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
440,059,362
1,287
filimonov
https://github.com/yandex/ClickHouse/blob/6aa2313decf619faf62c04c7bc8a7d6913831140/dbms/src/Interpreters/DDLWorker.cpp#L73
2018-11-19T22:08:20
2018-11-19T22:08:20
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,412,664
3,621
filimonov
Looks like you can't specify socket in connection settings for MySQL storage, only for dicts. Fast workaround - connect using tcp (just use 127.0.0.1 instead of localhost). Probably socket syntax should be added, as work via unix socket is faster.
2018-11-20T20:08:13
2018-11-20T20:08:13
{ "+1": 2 }
CONTRIBUTOR
ClickHouse
ClickHouse
440,512,932
3,589
alexey-milovidov
Fixed in master.
2018-11-21T03:10:23
2018-11-21T03:10:23
{}
MEMBER
ClickHouse
ClickHouse
440,680,497
3,621
alexey-milovidov
In previous versions we have used `libmysqlclient` from Ubuntu packages (statically linked). It has `/var/run/mysqld/mysqld.sock` as default socket path. Probably it is Debian/Ubuntu specific. Then we have changed `libmysqlclient` from Ubuntu packages to `libmariadbclient` from sources for several reasons: - to ma...
2018-11-21T14:28:52
2018-11-21T14:30:42
{}
MEMBER
ClickHouse
ClickHouse
440,789,752
3,622
alexey-milovidov
Yes, it means that materialized view is not working correctly. Possible reason is that thread group is not initialized when processing insert into query_log.
2018-11-21T19:52:03
2018-11-21T19:52:03
{}
MEMBER
ClickHouse
ClickHouse
440,888,706
3,617
alexey-milovidov
https://clickhouse-builds.s3.yandex.net/3617/86f1a18185eb1a67e1c0d445352458838fa02804/report.html
2018-11-22T01:54:49
2018-11-22T01:54:49
{}
MEMBER
ClickHouse
ClickHouse
440,099,800
3,242
simPod
The response would be larger in data size and those % are implicit and can be calculated everywhere. I think that would be a step back as there's no really need for that.
2018-11-20T01:11:21
2018-11-20T01:11:21
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,200,582
3,589
pilosus
I've run into the same problem with arrays after updating ClickHouse from 18.10.3 to the latest stable. Queries with arrays that used to be performing well now crash with Memory limit (for query) exceeded. The exception seems to be thrown early, at the very beginning of execution
2018-11-20T09:19:10
2018-11-20T09:19:10
{}
NONE
ClickHouse
ClickHouse
440,244,672
3,587
filimonov
> My main goal is to have some good official lib aka driver and build libs for other langs around it. > Another option is try to write good TCP based client, but I haven't found any relevant docs about TCP protocol. I like the idea of having such a client. But AFAIK it's not a priority, there are a lot of server ta...
2018-11-20T11:43:49
2018-11-20T11:50:01
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,478,150
3,627
kmatt
Is dictGetT still a valid function? dictGetString seems to work in this case. The ip_trie docs mention dictGetT, but its not listed in https://clickhouse.yandex/docs/en/query_language/functions/ext_dict_functions/ DB::Exception: Unknown function dictGetT
2018-11-21T00:17:50
2018-11-21T00:17:50
{}
NONE
ClickHouse
ClickHouse
440,522,168
3,617
chenxing-xc
> Could you please add an integration test? > > If you have some troubles, you can instead write step-by-step instruction right here on how this feature can be tested (roll our HDFS in Docker, write files into it, and read them with ClickHouse). 1. This feature depends on HDFS cluster, I am not sure how to add re...
2018-11-21T04:03:53
2018-11-21T04:03:53
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,690,095
3,627
filimonov
For matching the most specific subnet there is also ip_trie dictionaries (handy for geoip or similar tasks).
2018-11-21T14:51:04
2018-11-21T14:51:18
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
440,806,480
3,617
alexey-milovidov
None of the builds succeed.
2018-11-21T20:58:59
2018-11-21T20:58:59
{}
MEMBER
ClickHouse
ClickHouse
440,101,635
3,598
zhang2014
How to choose username and password : ``` xml <remote_servers> <database_a_cluster> <shard> <replica> <host>192.168.0.2</host> <user>username_a</user> <password>password_a</password> </replica> </shard> </database...
2018-11-20T01:20:24
2018-11-20T01:20:24
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,241,386
3,607
Felixoid
I know how to do it properly. And did use everything which you suggested. This would be a good advise, BTW, but the problem definitely in XML by itself. As was described, you have to write ugly templates to generate them instead of just objects serializing like https://docs.saltstack.com/en/latest/ref/renderers/all/sal...
2018-11-20T11:30:55
2018-11-20T11:31:30
{}
MEMBER
ClickHouse
ClickHouse
440,303,084
3,587
SkeLLLa
@filimonov, thanks. And as far as I can see not all libs use POST :).
2018-11-20T15:00:08
2018-11-20T15:00:08
{}
NONE
ClickHouse
ClickHouse
440,366,285
3,613
alexey-milovidov
But we also have X-ClickHouse-User, X-ClickHouse-Key, X-ClickHouse-Quota. And I cannot change these headers, it will be more harmful than changing X-ClickHouse-Progress.
2018-11-20T17:43:18
2018-11-20T17:43:55
{}
MEMBER
ClickHouse
ClickHouse
440,367,349
3,613
alexey-milovidov
Adding a configuration option for HTTP headers prefix (`X-ClickHouse-`) is Ok, but looks like overkill, because we will have to keep the default with `X` and almost nobody will have any reason to change it.
2018-11-20T17:45:53
2018-11-20T17:46:13
{}
MEMBER
ClickHouse
ClickHouse
440,370,467
3,612
levonet
With original initscript: ```sh [root@reporting-dev1 init.d]# ./clickhouse-server start Start clickhouse-server service: Path to data directory in /etc/clickhouse-server/config.xml: /var/lib/clickhouse/ DONE [root@reporting-dev1 init.d]# ps axw | grep clickhouse 18665 ? Ssl 0:00 clickhouse-server --daem...
2018-11-20T17:53:30
2018-11-20T17:53:30
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,374,264
3,612
alexey-milovidov
Yes, we have confirmed that it is not terminated by default. We will try to find the reason.
2018-11-20T18:05:34
2018-11-20T18:05:34
{}
MEMBER
ClickHouse
ClickHouse
440,650,055
3,143
dqminh
@alexey-milovidov should we reopen this ? The metrics are still incorrect for us.
2018-11-21T12:44:56
2018-11-21T12:44:56
{}
CONTRIBUTOR
ClickHouse
ClickHouse
439,815,709
3,585
EvgenyVinogradov
18.12.17 is not crashing by SegFault All builds I assembling as RPMs by my self. OS - ContOS 7 Queries that leads SegFault are so easy and selecting max(RECORD_ID) value from default.ONLINE_LOG_local I guess problems are related to calculated fields. I'll try to upgrade to 18.14.13, drop calculated fields and relo...
2018-11-19T08:52:12
2018-11-19T08:52:12
{}
NONE
ClickHouse
ClickHouse
439,928,767
3,605
filimonov
@blinkov yep, when 'hot' data fits in memory - HDD is not an issue. But it you have a lot of data you really need to read some data from storage, and in that case read speed of different storages can vary very much. Costs of differents storages also differs very much. That feature will also make possible to create '...
2018-11-19T15:19:51
2018-11-19T20:52:36
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,172,131
3,605
Imaskar
Also a big benefit here would be an ability to ingest data more reliably. Superior NVMe performance would help with writes and then merged partitions could be placed on HDD. That's actually a way to implement the feature - when doing merge, place resulting partition on HDD if it is cold enough now.
2018-11-20T07:31:43
2018-11-20T07:32:05
{}
NONE
ClickHouse
ClickHouse
440,197,934
2,420
nullquery
toDate('2028-09-25') = 2028-09-25 toDate('2118-09-25') = 0000-00-00 Why? 😭
2018-11-20T09:10:42
2018-11-20T09:10:42
{}
NONE
ClickHouse
ClickHouse
440,393,505
3,617
alexey-milovidov
`INSERT INFILE` is very controversial feature in this PR: 1. It requires special support in client (to read data from local file), but it is not implemented. It is expected that I can send data from local file with clickhouse-client using this feature (analogy to `SELECT ... INTO OUTFILE`). 2. It requires special...
2018-11-20T19:05:19
2018-11-20T19:18:19
{}
MEMBER
ClickHouse
ClickHouse
440,435,843
3,044
Farfaday
found it in the doc here : https://clickhouse.yandex/docs/en/operations/settings/permissions_for_queries/#allow_ddl
2018-11-20T21:28:37
2018-11-20T21:28:37
{}
NONE
ClickHouse
ClickHouse
440,223,003
2,482
Aricliang
system.query_log
2018-11-20T10:28:11
2018-11-20T10:28:11
{}
NONE
ClickHouse
ClickHouse
440,236,782
2,420
blinkov
@nullquery it's beyond max date, see https://clickhouse.yandex/docs/en/data_types/date/
2018-11-20T11:13:06
2018-11-20T11:13:06
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,385,921
3,613
simPod
Yes, thats definitely overkill. I guess it's only this or that, not both. Decision is yours.
2018-11-20T18:41:54
2018-11-20T18:41:54
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,477,397
3,601
alexey-milovidov
Performance is degraded by 12% on URLs from Yandex.Metrica. ``` clickhouse-benchmark <<< "SELECT count() FROM test.hits WHERE NOT ignore(firstSignificantSubdomain(URL))" ``` From 0.137 sec. to 0.153 sec. That's non satisfactory.
2018-11-21T00:13:29
2018-11-21T00:13:29
{}
MEMBER
ClickHouse
ClickHouse
440,981,651
3,614
KochetovNicolai
Issue may be the same as in #3615 . Can you check if #3632 fixes it?
2018-11-22T10:17:27
2018-11-22T10:17:27
{}
MEMBER
ClickHouse
ClickHouse
440,301,577
3,587
filimonov
> query string length limit (that could actually vary depending on language used). You can use POST and send the query in POST body instead of querystring (i think most of HTTP based libs do it like that). For POST body there is no limitation on size, also in POST you can just send unencoded queries as plain tex...
2018-11-20T14:56:05
2018-11-20T14:56:05
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,314,140
3,613
simPod
Sure but we should follow best practices. Ofc it's a BC break tho. Can be taken care of with version IMO. _About the hacking, yea. I had to bend Guzzle php as well but that's not relevant :)_
2018-11-20T15:29:50
2018-11-20T15:29:50
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,524,301
3,334
alexey-milovidov
Unfinished.
2018-11-21T04:18:34
2018-11-21T04:18:34
{}
MEMBER
ClickHouse
ClickHouse
440,159,897
3,598
nicelulu
I am trying to complete the integration test.I executed: "pytest -k test_replace_partition". The following error occurred. >/python2.7/subprocess.py:542: CalledProcessError --------------------------------------------------------------------------------------------------------------------------- Captured stderr ...
2018-11-20T06:31:06
2018-11-21T07:27:58
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,579,155
3,630
blinkov
Have you actually put different values for `<replica>` to the `<macros>` section of your configs?
2018-11-21T08:32:19
2018-11-21T08:32:19
{}
CONTRIBUTOR
ClickHouse
ClickHouse
439,749,585
3,598
zhang2014
@alexey-milovidov Maybe use `context.getInterserverCredentials()` is better? But that may require some refactor
2018-11-19T02:00:34
2018-11-19T02:00:34
{}
CONTRIBUTOR
ClickHouse
ClickHouse
439,907,262
3,608
abyss7
The test `00281_compile_sizeof_packed` has failed - not relevant to this PR
2018-11-19T14:17:43
2018-11-19T14:17:43
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,236,491
3,605
neilACC
> The big question is what is the best way to split data: on partition level inside each node or on shard level in the cluster? I think, doing both is twice the effort. Better pick one. @Imaskar yep, understand your view, i have new details updated , welcome to participate in the discussion and sharing your ideas. :...
2018-11-20T11:11:59
2018-11-20T11:11:59
{}
NONE
ClickHouse
ClickHouse
440,352,726
3,613
alexey-milovidov
@filimonov Maybe we can also add a header with query_id. It will be useful if the user has not provided their own query_id with query.
2018-11-20T17:11:42
2018-11-20T17:11:42
{}
MEMBER
ClickHouse
ClickHouse
440,588,760
3,630
xiaods
@blinkov thanks for your hints. your hints let me found a bug for my k8s deployment bug. thanks a lot. it very implicit bug. https://github.com/xiaods/k8s-clickhouse-v2/tree/master/build_image
2018-11-21T09:06:40
2018-11-21T09:06:40
{}
NONE
ClickHouse
ClickHouse
440,706,601
3,621
edonin
Thanks for the explanation Alexey. I will use 127.0.0.1 until it will be configurable.
2018-11-21T15:31:56
2018-11-21T15:31:56
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,170,529
3,605
neilACC
> @liuqian1989 it's kind of obvious possible feature, but in practice ClickHouse works fast enough even on 100% HDD given there's enough memory for page cache, so it's priority is not that high. > > I'd suggest you to try to put all those petabytes on HDD and see how it goes. Or you have already tried and are not sa...
2018-11-20T07:24:23
2018-11-20T07:34:07
{ "+1": 11 }
NONE
ClickHouse
ClickHouse
440,271,058
1,118
blinkov
@BlackEric001 nope, not scheduled yet. Though you can already use something like this to get some idea on how query is executed: ``` % clickhouse-client --send_logs_level=debug <<< 'select max(number) from numbers(100) where number % 2 union all select 123 as number;' > /dev/null [....yandex.net] 2018.11.20 16:18:29...
2018-11-20T13:20:58
2018-11-20T13:20:58
{ "hooray": 3 }
CONTRIBUTOR
ClickHouse
ClickHouse
440,519,706
3,617
alexey-milovidov
> BTW, this feature is heavily used in my team because our ETL task is scheduled on a platform without ClickHouse, and server side data access is needed. We should think how we can made this feature more consistent.
2018-11-21T03:45:24
2018-11-21T03:45:24
{}
MEMBER
ClickHouse
ClickHouse
440,581,033
2,878
blinkov
@AlexFridman похоже на это смотрел @4ertus2, но он сейчас в отпуске, разобрался ли не знаю
2018-11-21T08:39:25
2018-11-21T08:39:25
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,485,454
3,601
alexey-milovidov
Performance has been improved up to 0.116 sec by https://github.com/yandex/ClickHouse/commit/266259486551bf95c970fbafb48bc9573387c5ef
2018-11-21T00:48:47
2018-11-21T00:48:47
{}
MEMBER
ClickHouse
ClickHouse
440,614,384
3,592
javisantana
I understand this only works if the inserts are done through the Distributed table right?
2018-11-21T10:29:34
2018-11-21T10:29:34
{}
CONTRIBUTOR
ClickHouse
ClickHouse
439,817,828
3,605
blinkov
@liuqian1989 it's kind of obvious possible feature, but in practice ClickHouse works fast enough even on 100% HDD given there's enough memory for page cache, so it's priority is not that high. I'd suggest you to try to put all those petabytes on HDD and see how it goes. Or you have already tried and are not satisfie...
2018-11-19T08:59:33
2018-11-19T08:59:33
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,171,379
3,606
chenxing-xc
> @chenxing-xc , and what is physical format of data, stored in HDFS? Avro/ORC/Parquet, etc. The format which clickhouse support is ok. e.g. JSON, CSV ....
2018-11-20T07:28:27
2018-11-20T07:28:27
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,379,240
3,619
alexey-milovidov
Why it was redundant?
2018-11-20T18:20:35
2018-11-20T18:20:35
{}
MEMBER
ClickHouse
ClickHouse
440,397,392
3,617
alexey-milovidov
We can proceed first without `INSERT INFILE` feature.
2018-11-20T19:17:37
2018-11-20T19:17:37
{}
MEMBER
ClickHouse
ClickHouse
440,577,148
2,601
alex-zaitsev
I checked the trace, everything is the same, the only difference is in query pipeline that has 'Filter' stage in case with 'where': where, slow: ``` <Debug> executeQuery: Query pipeline: Expression Expression ParallelAggregating Expression × 4 Filter MergeTreeThread ``` prewhere, fast: ```...
2018-11-21T08:25:06
2018-11-21T08:25:06
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,588,932
3,630
xiaods
let me try to fix it and check the clickhouse's behavior
2018-11-21T09:07:19
2018-11-21T09:07:19
{}
NONE
ClickHouse
ClickHouse
440,749,501
3,623
alexey-milovidov
> but i think we should also fix the actual cause that causes thread_group to be null? Yes. This PR doesn't fix the root cause.
2018-11-21T17:31:37
2018-11-21T17:31:37
{}
MEMBER
ClickHouse
ClickHouse
441,006,858
1,793
xiaods
@tobad357 i am ok for distributed table. https://github.com/xiaods/k8s-clickhouse-v2.git
2018-11-22T11:52:29
2018-11-22T11:52:29
{}
NONE
ClickHouse
ClickHouse
440,279,875
3,587
SkeLLLa
@filimonov thanks for reply. As for HTTP it's ok, but I find it not very good, because of it's limitations: - query string length limit (that could actually vary depending on language used). - query string parameters encoding could be potential issue that is hard to predict. That why I was also looking to TCP pro...
2018-11-20T13:50:40
2018-11-20T13:50:40
{}
NONE
ClickHouse
ClickHouse
440,392,155
3,619
amosbird
CurrentThread::QueryScope handles that already
2018-11-20T19:01:18
2018-11-20T19:01:18
{}
COLLABORATOR
ClickHouse
ClickHouse
440,399,982
3,617
alexey-milovidov
LICENSE file should be located in the https://github.com/chenxing-xc/ClickHouse-Extras-libhdfs3 root.
2018-11-20T19:26:03
2018-11-20T19:26:03
{}
MEMBER
ClickHouse
ClickHouse
440,512,069
3,589
alexey-milovidov
The bug was introduced in f56d1676, #2830
2018-11-21T03:05:38
2018-11-21T03:05:38
{}
MEMBER
ClickHouse
ClickHouse
440,577,742
2,878
AlexFridman
Привет @blinkov! Подскажи, пожалуйста, что нам делать. Сам в это дело вчера уперся :(
2018-11-21T08:27:22
2018-11-21T08:27:22
{}
NONE
ClickHouse
ClickHouse
440,669,832
3,627
den-crane
>Is dictGetT still a valid function? No, T means Type `dictGet<T>` Replace with a type you need. String/Int64/UInt....
2018-11-21T13:55:28
2018-11-21T13:59:27
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,886,925
3,617
chenxing-xc
> None of the builds succeed. What 's the build error? there are some extra dependency mentioned in submodule libhdfs3 which need to be installed( by apt-get). e.g. libxml2, krb5, libuuid, libgsasl
2018-11-22T01:42:37
2018-11-22T01:42:37
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,211,920
3,589
Sahaquielxo
Confirmed
2018-11-20T09:53:18
2018-11-20T09:53:18
{}
NONE
ClickHouse
ClickHouse
440,237,379
3,606
blinkov
@chenxing-xc please clean up the diff, it is indeed unreadable.
2018-11-20T11:15:24
2018-11-20T11:15:24
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,400,511
3,617
alexey-milovidov
I have added you to the ClickHouse-Extras organization. You can create/move any repositories you want there.
2018-11-20T19:27:50
2018-11-20T19:27:50
{}
MEMBER
ClickHouse
ClickHouse
440,452,341
3,615
champtar
I have a "small" repro case, with 3 servers (actually 2 are enough I think) create a cluster: ``` <yandex> <remote_servers> <test_cluster> <shard> <weight>1</weight> <replica> <default_database>test</default_database> <host>server1</host> <port>9000</port> <user>dis...
2018-11-20T22:29:25
2018-11-20T22:29:25
{ "+1": 1 }
CONTRIBUTOR
ClickHouse
ClickHouse
440,524,927
3,629
alexey-milovidov
It is available starting from 18.14.9: https://github.com/yandex/ClickHouse/blob/master/CHANGELOG.md#clickhouse-release-18149-2018-10-16
2018-11-21T04:23:43
2018-11-21T04:23:43
{}
MEMBER
ClickHouse
ClickHouse
439,730,962
3,594
alexey-milovidov
> I don't know what is better default for per endpoint connection pools? I also don't know. But I think that 4 parallel transfers should be enough to utilize 10 GBit network even with moderate packet loss.
2018-11-18T22:17:53
2018-11-18T22:17:53
{}
MEMBER
ClickHouse
ClickHouse
439,760,785
3,256
champtar
Fixed in master ``` ┌────a─┬─count()─┐ │ 6019 │ 1 │ └──────┴─────────┘ 1 rows in set. Elapsed: 0.969 sec. Processed 1.00 million rows, 12.00 MB (1.03 million rows/s., 12.39 MB/s.) ``` Thanks @KochetovNicolai !
2018-11-19T03:25:33
2018-11-19T03:25:33
{}
CONTRIBUTOR
ClickHouse
ClickHouse
439,969,580
3,609
s-mx
Forgot add getters in StorateReplicatedMergeTree.
2018-11-19T17:11:13
2018-11-19T17:11:13
{}
CONTRIBUTOR
ClickHouse
ClickHouse
439,975,061
3,580
abyss7
The test `00411_long_accurate_number_comparison` has failed. Looks like it's flaky.
2018-11-19T17:27:24
2018-11-19T17:27:24
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,201,882
3,606
amosbird
This doesn't look like a reviewable PR. There are tons of unrelated conflicts hanging around. I'd suggest a full rebase at least (soft reset to origin/master).
2018-11-20T09:23:18
2018-11-20T09:23:18
{}
COLLABORATOR
ClickHouse
ClickHouse
440,247,548
3,606
blinkov
#3617 looks way better :)
2018-11-20T11:55:21
2018-11-20T11:55:21
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,311,755
3,613
filimonov
Avoiding X- prefixes is just a 'best practice' / 'recommendation', not a hard rule. Changing is a bad idea, as will broke backward compatibility. But it's easy to make it configurable. Generally those header are kinda controversial : streaming headers is not something normal for HTTP exchange, and usually you ne...
2018-11-20T15:23:48
2018-11-20T15:23:48
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,577,261
3,598
alesapin
@nicelulu Error `.IOError: [Errno 2] No such file or directory: ... docker_compose.yml` often is consequence of some other errors, which occurred during preparation of compose file (it's generated every time). Can you run same pytest command, but with `-ss` option and provide full output? I guess that error caused ...
2018-11-21T08:25:36
2018-11-21T08:25:36
{}
MEMBER
ClickHouse
ClickHouse
440,890,714
3,617
chenxing-xc
> https://clickhouse-builds.s3.yandex.net/3617/86f1a18185eb1a67e1c0d445352458838fa02804/report.html I saw the following error in cmake output, "CMake Error at /usr/share/cmake-3.10/Modules/FindPackageHandleStandardArgs.cmake:137 (message): Could NOT find CURL (missing: CURL_LIBRARY CURL_INCLUDE_DIR) Call Stack (...
2018-11-22T02:07:54
2018-11-22T02:07:54
{}
CONTRIBUTOR
ClickHouse
ClickHouse
439,803,365
3,605
Imaskar
The big question is what is the best way to split data: on partition level inside each node or on shard level in the cluster? I think, doing both is twice the effort. Better pick one.
2018-11-19T08:03:05
2018-11-19T08:03:05
{}
NONE
ClickHouse
ClickHouse
439,930,285
3,598
alexey-milovidov
@zhang2014 InterserverCredentials are used for HTTP(s) request made by replicas on replication port. It won't work for queries.
2018-11-19T15:24:05
2018-11-19T15:24:05
{}
MEMBER
ClickHouse
ClickHouse
440,247,574
1,118
jackpgao
Looking forward.
2018-11-20T11:55:26
2018-11-20T11:55:26
{ "+1": 1 }
NONE
ClickHouse
ClickHouse
440,280,749
3,612
alexey-milovidov
We cannot find the reason why the child process is not terminated with parent. Could you please provide more details: a process tree before and after terminating clickhouse-server and details about what is `sh` in your system.
2018-11-20T13:53:28
2018-11-20T13:53:28
{}
MEMBER
ClickHouse
ClickHouse
440,354,051
3,613
alexey-milovidov
@simPod I see that naming headers with `X-` prefix is common practice. Example: S3 API https://docs.aws.amazon.com/AmazonS3/latest/API/RESTCommonRequestHeaders.html Maybe we simply should not make this change right now.
2018-11-20T17:14:28
2018-11-20T17:14:46
{}
MEMBER
ClickHouse
ClickHouse
440,518,966
3,617
alexey-milovidov
Could you please add an integration test? If you have some troubles, you can instead write step-by-step instruction right here on how this feature can be tested (roll our HDFS in Docker, write files into it, and read them with ClickHouse).
2018-11-21T03:39:57
2018-11-21T03:40:07
{}
MEMBER
ClickHouse
ClickHouse
440,609,037
3,592
kanterov
@alexey-milovidov I checked why tests are failing, it seems it didn't pick up changes in server-test.xml
2018-11-21T10:13:48
2018-11-21T10:13:48
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,678,601
3,623
dqminh
The guard helps the exception in #3622, but i think we should also fix the actual cause that causes thread_group to be null ? Or was it already known ?
2018-11-21T14:23:24
2018-11-21T14:23:24
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,695,771
3,630
alexey-milovidov
You can check the effective values by `SELECT * FROM system.macros`.
2018-11-21T15:04:30
2018-11-21T15:04:30
{ "+1": 1 }
MEMBER
ClickHouse
ClickHouse
439,729,921
3,599
alexey-milovidov
Missing integration test.
2018-11-18T22:04:36
2018-11-18T22:04:36
{}
MEMBER
ClickHouse
ClickHouse
439,730,330
3,598
alexey-milovidov
Missing integration test.
2018-11-18T22:09:52
2018-11-18T22:09:52
{}
MEMBER
ClickHouse
ClickHouse
439,899,920
3,585
filimonov
Can you also try Altinity builds? https://packagecloud.io/Altinity/clickhouse (build scripts are there https://github.com/Altinity/clickhouse-rpm )
2018-11-19T13:53:51
2018-11-19T13:53:51
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,375,183
3,612
levonet
Ok, I remove this code.
2018-11-20T18:08:28
2018-11-20T18:08:28
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,403,715
3,619
alexey-milovidov
Ok, thank you!
2018-11-20T19:37:47
2018-11-20T19:37:47
{}
MEMBER
ClickHouse
ClickHouse
440,498,249
3,617
chenxing-xc
> We can proceed first without `INSERT INFILE` feature. Ok, I can withdraw this feature, and put new submodules into ClickHouse-Extras. The original goal of "INFILE' feature is to simulate COPY syntax in PG. BTW, this feature is heavily used in my team because our ETL task is scheduled on a platform without ClickHo...
2018-11-21T01:48:45
2018-11-21T01:48:45
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,523,018
3,617
zhang2014
@chenxing-xc Maybe: 1. INSERT INTO TABLE FUNCTION hdfs('url', format, column-description) 2. SELECT * FROM hdfs('url', format, column-description) But this needs implement `write` method in StorageHDFS.
2018-11-21T04:08:43
2018-11-21T04:08:43
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,170,701
3,606
alex-krash
@chenxing-xc , and what is physical format of data, stored in HDFS? Avro/ORC/Parquet, etc.
2018-11-20T07:25:17
2018-11-20T07:25:17
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,244,973
3,606
chenxing-xc
I rebase master and create a new PR.
2018-11-20T11:45:05
2018-11-20T11:45:05
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,358,703
3,613
simPod
@alexey-milovidov it's rather common mispractice everyone copies wrong ;) Every argument I can further provide is in [IETF RFC 6648](https://tools.ietf.org/html/rfc6648) that is very brief but also clear. My point here is that the API design can be improved and made RFC-copliant and it's easier to do it in earlier s...
2018-11-20T17:25:44
2018-11-20T17:25:44
{}
CONTRIBUTOR
ClickHouse
ClickHouse
440,365,411
3,613
alexey-milovidov
--We can change it without any compatibility configuration settings. I think, this is Ok in this case.--
2018-11-20T17:41:19
2018-11-20T17:43:27
{}
MEMBER
ClickHouse
ClickHouse
440,461,416
3,627
den-crane
How many rows in `ranges` ? Check ip_trie dict https://clickhouse.yandex/docs/en/query_language/dicts/external_dicts_dict_layout/#ip_trie Your query will work much faster SELECT d.*, dictGetT('dict_name', 'id', tuple(address)) as id FROM data AS d
2018-11-20T23:07:13
2018-11-20T23:07:13
{ "+1": 2 }
CONTRIBUTOR
ClickHouse
ClickHouse
440,622,913
3,592
kanterov
@javisantana yes, it assumes that somehow data is distributed according to the sharding key
2018-11-21T10:58:08
2018-11-21T10:58:08
{}
CONTRIBUTOR