Datasets:
open-index
/
open-github-issues

Dataset card Data Studio Files
xet
 Community
1
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
560,963,571
7,973
sundy-li
> Maybe our new CustomSeparated format will suite better for this case I do agree with that.
2019-12-03T01:59:20
2019-12-03T02:03:24
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,680,607
8,028
alexey-milovidov
@akuzm has not finished support for MSan.
2019-12-04T14:53:29
2019-12-04T14:53:29
{}
MEMBER
ClickHouse
ClickHouse
561,710,363
7,859
alesapin
> just reuse system.merges as the nature of moves is quite similar to merges. Moves absolutely different from merges and work at another level of abstraction (`MergeTreeData` level). `system.(background_)jobs` or `system.(background_)operations` looks good to me. I think this table should contain data from `system...
2019-12-04T16:01:22
2019-12-04T16:01:22
{}
MEMBER
ClickHouse
ClickHouse
561,166,520
7,795
alexey-milovidov
This performance regression is almost acceptable but not quite.
2019-12-03T13:25:45
2019-12-03T13:25:45
{}
MEMBER
ClickHouse
ClickHouse
561,467,498
8,017
den-crane
CH starts to execute`Alter delete` in background right after you called it (if it is not blocked by other running delete or a running merge). It takes so long because it rewrites all data. Try `truncate table` or `drop partition` if you need to remove all rows.
2019-12-04T04:07:00
2019-12-04T04:07:00
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,652,721
8,027
alexey-milovidov
`-DENABLE_TCMALLOC=ON -DENABLE_JEMALLOC=OFF` That's the reason. Right now the build with tcmalloc is broken.
2019-12-04T13:47:25
2019-12-04T13:47:25
{}
MEMBER
ClickHouse
ClickHouse
560,420,190
7,970
nikitamikhaylov
Sorry. FYI: Clang is another compiler for C++. You can install **clang 8** and then compile without any problems. We will update our documentation ASAP.
2019-12-02T14:30:24
2019-12-02T14:30:24
{ "+1": 2 }
MEMBER
ClickHouse
ClickHouse
560,949,719
7,535
SaltTan
I observe the same behaviour in 19.16.6.17 and 19.17.4.11. I think your test is incorrect. When you create the source table for the second time it is empty, so the invalidate query returns a different value (in your test '0000-00-00 00:00:00' instead of '2019-12-02 16:22:05') and the dictionary gets reloaded (at 2019...
2019-12-03T01:00:50
2019-12-03T01:00:50
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
560,957,799
7,307
alexey-milovidov
@stavrolia You did not answered how big is the improvement.
2019-12-03T01:34:52
2019-12-03T01:35:07
{}
MEMBER
ClickHouse
ClickHouse
560,969,850
6,802
alexey-milovidov
@Akazz ?
2019-12-03T02:27:22
2019-12-03T02:27:22
{}
MEMBER
ClickHouse
ClickHouse
561,330,063
6,802
Akazz
I am going to pick this up _asap_! > @Akazz This is still to do. And only a few steps remain.
2019-12-03T19:54:38
2019-12-03T19:54:38
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,380,538
8,014
volfco
Removing the bloom filter index (alter table histograms on cluster metrics_qa drop index tvl_bf;) drops the memory usage massively: ``` metrics_qa :) select database, table, elapsed, progress*100, num_parts, formatReadableSize(memory_usage) from system.merges; SELECT database, table, elapsed, ...
2019-12-03T22:11:32
2019-12-03T22:11:53
{}
NONE
ClickHouse
ClickHouse
561,509,676
6,168
YiuRULE
Does we got update on the PR ? :(
2019-12-04T07:12:35
2019-12-04T07:12:35
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,714,009
8,031
den-crane
It's expected, not a bug. https://clickhouse.yandex/docs/ _The behavior of parsing incorrect dates is implementation specific. ClickHouse may return zero date, throw an exception or do "natural" overflow._ ``` SELECT toDate('2019-09-31') ┌─toDate('2019-09-31')─┐ │ 2019-10-01 │ └────────────────...
2019-12-04T16:09:28
2019-12-04T16:10:44
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,793,620
8,003
niting112
@brefre @filimonov Is this a bug for some particular use case or in general when using kafka engine ? We are also planning to update to version 19.17.4.11 because there is another important bug fix in this version related to kafka engine (Prevent message duplication when producing Kafka table has any MVs selecting from...
2019-12-04T19:07:41
2019-12-04T19:10:37
{}
NONE
ClickHouse
ClickHouse
560,970,788
7,973
alexey-milovidov
BTW the proposed alias `HiveTextFile` for a preconfigured `CustomSeparated` format can still make sense.
2019-12-03T02:31:49
2019-12-03T02:31:49
{}
MEMBER
ClickHouse
ClickHouse
561,050,196
7,646
ghuname
@abyssz Can you at least report table/topic that is relevant to the message for: ``` 2019.11.06 01:16:20.394819 [ 48 ] {} void DB::StorageKafka::threadFunc(): std::exception. Code: 1001, type: cppkafka::HandleException, e.what() = Local: Timed out, version = 19.15.3.6 (official build) ```
2019-12-03T08:12:22
2019-12-03T08:12:22
{}
NONE
ClickHouse
ClickHouse
561,277,879
8,003
filimonov
Thanks for the report i think i can fix that, right now working with code around that place. Just to ensure - your messages in kafka have no timestamps?
2019-12-03T17:41:43
2019-12-03T17:41:43
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,353,366
7,989
andrelec1
> 2020-03-19 02:00:00 and 2020-03-19 03:00:00 or 2020-10-25 02:00:00 and 2020-10-25 03:00:00 This is a specifique date where day as 23h / 25h ... But like a say,i think i will reaply timezone in the back end and keep utc date in db ... But like you say, i start making a poc ;) Thank again.
2019-12-03T20:56:47
2019-12-03T20:57:13
{}
NONE
ClickHouse
ClickHouse
561,652,847
8,027
alexey-milovidov
Why do you need tcmalloc instead of jemalloc?
2019-12-04T13:47:46
2019-12-04T13:47:46
{}
MEMBER
ClickHouse
ClickHouse
560,952,035
7,900
alexey-milovidov
> Can we add AsynchronousMetrics to the system.metric_log ? We have not added it, because they are updated in longer intervals (60 seconds by default) in contrast to other values, that are always up to date and can be written as frequent as needed.
2019-12-03T01:10:44
2019-12-03T01:10:44
{}
MEMBER
ClickHouse
ClickHouse
561,135,851
6,528
zombiemonkey
Have realized this isn't going to work because of processing per block. Is there any way to force consecutive blocks to synchronize?
2019-12-03T11:57:06
2019-12-03T11:57:06
{}
NONE
ClickHouse
ClickHouse
561,272,568
7,989
den-crane
Make a POC. CH compress data 10-100 times. > like moosefs Don't even try. CH is superfast because the data is near on a local disk. Network will be huge bottleneck. >About DST Did you check my example? What do you think about this 2019-01-01 **00:00:00** / 2019-08-01 **00:00:00** 2019-01-01 **01:00:00*...
2019-12-03T17:28:03
2019-12-03T17:30:29
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,626,826
7,998
alesapin
Fix https://github.com/ClickHouse/ClickHouse/pull/8028.
2019-12-04T12:41:50
2019-12-04T12:41:50
{}
MEMBER
ClickHouse
ClickHouse
561,627,594
8,028
alesapin
Add check of `clickhouse-odbc-bridge` to GlibcCompatibilityCheck.
2019-12-04T12:44:33
2019-12-04T12:44:33
{}
MEMBER
ClickHouse
ClickHouse
561,678,701
7,535
alesapin
Fix https://github.com/ClickHouse/ClickHouse/pull/8029.
2019-12-04T14:49:17
2019-12-04T14:49:32
{}
MEMBER
ClickHouse
ClickHouse
561,695,681
7,998
alesapin
Merged, we will backport this fix to 19.17, 19.16 and 19.15 if possible.
2019-12-04T15:28:08
2019-12-04T15:28:08
{}
MEMBER
ClickHouse
ClickHouse
560,601,658
7,986
filimonov
What do you mean by crash? That looks like a normal error report.
2019-12-02T21:32:31
2019-12-02T21:32:31
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
560,881,658
7,981
alexey-milovidov
@stavrolia failed to answer.
2019-12-02T23:23:25
2019-12-02T23:23:25
{}
MEMBER
ClickHouse
ClickHouse
561,230,094
7,785
nvartolomei
Let's have a `FREEZE TABLE [db.]name [PARTITION partition_expr]`, @alexey-milovidov?
2019-12-03T15:52:37
2019-12-03T15:52:37
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,279,721
6,731
filimonov
Reproduced, was broken after introducing virtual columns. https://github.com/ClickHouse/ClickHouse/pull/5382/ Will address that soon.
2019-12-03T17:46:13
2019-12-03T17:46:13
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,573,936
8,025
alexey-milovidov
`lld` must be selected by default if it is available. Please check the CMakeLists. If it's not true, we should fix that issue first.
2019-12-04T10:22:15
2019-12-04T10:23:01
{ "+1": 1 }
MEMBER
ClickHouse
ClickHouse
561,603,473
8,024
alesapin
One stateless test is enough.
2019-12-04T11:27:10
2019-12-04T11:27:10
{}
MEMBER
ClickHouse
ClickHouse
561,796,638
8,014
volfco
``` CREATE TABLE default.histograms (`Timestamp` DateTime CODEC(DoubleDelta, LZ4), `Client` LowCardinality(String), `Path` LowCardinality(String), `Tags.Key` Array(LowCardinality(String)), `Tags.Value` Array(String) CODEC(LZ4HC(0)), `Range` Array(Float32), `Steps` Float32, `IndexNotation` Int32, `Histogram` Array(Floa...
2019-12-04T19:15:04
2019-12-04T19:15:24
{}
NONE
ClickHouse
ClickHouse
560,889,603
7,985
alexey-milovidov
Fixed.
2019-12-02T23:27:03
2019-12-02T23:27:03
{}
MEMBER
ClickHouse
ClickHouse
561,099,407
7,307
stavrolia
@alexey-milovidov two times faster ![image_2019-12-03_13-23-06](https://user-images.githubusercontent.com/4096393/70042543-1ea91600-15d0-11ea-8377-ca81551a27ac.png)
2019-12-03T10:19:28
2019-12-03T10:23:44
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,150,039
7,946
alesapin
And of course, this is very small step to the described task. MergeTree codebase havily relies on local filesystem and the proposed idea seems very challenging.
2019-12-03T12:41:06
2019-12-03T12:41:06
{}
MEMBER
ClickHouse
ClickHouse
560,434,861
7,440
den-crane
the old behavior of `system reload dictionary` led to SELECT `dummy`, (dummy) AS `two` FROM `system`.`one` WHERE now() >= **'0000-00-00 00:00:00'**
2019-12-02T15:04:36
2019-12-02T15:14:40
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,689,443
7,820
alesapin
Fix https://github.com/ClickHouse/ClickHouse/pull/8032.
2019-12-04T15:13:54
2019-12-04T15:13:54
{}
MEMBER
ClickHouse
ClickHouse
561,008,761
7,982
excitoon
> @excitoon build on master is run on every commit and you can check the results here: > https://github.com/ClickHouse/ClickHouse/commits/master Okay, thank you!
2019-12-03T05:43:53
2019-12-03T05:43:53
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,338,934
4,729
aadant
This is similar to ALTER TABLE ... EXCHANGE PARTITION in MySQL see https://dev.mysql.com/doc/refman/5.6/en/partitioning-management-exchange.html
2019-12-03T20:17:44
2019-12-03T20:17:44
{}
NONE
ClickHouse
ClickHouse
561,544,860
8,002
herokillerJ
i write wrong task config,i fix it in the config file and retry but still not work,then i was change the taskpath to a new dir,then it work, if use same task path many times,config file is the first time to upload,
2019-12-04T09:06:33
2019-12-04T09:06:33
{}
NONE
ClickHouse
ClickHouse
560,430,456
7,440
den-crane
@alesapin @vitlibar >This behavior seems to be expected This is the **new** behavior and this new behavior causes a problem with maintaining of dictionaries with partial updates. Now `system reload dictionary` initiates a new cycle of an update of earlier loaded dictionaries and for dictionaries with partial upd...
2019-12-02T14:54:40
2019-12-02T17:34:49
{}
CONTRIBUTOR
ClickHouse
ClickHouse
560,964,876
7,795
alexey-milovidov
@Akazz have you checked the performance manually?
2019-12-03T02:05:15
2019-12-03T02:05:15
{}
MEMBER
ClickHouse
ClickHouse
561,073,528
7,998
huang--yijian
Also Altinity does not provide RPM of ClickHouse 19.17 for el/6.
2019-12-03T09:17:07
2019-12-03T09:17:07
{}
NONE
ClickHouse
ClickHouse
561,530,710
8,014
alesapin
Show your table definition (`CREATE QUERY`), please.
2019-12-04T08:24:27
2019-12-04T08:24:27
{}
MEMBER
ClickHouse
ClickHouse
561,548,772
8,018
alesapin
You use an outdated ClickHouse version, which doesn't support modern MergeTree DDL. It was introduced more than 2 years ago, so I recommend you to update your version to current stable (and also read compatibility notes in changelog).
2019-12-04T09:17:06
2019-12-04T09:17:30
{}
MEMBER
ClickHouse
ClickHouse
561,581,241
7,998
alesapin
Dependency comes from `clickhouse-odbc-bridge` :( ``` $ readelf -s clickhouse-odbc-bridge | grep '@GLIBC_' | grep '2.16' 120: 0000000000000000 0 FUNC GLOBAL DEFAULT UND aligned_alloc@GLIBC_2.16 (6) ```
2019-12-04T10:32:43
2019-12-04T10:32:43
{}
MEMBER
ClickHouse
ClickHouse
561,659,251
7,535
alesapin
Tried to reproduce with test https://github.com/ClickHouse/ClickHouse/pull/8029/files. I think I'm doing something wrong :(
2019-12-04T14:03:57
2019-12-04T14:03:57
{}
MEMBER
ClickHouse
ClickHouse
561,710,624
7,837
filimonov
> With removing of kafka_row_delimiter you could probably also remove "1 kafka message = possible multiple CH rows" and go to "1 kafka message= 1 CH row". Makes kafka offsets etc simpler. One size does not suit all. It should support both options (and is already support that), and user can choose.
2019-12-04T16:01:55
2019-12-04T16:01:55
{}
CONTRIBUTOR
ClickHouse
ClickHouse
560,478,604
7,971
petrushenkovtb
Latest Clickhouse in container
2019-12-02T16:44:46
2019-12-02T16:44:46
{}
NONE
ClickHouse
ClickHouse
560,491,328
7,977
den-crane
2. Kafka engine does not store data. Select from KafkaEngine reads kafka topic and moves offset. >Does MATERIALIZED VIEW implement select query Kafka engine pulls data from a topic and pushes to MV.
2019-12-02T17:15:52
2019-12-02T17:15:52
{}
CONTRIBUTOR
ClickHouse
ClickHouse
560,967,758
7,621
alexey-milovidov
@abyss7 You have failed to review this PR.
2019-12-03T02:17:58
2019-12-03T02:17:58
{}
MEMBER
ClickHouse
ClickHouse
561,114,679
7,989
andrelec1
thank for your answere ... 400k rows by hit, actualy something like 50hits day ^^ ... so isn't a constant charge, just a spades on request ... but actualy on the mariadb, the request is realy long, and if two client request the db in same time, everything slow down and become unsuable ... ( cache is very big or/...
2019-12-03T10:56:07
2019-12-03T10:56:07
{}
NONE
ClickHouse
ClickHouse
561,121,722
7,050
jackpgao
``` ┌─version()──┐ │ 19.15.5.18 │ └────────────┘ ``` ``` 2019.12.03 16:15:53.124934 [ 500 ] {} <Fatal> BaseDaemon: ######################################## 2019.12.03 16:15:53.125004 [ 500 ] {} <Fatal> BaseDaemon: (version 19.15.5.18 (official build)) (from thread 424) Received signal Segmentation fault (11)...
2019-12-03T11:13:45
2019-12-03T11:13:45
{}
NONE
ClickHouse
ClickHouse
561,145,601
7,795
alexey-milovidov
> Fix logical functions with UInt8 numbers, when they not equal to 1. What was the cause of this error? Maybe we are using bit ANDs instead of logical ANDs to avoid less efficient short-curcuit code in multiple AND chains?
2019-12-03T12:27:51
2019-12-03T12:27:51
{}
MEMBER
ClickHouse
ClickHouse
561,158,684
7,795
CurtizJ
> Maybe we are using bit ANDs instead of logical ANDs Yes, we use them, but for another reason. I've tried to expalin it in [comment](https://github.com/ClickHouse/ClickHouse/issues/7772#issuecomment-554399052).
2019-12-03T13:07:01
2019-12-03T13:09:03
{}
MEMBER
ClickHouse
ClickHouse
561,268,475
7,996
den-crane
it does not. CH does not support prepared queries.
2019-12-03T17:18:28
2019-12-03T17:18:28
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,699,915
6,989
alesapin
I have fixed a very similar issue here https://github.com/ClickHouse/ClickHouse/pull/6452. But 19.4.3.3 should have this fix. I have no idea, how the adaptive part (`201908_0_77_21_78`) ended up in your fixed granularity table. Reopen this issue, if you will be able to reproduce this.
2019-12-04T15:37:53
2019-12-04T15:37:53
{}
MEMBER
ClickHouse
ClickHouse
561,716,652
3,929
alesapin
@zhang2014 Do you have any plans to continue to work on this PR?
2019-12-04T16:15:39
2019-12-04T16:15:39
{}
MEMBER
ClickHouse
ClickHouse
561,727,257
8,017
den-crane
Yes. CH evaluates which parts should be mutated (using where conditions) and do a special merge for those parts (CH writes each parts' columns to a new file excluding removed rows -> replaces old parts with new).
2019-12-04T16:38:36
2019-12-04T16:38:36
{}
CONTRIBUTOR
ClickHouse
ClickHouse
560,979,298
7,778
hustnn
In addition to memory profiler, how about also add the cpu usage profiler for each stream (operator)? I am looking at it. @filimonov
2019-12-03T03:14:00
2019-12-03T03:14:13
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,137,738
8,000
hombit
@javisantana Thank you for your answer! In my case I need spherical coordinates not for common geographical problems so I cannot use this approach. I have to care of high latitude points and it could be very tricky to solve this problems on client side.
2019-12-03T12:03:10
2019-12-03T12:59:54
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,245,796
7,906
smirnov-vs
Hello. I have two distributed tables and I need to JOIN them without GLOBAL, cause data is sharded by same key. So I wrote something like ``` SELECT * FROM ( SELECT id, * FROM table1 WHERE <...> ) ALL LEFT JOIN ( SELECT id, * FROM table2_local WHERE <...> ) USING (id) ``` table1 is Distrubuted, ta...
2019-12-03T16:27:17
2019-12-03T16:27:17
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,400,207
8,003
brefre
Hi, This is a good question. I assume you are talking about the internal timestamps in the kafka messages, and not my JSON structure. Many months ago, we upgraded our kafka cluster from a older version. I have found that topics created in the older version of kafka do not have timestamps, but newly created topic...
2019-12-03T23:14:27
2019-12-03T23:14:27
{}
NONE
ClickHouse
ClickHouse
561,651,727
8,028
alexey-milovidov
We can treat malloc libraries in the same way we do with glibc-compatibility library (add as default library).
2019-12-04T13:44:34
2019-12-04T13:44:34
{}
MEMBER
ClickHouse
ClickHouse
560,904,144
7,977
chu1070y
> Kafka engine pulls data from a topic and pushes to MV. You are saying that Kafka engine is the one which sends the data to MV? I was thinking MV is the one that pull data of itself from Kafka engine.
2019-12-02T23:33:53
2019-12-02T23:35:58
{}
NONE
ClickHouse
ClickHouse
561,007,801
7,981
stavrolia
@alexey-milovidov now we don't
2019-12-03T05:39:49
2019-12-03T05:39:49
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,144,948
7,795
alexey-milovidov
@CurtizJ Let's add this query to performance tests in separate PR?
2019-12-03T12:25:34
2019-12-03T12:25:34
{}
MEMBER
ClickHouse
ClickHouse
561,214,658
7,785
nvartolomei
Can't think of a good way to change current behaviour as ALTER commands can be chained. Ie what should be returned in the following case? ``` ALTER TABLE t FREEZE, ADD COLUMN `x` UInt8 Ok. 0 rows in set. Elapsed: 0.004 sec ```
2019-12-03T15:19:35
2019-12-03T15:19:42
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,480,507
8,017
byx313
> CH starts to execute`Alter delete` in background right after you called it (if it is not blocked by other running delete or a running merge). It takes so long because it rewrites all data. > > Try `truncate table` or `drop partition` if you need to remove all rows. Do you mean that （1）CH get all data （2） check ...
2019-12-04T05:15:00
2019-12-04T05:57:19
{}
NONE
ClickHouse
ClickHouse
561,547,255
8,012
alesapin
All required tests passed. Rest of them failed because of CI problems.
2019-12-04T09:13:09
2019-12-04T09:13:09
{}
MEMBER
ClickHouse
ClickHouse
561,627,176
8,028
alesapin
Maybe it's not the best fix for this problem, but my CMake skills is poor. I hope @abyss7 or @proller can propose something better.
2019-12-04T12:42:59
2019-12-04T12:42:59
{}
MEMBER
ClickHouse
ClickHouse
561,690,037
8,028
alesapin
Glibc check works https://clickhouse-test-reports.s3.yandex.net/0/a651a1e05bfecfa09fe6e498f8fa78dd8748b92d/compatibility_check.html.
2019-12-04T15:15:08
2019-12-04T15:15:08
{}
MEMBER
ClickHouse
ClickHouse
561,690,138
8,028
alesapin
Other tests are not related to changes.
2019-12-04T15:15:21
2019-12-04T15:15:21
{}
MEMBER
ClickHouse
ClickHouse
561,851,392
8,025
excitoon
So I guess I shall add a comment about that one need not only `clang` but also `lld` package.
2019-12-04T21:40:13
2019-12-04T21:40:13
{}
CONTRIBUTOR
ClickHouse
ClickHouse
560,951,646
5,773
alexey-milovidov
@abyss7 This PR cannot be finished before adding LLVM from contrib.
2019-12-03T01:09:08
2019-12-03T01:09:08
{}
MEMBER
ClickHouse
ClickHouse
560,967,347
7,386
alexey-milovidov
@akuzm ?
2019-12-03T02:16:15
2019-12-03T02:16:15
{}
MEMBER
ClickHouse
ClickHouse
561,142,107
7,795
CurtizJ
I've ran not very presice benchmark on 19.17 vs version with this patch and it showed about 10% performance degradation. Query with 100 `AND` operands: ``` select count() from numbers(10000000) where number != 96594 AND number != 18511 AND number != 98085 AND number != 84177 AND number != 70314 AND number !...
2019-12-03T12:17:05
2019-12-03T12:17:05
{}
MEMBER
ClickHouse
ClickHouse
561,156,161
7,998
alexey-milovidov
The binary itself does not require this libc version: ``` $ ldd -v /usr/bin/clickhouse linux-vdso.so.1 (0x00007ffda6934000) libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007fbdbd2c6000) librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007fbdbd2bc000) libdl.so.2 =...
2019-12-03T12:59:29
2019-12-03T12:59:29
{}
MEMBER
ClickHouse
ClickHouse
561,460,966
7,998
huang--yijian
> ldconfig -p |grep libc.so [root@storage10 ~]# ldconfig -p |grep libc.so libc.so.6 (libc6,x86-64, OS ABI: Linux 2.6.18) => /lib64/libc.so.6 ![image](https://user-images.githubusercontent.com/7484190/70110349-a30b9f80-1689-11ea-86cc-f9a6e1f67a22.png)
2019-12-04T03:31:52
2019-12-04T03:33:50
{}
NONE
ClickHouse
ClickHouse
561,622,757
7,207
barrkel
My use case for ClickHouse is as a cache alternative to MySQL for interactive sorts and filters. Expected row counts are in the low millions, and data size in single digit GB. CH being able to scale down to less memory is its primary selling case for me.
2019-12-04T12:28:47
2019-12-04T12:28:47
{}
NONE
ClickHouse
ClickHouse
561,722,328
3,929
alexey-milovidov
This PR should not be abandoned because it contains valuable feature.
2019-12-04T16:27:56
2019-12-04T16:27:56
{}
MEMBER
ClickHouse
ClickHouse
560,455,417
7,738
blinkov
also there's a file conflict
2019-12-02T15:51:57
2019-12-02T15:51:57
{}
CONTRIBUTOR
ClickHouse
ClickHouse
560,478,968
7,971
petrushenkovtb
ClickHouse client version 19.15.2.2 (official build). Connecting to localhost:9000 as user default. Connected to ClickHouse server version 19.15.2 revision 54426.
2019-12-02T16:45:42
2019-12-02T16:45:42
{}
NONE
ClickHouse
ClickHouse
560,970,122
7,120
alexey-milovidov
Is it abandoned?
2019-12-03T02:28:45
2019-12-03T02:28:45
{}
MEMBER
ClickHouse
ClickHouse
561,132,958
7,998
nicelulu
try ```ldconfig -p |grep libc.so``` cc.
2019-12-03T11:48:57
2019-12-03T11:48:57
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,156,866
7,050
alexey-milovidov
@jackpgao This is another issue, you should create a separate task for it (possibly with info for reproducing).
2019-12-03T13:01:29
2019-12-03T13:01:29
{}
MEMBER
ClickHouse
ClickHouse
561,219,277
8,004
amosbird
> Haven't provided an example/steps to reproduce A test case is very appreciated.
2019-12-03T15:29:40
2019-12-03T15:29:40
{}
COLLABORATOR
ClickHouse
ClickHouse
560,570,096
7,989
den-crane
400k rows per second? >column can be added before performance reduce up to 1000. Make a POC. Now it sounds like one CH shard is enough for good. >How is hard addind new node to the cluster ? Very. No re-sharding. Best option is too have a right number of shards from the beginning. If you don't bother ...
2019-12-02T20:37:08
2019-12-02T20:39:13
{}
CONTRIBUTOR
ClickHouse
ClickHouse
560,954,604
7,900
alexey-milovidov
Can we avoid duplicating metric names in snake_case? snake_case is just a guideline from Prometheus, not a strict requirement.
2019-12-03T01:21:09
2019-12-03T01:21:09
{}
MEMBER
ClickHouse
ClickHouse
560,965,010
7,835
alexey-milovidov
Yes, the cache cleanup actually helps.
2019-12-03T02:05:50
2019-12-03T02:05:50
{}
MEMBER
ClickHouse
ClickHouse
561,019,847
7,986
alxschwarz
Sorry, maybe I'm wrong... The problem is that at the moment of these errors it is impossible to get any other data from the database.
2019-12-03T06:21:25
2019-12-03T06:21:25
{}
NONE
ClickHouse
ClickHouse
560,531,764
7,859
filimonov
> Why information about moves stored in `system.merges` table? And why mutations are there? :) More seriously - there are more options. We are open to discuss the pros & cons. From a user perspective, we just want to know what the ClickHouse is doing in the background right now. Especially if it is a long-run...
2019-12-02T18:56:35
2019-12-03T08:26:42
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,098,569
8,000
javisantana
Some suggestions to workaround that. I think with the regular indexing features the performance is good enough, I wrote [this about geospatial processing with CH](https://carto.com/blog/geospatial-processing-with-clickhouse/) some time ago but you have some possibilities: - having `ORDER BY (lon, lat)` provides a ...
2019-12-03T10:17:15
2019-12-03T10:17:15
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,281,083
7,337
filimonov
That is a bug, it's a duplicate of https://github.com/ClickHouse/ClickHouse/issues/6731
2019-12-03T17:49:47
2019-12-03T17:49:47
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,352,726
7,944
filimonov
#8004
2019-12-03T20:55:10
2019-12-03T20:55:10
{}
CONTRIBUTOR
ClickHouse
ClickHouse
561,675,374
8,028
alesapin
`Functional stateful tests (memory) — fail:20, passed:50` O_o
2019-12-04T14:41:33
2019-12-04T14:41:33
{}
MEMBER
ClickHouse
ClickHouse
561,722,794
7,814
doom369
@den-crane thanks, but is hides only 1 field. So problem remains.
2019-12-04T16:28:55
2019-12-04T16:28:55
{}
NONE
ClickHouse
ClickHouse
566,602,857
8,247
den-crane
it's very unusual to put arrays into primary key. ``` create table XX (A Int64, `B.ID` Array(String), `C.ID` Array(String)) Engine=MergeTree order by (A, B.ID); insert into XX values(1, ['1'], ['1']); select * from XX array join C where C.ID='1' ┌─A─┬─B.ID──┬─C.ID─┐ │ 1 │ ['1'] │ 1 │ └───┴───────┴──────...
2019-12-17T15:54:22
2019-12-17T15:54:22
{ "+1": 1 }
CONTRIBUTOR