexey-milovidov | I prefer to simply pass all settings as strings in new format. This will make code easier. | 2019-11-06T11:46:42 | 2019-11-06T11:46:42 | {} | MEMBER |
ClickHouse | ClickHouse | 550,312,199 | 7,611 | RaulGuo | > > performance
>
> you need ask boss for more money......
more budget, less salary, lol and crying…… | 2019-11-06T13:38:18 | 2019-11-06T13:38:18 | {} | NONE |
ClickHouse | ClickHouse | 550,108,897 | 7,642 | den-crane | Join does not use indexes. Because currently CH supports only HashJoin. And HashJoin puts a right table into memory into HashTable with joinKey as a lookup key. | 2019-11-06T02:08:29 | 2019-11-06T02:08:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,220,097 | 7,643 | ghost | @alexey-milovidov
Thank you for your help.
I have already made a function like `javaHashUTF16LE` for my job.
If you don't mind. I will make PR about `javaHashUTF16LE` as soon as possible. | 2019-11-06T09:18:56 | 2019-11-06T09:19:51 | {} | NONE |
ClickHouse | ClickHouse | 550,465,910 | 7,663 | alexey-milovidov | Trivial solution is to iterate over all config values and expose them in system table.
But it's unacceptable because too much info will be exposed (various auth info, path to SSL certificate, etc...).
Maybe we can expose only nodes at the first level of XML. This variant looks fragile. | 2019-11-06T19:33:06 | 2019-11-06T19:33:06 | {} | MEMBER |
ClickHouse | ClickHouse | 549,624,421 | 7,627 | chu1070y | >One partition (tuple()) == all. No partition pruning / elimination.
I don't understand tuple() mean.
Does `tuple()` mean do not use partition? | 2019-11-05T01:35:48 | 2019-11-06T00:21:06 | {} | NONE |
ClickHouse | ClickHouse | 550,137,650 | 7,627 | chu1070y | Is there any possible that the Clickhouse automatically creates more than one partition. becuz, when I created tables without partition key, partition is created like below.
> 2019.11.06 13:24:28.903168 [ 54 ] {1719ba64-2f5f-443b-84ad-25a0eeeeda2d} <Debug> demo.order_product (SelectExecutor): Selected 4 parts by dat... | 2019-11-06T04:26:12 | 2019-11-06T04:26:12 | {} | NONE |
ClickHouse | ClickHouse | 550,573,459 | 7,663 | alexey-milovidov | > Maybe something like system.zookeeper?
You just reminded me that `system.zookeeper` is also unsafe :) I'm sure that our information security team is not happy with it.
| 2019-11-07T01:09:58 | 2019-11-07T01:09:58 | {} | MEMBER |
ClickHouse | ClickHouse | 549,889,338 | 1,981 | alexey-milovidov | @lexerom You can also use sha256 password. | 2019-11-05T16:06:23 | 2019-11-05T16:06:23 | {} | MEMBER |
ClickHouse | ClickHouse | 550,109,631 | 7,627 | den-crane | tuple means 1-ary.
All data will be placed into the same one partition. | 2019-11-06T02:11:52 | 2019-11-06T02:11:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,197,557 | 7,540 | alexey-milovidov | `SELECT bitTestAny(-103, 422)` | 2019-11-06T08:18:09 | 2019-11-06T08:18:09 | {} | MEMBER |
ClickHouse | ClickHouse | 550,314,694 | 7,611 | den-crane | > And I have read about the article, but it doesn't work for me, same error still report. I think I cannot find a solution for this. I am using version 19.14.6, I think this article is overtime.
Do you see why in this article in a cluster description replicas have <default_database> section?
I don't see this <def... | 2019-11-06T13:44:43 | 2019-11-06T13:44:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,319,798 | 7,657 | alexey-milovidov | This task is assigned to @RuS2m | 2019-11-06T13:56:34 | 2019-11-06T13:56:34 | {} | MEMBER |
ClickHouse | ClickHouse | 549,884,433 | 6,546 | KochetovNicolai | Good! | 2019-11-05T15:55:16 | 2019-11-05T15:55:16 | {} | MEMBER |
ClickHouse | ClickHouse | 549,889,312 | 7,610 | den-crane | > It works. Thank you for your help!
> What's the root cause ? And is it fixed in later versions of ClickHouse ?
As I understood your first message it was a bit rot? Isn't it? | 2019-11-05T16:06:18 | 2019-11-05T16:06:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,087,492 | 7,608 | amosbird | Ok, it seems some compiler tweaks are required. I'll investigate...
> This is the new feature of clickhouse-benchmark that should guarantee statistical significant results
That's awesome! | 2019-11-06T00:35:54 | 2019-11-06T00:35:54 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 550,204,496 | 7,643 | alexey-milovidov | To avoid BOM, you should specify `utf16be` or `utf16le`:
```
milovidov-Pro-P30 :) SELECT hex(convertCharset('1', 'utf-8', 'utf-16'))
SELECT hex(convertCharset('1', 'utf-8', 'utf-16'))
┌─hex(convertCharset('1', 'utf-8', 'utf-16'))─┐
│ FFFE3100 │
└───────────────────────────... | 2019-11-06T08:38:38 | 2019-11-06T08:38:38 | {} | MEMBER |
ClickHouse | ClickHouse | 550,205,307 | 7,624 | azat | >Just wrap it in toUInt32(N) AS _shard_num
I though that it is too clumsy, but ok
>We can release this feature as experimental and if some difficult issues will be found
Queries without _shard_num does not modified, so it should not produce any problems for basic usage.
>we will revert
or fix? :) | 2019-11-06T08:40:50 | 2019-11-06T08:40:50 | {} | MEMBER |
ClickHouse | ClickHouse | 550,298,144 | 7,647 | yuzhichang | @mrAndersen You can cast array to UInt32 explicitly, for example `bitmapBuild(cast([1,2,3,4] as Array(UInt32)))`. | 2019-11-06T12:58:59 | 2019-11-06T12:58:59 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,430,707 | 7,601 | alesapin | Another one https://github.com/ClickHouse/ClickHouse/pull/7664 | 2019-11-06T18:04:02 | 2019-11-06T18:04:02 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 550,466,487 | 7,663 | alexey-milovidov | We don't have a common repository of available config entries in code. Instead, any place in code can look up for any value in config.
This is unconvenient also because we cannot give user a warning about unknown config elements (in most cases). | 2019-11-06T19:34:38 | 2019-11-06T19:37:41 | {} | MEMBER |
ClickHouse | ClickHouse | 550,960,873 | 6,256 | YiuRULE | Any news about this PR ? most of the changes who has been made was related to this PR https://github.com/ClickHouse/ClickHouse/pull/5045 | 2019-11-07T07:37:11 | 2019-11-07T07:37:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,965,310 | 7,667 | chu1070y | When I INSERT a data to one table not using DISTRIBUTED table engine, The data never be sharding. Am I right? | 2019-11-07T07:53:26 | 2019-11-07T07:53:34 | {} | NONE |
ClickHouse | ClickHouse | 550,975,952 | 7,667 | Slach | @chu1070y yer you are right, DISTRIBUTED table engine split input data by sharding criteria when you insert data to distributed data on any node that's part of the cluster
when you make SELECT from the distributed table then distributed engine make select subqueries to each available nodes in the cluster which descr... | 2019-11-07T08:28:56 | 2019-11-07T08:28:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 549,866,021 | 7,636 | excitoon | Testing 0df4a0208f7650c664dc9a3d1def8db322f69560 | 2019-11-05T15:12:11 | 2019-11-05T15:12:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 549,954,240 | 6,642 | CurtizJ | But thanks for reproducable example. This bug reproduces very rarely. | 2019-11-05T18:25:13 | 2019-11-05T18:25:13 | {} | MEMBER |
ClickHouse | ClickHouse | 549,996,837 | 7,630 | den-crane | streaming parser:
```
INSERT INTO json_load_test (idx) VALUES (2147483648);
INSERT INTO json_load_test FORMAT JSONEachRow {"idx": 2147483648}
SELECT * FROM json_load_test
┌─────────idx─┐
│ -2147483648 │
└─────────────┘
┌─────────idx─┐
│ -2147483648 │
└─────────────┘
```
----------
not streaming parser... | 2019-11-05T20:06:11 | 2019-11-05T20:06:33 | {
"+1": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,194,221 | 7,624 | azat | > > it does not support qualified columns
> Not too strange
So do you suggest to ignore this for now? | 2019-11-06T08:07:55 | 2019-11-06T08:07:55 | {} | MEMBER |
ClickHouse | ClickHouse | 550,317,512 | 7,611 | RaulGuo | > > And I have read about the article, but it doesn't work for me, same error still report. I think I cannot find a solution for this. I am using version 19.14.6, I think this article is overtime.
>
> Do you see why in this article in a cluster description replicas have <default_database> section?
> I don't see thi... | 2019-11-06T13:51:32 | 2019-11-06T13:51:56 | {} | NONE |
ClickHouse | ClickHouse | 550,958,398 | 7,667 | Slach | you can INSERT into each table in the shard and SELECT from each table in the shard with using a direct connection to selected host
but if need SELECT from all tables which shard contains you should use DISTRIBUTED table engine | 2019-11-07T07:28:07 | 2019-11-07T07:28:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 549,898,417 | 7,633 | den-crane | ERROR: type should be string, got "\r\nhttps://clickhouse.yandex/docs/en\r\n\r\n_The behavior of parsing incorrect dates is implementation specific. ClickHouse may return zero date, throw an exception or do \"natural\" overflow._" | 2019-11-05T16:26:54 | 2019-11-05T16:26:54 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 549,958,394 | 6,526 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/pull/7183
19.16. | 2019-11-05T18:34:34 | 2019-11-05T18:35:05 | {} | MEMBER |
ClickHouse | ClickHouse | 550,204,684 | 7,643 | alexey-milovidov | And probably we should add a function `javaHashUTF16LE`. | 2019-11-06T08:39:12 | 2019-11-06T08:39:12 | {} | MEMBER |
ClickHouse | ClickHouse | 550,209,634 | 7,624 | alexey-milovidov | > Queries without _shard_num does not modified, so it should not produce any problems for basic usage.
That's perfect.
PS. I've already reviewed the code. | 2019-11-06T08:52:47 | 2019-11-06T08:52:47 | {} | MEMBER |
ClickHouse | ClickHouse | 550,327,153 | 7,658 | akuzm | https://github.com/ClickHouse/ClickHouse/pull/7655 -- that one looks better in that it ensures a consistent schedule between has/setTotals(). | 2019-11-06T14:13:44 | 2019-11-06T14:14:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,348,438 | 7,611 | den-crane | Please provide your <distribute_cluster> XML
And you should omit '' database in Distributed table ENGINE = Distributed( testcluster, '', tc_shard, rand() )
Also **DDL on cluster won't work.** https://github.com/ClickHouse/ClickHouse/issues/7611#issuecomment-549432847 | 2019-11-06T15:00:12 | 2019-11-06T15:03:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,430,442 | 7,638 | alesapin | Fix https://github.com/ClickHouse/ClickHouse/pull/7664 | 2019-11-06T18:03:24 | 2019-11-06T18:03:24 | {} | MEMBER |
ClickHouse | ClickHouse | 550,962,396 | 4,959 | jackpgao | ```
┌─version()──┐
│ 19.13.3.26 │
└────────────┘
```
```
2019.11.07 15:35:47.483092 [ 97 ] {97a4ed62-edb7-44ac-92eb-293657785370} <Error> executeQuery: Code: 33, e.displayText() = DB::Exception: Cannot read all data. Bytes read: 293. Bytes expected: 1070. (version 19.13.3.26)
```
| 2019-11-07T07:42:52 | 2019-11-07T07:42:52 | {} | NONE |
ClickHouse | ClickHouse | 549,875,871 | 7,546 | ariesdevil | CMake with the following instruction looks good:
```shell
cmake .. -DCMAKE_CXX_COMPILER=`which clang++` -DCMAKE_C_COMPILER=`which clang` -DCMAKE_BUILD_TYPE=Debug -DENABLE_MYSQL=0 -DUSE_STATIC_LIBRARIES=0 -DNO_WERROR=1 -DENABLE_TESTS=0 -DUSE_RDKAFKA=0
```
the cmake output is:
```
-- The C compiler identification i... | 2019-11-05T15:34:51 | 2019-11-05T15:37:22 | {} | NONE |
ClickHouse | ClickHouse | 549,899,652 | 7,630 | den-crane | Not related to JSON. CH does overflow by performance reasons.
```
SELECT toInt8(2147483644)
┌─toInt8(2147483644)─┐
│ -4 │
└────────────────────┘
SELECT toInt32('2147483648')
┌─toInt32('2147483648')─┐
│ -2147483648 │
└───────────────────────┘
``` | 2019-11-05T16:29:42 | 2019-11-05T16:30:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,030,444 | 6,526 | boqu | @alexey-milovidov Thank you! | 2019-11-05T21:32:36 | 2019-11-05T21:32:36 | {} | NONE |
ClickHouse | ClickHouse | 550,315,974 | 7,610 | den-crane | > What's the root cause of this damaged part?
I don't know. Accident memory / disk corruption? | 2019-11-06T13:47:58 | 2019-11-06T13:47:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,350,833 | 7,651 | ghost | @alexey-milovidov
I learned a lot during code-review.
Thank you very much. | 2019-11-06T15:05:14 | 2019-11-06T15:05:14 | {} | NONE |
ClickHouse | ClickHouse | 550,198,115 | 7,624 | alexey-milovidov | We can release this feature as experimental and if some difficult issues will be found, we will revert. | 2019-11-06T08:19:58 | 2019-11-06T08:19:58 | {} | MEMBER |
ClickHouse | ClickHouse | 550,248,098 | 7,646 | alesapin | Please, provide full traces, describe in details which messages format you use and output of `SHOW CREATE TABLE` for Kafka table. | 2019-11-06T10:29:44 | 2019-11-06T10:29:44 | {} | MEMBER |
ClickHouse | ClickHouse | 550,298,633 | 7,647 | mrAndersen | Thanks! | 2019-11-06T13:00:26 | 2019-11-06T13:00:26 | {} | NONE |
ClickHouse | ClickHouse | 550,427,275 | 7,653 | vitlibar | > I prefer to simply pass all settings as strings in new format. This will make code easier.
Done. | 2019-11-06T17:56:26 | 2019-11-06T17:56:26 | {} | MEMBER |
ClickHouse | ClickHouse | 550,067,338 | 7,601 | alexey-milovidov | I remember this issue (it has long history and is difficult to fix in a proper way).
Let's summarize:
- we copy Context inside InterpreterSelectQuery to support modification of context (Settings) for subqueries;
- InterpreterSelectQuery lives shorter than query pipeline. It's used just to build the query pipeline ... | 2019-11-05T23:17:38 | 2019-11-05T23:17:38 | {} | MEMBER |
ClickHouse | ClickHouse | 550,072,003 | 7,518 | alexey-milovidov | Let's optimize only typical cases:
- start, end and step can be converted to the common supertype before processing.
It will not have any performance impact if they already have the same type or some arg of another type but constant.
Now we avoid cubic/quadradic template instantiations.
How different arguments ... | 2019-11-05T23:34:13 | 2019-11-05T23:34:13 | {} | MEMBER |
ClickHouse | ClickHouse | 550,073,372 | 7,624 | alexey-milovidov | > will this functionality be helpful
I always wanted this functionality. We event tried to implement something like that long time ago but faced some difficulties. (We tried to implement it in almost the same way but there was no WITH clause). | 2019-11-05T23:39:04 | 2019-11-05T23:47:00 | {} | MEMBER |
ClickHouse | ClickHouse | 550,075,861 | 7,624 | alexey-milovidov | > type of the _shard_num is not defined
Just wrap it in `toUInt32(N) AS _shard_num`
> it does not support qualified columns
Not too strange. | 2019-11-05T23:48:42 | 2019-11-05T23:48:42 | {} | MEMBER |
ClickHouse | ClickHouse | 550,076,424 | 7,624 | alexey-milovidov | If we will not find some other weird cases of this implementation, I think, it is acceptable. | 2019-11-05T23:51:00 | 2019-11-05T23:51:00 | {} | MEMBER |
ClickHouse | ClickHouse | 550,197,821 | 7,624 | alexey-milovidov | > So do you suggest to ignore this for now?
Yes. | 2019-11-06T08:19:04 | 2019-11-06T08:19:04 | {} | MEMBER |
ClickHouse | ClickHouse | 550,311,642 | 7,611 | RaulGuo | > https://www.altinity.com/blog/2018/5/10/circular-replication-cluster-topology-in-clickhouse
>
> But it's not fully supported:
>
> * DDL on cluster won't work.
> * Double distributed joins will fail with exception "Table not found"
>
> Personally I don't recommend this schema because second replicas kill per... | 2019-11-06T13:36:48 | 2019-11-06T13:36:48 | {} | NONE |
ClickHouse | ClickHouse | 550,394,498 | 7,627 | den-crane | `4 parts `
part != partition (parts are pieces of a partition)
Check the doc https://clickhouse.yandex/docs/en/operations/table_engines/mergetree/
| 2019-11-06T16:40:08 | 2019-11-06T16:40:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 549,953,084 | 7,632 | alexey-milovidov | Could you please also add an integration test? | 2019-11-05T18:22:35 | 2019-11-05T18:22:35 | {} | MEMBER |
ClickHouse | ClickHouse | 550,247,164 | 7,518 | sundy-li | @alexey-milovidov Thanks for your detail advice.
I have updated the code and run the performance tests, it's so wried that the second performance-test outperforms the first in my ubuntu 6core server.
```
[
{
"hostname": "abc.ubuntu.local",
"main_metric": "max_rows_per_second",
"num_cores": 6,
... | 2019-11-06T10:27:24 | 2019-11-06T10:27:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,378,364 | 7,646 | filimonov | I saw that error several times when there were more consumers than partitions in the Kafka topic.
Can you confirm that you have enough partitions for all the consumers?
You can check partitions and consumers assignment with `kafka-consumer-groups` command-line tool.
PS. Probably we can/should try to report t... | 2019-11-06T16:04:59 | 2019-11-06T16:06:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,415,443 | 7,662 | amosbird | Second part of https://github.com/ClickHouse/ClickHouse/pull/7459 | 2019-11-06T17:27:41 | 2019-11-06T17:27:41 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 550,522,292 | 7,660 | filimonov | > I don't know what you mean by MV.
So rather you don't use that :) Materialized views.
| 2019-11-06T22:06:13 | 2019-11-06T22:06:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,565,114 | 7,666 | amosbird | It seems some scalar function is not suitable for scalar optimization. Try setting `enable_scalar_subquery_optimization` to false for the moment. | 2019-11-07T00:33:25 | 2019-11-07T00:33:25 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 550,866,104 | 7,660 | ishirav | @filimonov - ahh, materialized views. We don't currently use them.
@excitoon - good idea. Here are the OPS of the two EBS volumes:
Here is their throughput:
![image](https://user-images.githubusercon... | 2019-11-07T06:16:40 | 2019-11-07T06:16:40 | {
"heart": 2
} | NONE |
ClickHouse | ClickHouse | 549,803,457 | 6,546 | KochetovNicolai | #7156
Should be fixed since 19.14 | 2019-11-05T12:31:01 | 2019-11-05T15:42:56 | {} | MEMBER |
ClickHouse | ClickHouse | 549,882,470 | 7,546 | alexey-milovidov | @ariesdevil
It's enough to build just `clickhouse` binary:
`ninja clickhouse` | 2019-11-05T15:50:32 | 2019-11-05T15:50:32 | {} | MEMBER |
ClickHouse | ClickHouse | 549,993,185 | 7,630 | dmlogv | But why `INSERT INTO ... FORMAT JSONEachRow` overflow behaviour differs from `INSERT INTO ... VALUES`?
There is an explanation in the documentation? | 2019-11-05T19:57:33 | 2019-11-05T19:57:33 | {} | NONE |
ClickHouse | ClickHouse | 550,005,307 | 7,641 | 4ertus2 | #4219 | 2019-11-05T20:27:34 | 2019-11-05T20:27:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,258,740 | 7,646 | jsding | Hi,
Thanks for your reply.
```
CREATE TABLE gateway_error (
ip String,
message String,
gateway String,
timestamp UInt64
) ENGINE = Kafka SETTINGS kafka_broker_list = '127.0.0.1:9092',
kafka_topic_list = 'gateway_error',
kafka_group_name = 'ck... | 2019-11-06T10:57:53 | 2019-11-06T10:57:53 | {} | NONE |
ClickHouse | ClickHouse | 550,029,382 | 6,526 | alexey-milovidov | Only to 19.15 and 19.14. | 2019-11-05T21:29:54 | 2019-11-05T21:29:54 | {} | MEMBER |
ClickHouse | ClickHouse | 550,241,638 | 7,645 | excitoon | I will need to add tests on this setting. | 2019-11-06T10:13:14 | 2019-11-06T10:13:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,373,382 | 7,660 | filimonov | There are 2 problems here:
1) for some reason parts were not removed from the old location - that can happen if some long-running query still reading old parts, and locking them. We will recheck that part.
2) when the same part exists on one (or more) disk - that can happen when the source data part was not removed f... | 2019-11-06T15:54:10 | 2019-11-06T15:56:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,382,963 | 7,660 | filimonov | And how many parts (including / excluding duplicates) do you have in that table? (You can just count the folders)
Can you explain a bit you data pipeline? It looks like some MV & Buffer tables were involved. | 2019-11-06T16:15:16 | 2019-11-06T16:17:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,394,414 | 7,660 | ishirav | @filimonov the problem was not specific to a single table, there were duplicates in many tables.
We have 24 tables, some have only a few parts while the largest 5 tables have 150 to 250 parts. Across all table there are ~1500 parts.
Interestingly, if I look at the parts that ClickHouse moved to the /data-2 volume, ... | 2019-11-06T16:39:57 | 2019-11-06T16:39:57 | {} | NONE |
ClickHouse | ClickHouse | 550,467,490 | 7,663 | alexey-milovidov | Another solution is to have a system table that will expose only a fixed list of settings specified directly in code. I don't like this solution because it will be hard to support (to keep it in mind when we add every new config value). But this variant can be acceptable.
| 2019-11-06T19:37:10 | 2019-11-06T19:38:30 | {} | MEMBER |
ClickHouse | ClickHouse | 550,517,541 | 7,660 | excitoon | @ishirav Can you share I/O utilization statistics of disks during move? | 2019-11-06T21:53:39 | 2019-11-06T21:53:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 550,524,683 | 7,663 | filimonov | Maybe something like system.zookeeper? + limit access by some xml attribute?
I.e.
```sql
SELECT * FROM system.server_settings WHERE path='/max_connections'.
```
and in config
```
<password hide="1">secret</password>
```
? | 2019-11-06T22:12:55 | 2019-11-06T22:13:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 556,978,837 | 7,763 | alexey-milovidov | Status:
- we found that address hint is still needed.
What's remaining:
- to fix a few typos;
- to add a test to validate future changes.
Do you want to open another pull request for that? | 2019-11-21T08:40:45 | 2019-11-21T08:40:45 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.