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 | 689,034,774 | 14,566 | ritaank | I mean I guess I find it strange as well, but I am just going off of the privilege hierarchy as listed [here](https://clickhouse.tech/docs/en/sql-reference/statements/grant/#grant-privileges)
`TTL` and `... | 2020-09-08T17:44:13 | 2020-09-08T17:44:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,126,838 | 14,045 | radiatorius | `SELECT * FROM test_table` query returns such result (text trimmed):
```
361a8df7-f6e6-4fc2-9708-9bbf420f0a81 1 2 N2DB19BufferWithOwnMemoryINS_11WriteBufferEEE\0\0\0??r\0\0\0\0 ?u\0\0\0\0??u\0\0\0\0\0\0\0\0\0\0\0\0N2DB23WriteBufferFromFileBaseE\0\0??r\0\0\0\0p?u\0\0\0\0P?u\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\... | 2020-09-08T20:46:17 | 2020-09-08T20:46:17 | {} | NONE |
ClickHouse | ClickHouse | 689,133,593 | 14,547 | filimonov | Scripts defining sequence of building docker images & passing envs and other parameters.
Also some html report building scripts.
- will you share those? | 2020-09-08T21:00:11 | 2020-09-08T21:00:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,482,282 | 14,544 | alexey-milovidov | Not quite fixed. | 2020-09-07T19:22:34 | 2020-09-07T19:22:34 | {} | MEMBER |
ClickHouse | ClickHouse | 688,538,989 | 14,562 | CurtizJ | Probably there are some places in code with the same mistake. | 2020-09-07T23:25:04 | 2020-09-07T23:25:04 | {} | MEMBER |
ClickHouse | ClickHouse | 688,558,206 | 14,251 | bohutang | You can try the new build from:
https://clickhouse-builds.s3.yandex.net/14523/576c6e97570cd3aa49ae69c06b9e7a4b930f52c4/clickhouse_build_check/report.html
| 2020-09-08T00:48:56 | 2020-09-08T00:49:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,574,062 | 14,251 | bohutang | > @BohuTANG , do you know why there is no rpm package in your URL?
rpm needs wait for https://repo.clickhouse.tech/rpm/testing/x86_64/ daily updates | 2020-09-08T01:55:38 | 2020-09-08T01:55:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,775,463 | 14,575 | alesapin | Build finished, oom is not related. | 2020-09-08T10:25:06 | 2020-09-08T10:25:06 | {} | MEMBER |
ClickHouse | ClickHouse | 689,177,019 | 14,538 | den-crane | Do you have a `data sample / create table` to repro the issue ? | 2020-09-08T22:50:17 | 2020-09-08T22:50:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,183,808 | 14,547 | alexey-milovidov | Yes, we must share these scripts https://github.com/ClickHouse/ClickHouse/issues/10825 | 2020-09-08T23:12:07 | 2020-09-08T23:12:07 | {} | MEMBER |
ClickHouse | ClickHouse | 688,480,873 | 14,519 | alexey-milovidov | > but would it work for the binary? I'm not sure, couldn' find any reference for this.
Me neither. Probably not. But we can use the trick with copy. | 2020-09-07T19:17:11 | 2020-09-07T19:17:11 | {} | MEMBER |
ClickHouse | ClickHouse | 688,844,770 | 14,565 | filimonov | > Or any other way to export to redis?
check that (not very smart, but straight-forward).
```
SELECT arrayStringConcat(arrayMap(x -> concat('BITSET rediskey ', toString(x), ' 1'), bitmapToArray(bitmapBuild([1, 2, 3, 4, 5, 1000])) AS res), '\n')
FORMAT TSVRaw
BITSET rediskey 1 1
BITSET rediskey 2 1
BITSET re... | 2020-09-08T12:50:56 | 2020-09-08T12:51:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,112,263 | 14,546 | simPod | Wow, works. Thank you. I'll try to backtrack what confused me in docs and eventually send PR with improvement. I'll close this after. | 2020-09-08T20:16:38 | 2020-09-08T20:16:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,563,553 | 14,517 | bharatnc | @filimonov for your information, since you added the `comp-zookeeper` label :) - might be useful to have your suggestions. | 2020-09-08T01:11:32 | 2020-09-08T01:11:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,573,102 | 14,251 | liangli1213 | @BohuTANG , do you know why there is no rpm package in your URL? | 2020-09-08T01:51:02 | 2020-09-08T01:51:02 | {} | NONE |
ClickHouse | ClickHouse | 688,604,448 | 13,961 | yancheng44 | @azat
Thank you for your reply.
It already help improve the performance a lot.
Is there anyway to let the IN works like Global in in a single server Or let the Global in not combine temp table.
I have repartition the user to different server, I just want the query merge result at the last step.
Tha... | 2020-09-08T03:59:22 | 2020-09-08T03:59:22 | {} | NONE |
ClickHouse | ClickHouse | 688,840,451 | 14,490 | alesapin | No performance changes in PR | 2020-09-08T12:42:16 | 2020-09-08T12:42:16 | {} | MEMBER |
ClickHouse | ClickHouse | 688,993,861 | 14,583 | nikitamikhaylov | OOM in builds are unrelated to PR. | 2020-09-08T16:29:07 | 2020-09-08T16:29:07 | {} | MEMBER |
ClickHouse | ClickHouse | 688,995,725 | 2,605 | Matt-Texier | Support of IPv4 and IPv6 types has been added and it works fine.
See this snippet of an example of model using those types:
```sql
CREATE TABLE telemetry.enriched_flows
(
full_time_stamp DateTime,
bin_cycle_sec UInt16,
as_src UInt32,
as_dst UInt32,
comms String,
as_path String... | 2020-09-08T16:32:31 | 2020-09-08T16:32:31 | {} | NONE |
ClickHouse | ClickHouse | 689,104,135 | 14,513 | filimonov | > Hm, afair i've used intDiv in `partition by` keys and it was working properly...
>
> And i've was also looking for the info about binary ops monotonicity, and didn't find it in code...
>
> Loosely coupled with #11796
Rechecked. It seems it somehow works for PARTITION key, but doesn't work for primary key - f... | 2020-09-08T19:58:45 | 2020-09-08T19:58:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,111,256 | 13,668 | simPod | Thank you | 2020-09-08T20:14:21 | 2020-09-08T20:14:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,136,206 | 14,574 | filimonov | ClickHouse distibution has a file to adjust the limits:
https://github.com/ClickHouse/ClickHouse/blob/master/debian/clickhouse.limits
Docker readme file also show how to increate ulimit for number of files.
So most probably you installed clickhouse in some non-standard way. | 2020-09-08T21:05:44 | 2020-09-08T21:05:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,033,080 | 14,494 | CurtizJ | Ok. I think it will be useful to implement separate pool with customizable priority later. | 2020-09-08T17:40:50 | 2020-09-08T17:40:50 | {} | MEMBER |
ClickHouse | ClickHouse | 689,061,747 | 14,586 | alexey-milovidov | One build has failed with OOM - ignore. | 2020-09-08T18:37:26 | 2020-09-08T18:37:26 | {} | MEMBER |
ClickHouse | ClickHouse | 688,543,573 | 13,961 | azat | @yancheng44
Here there are the problems that affect the query execution (if I'm not missing anything):
- subquery is the heaviest part, and it will be execute before the main query, as a workaround for now you can use `prefer_localhost_replica=0`, then query on the local node will be execute as a remote query and ... | 2020-09-07T23:44:55 | 2020-09-07T23:44:55 | {} | MEMBER |
ClickHouse | ClickHouse | 688,598,499 | 4,060 | YuYuYan | JDBC occurs the same leak, how do fix it | 2020-09-08T03:34:19 | 2020-09-08T03:34:19 | {} | NONE |
ClickHouse | ClickHouse | 688,910,583 | 14,584 | den-crane | This is the expected behavior by design. | 2020-09-08T14:22:11 | 2020-09-08T14:22:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,912,511 | 14,576 | den-crane | Parts in these tables (old and new formats) are incompatible.
The only way is `insert into new select from old`
A table with the new format has another part naming and require minmax_partition_column.idx files and so on. | 2020-09-08T14:24:53 | 2020-09-08T14:27:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,916,873 | 14,574 | den-crane | This version 18.14.8 is unsupported. Try to upgrade to the stable (20.8, 20.7, 20.6) or LTS (20.3) versions. | 2020-09-08T14:31:17 | 2020-09-08T14:39:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,971,796 | 14,508 | blinkov | converstion => conversion | 2020-09-08T15:54:13 | 2020-09-08T15:54:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,003,568 | 2,605 | simPod | @alex-zaitsev ipv4 and ipv6 are great. Now I'm looking for a way to store '0.0.0.0/0' properly. I.e. for Inet type. | 2020-09-08T16:45:29 | 2020-09-08T16:45:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,481,757 | 14,201 | azat | First of all https://github.com/ClickHouse/ClickHouse/issues/14201#issuecomment-683350463 was just a work around, I'm not insisting, but
>but system.text_log has no save the msg body
There is `message` column.
>and the log will more and more
You can use materialize view to duplicate the into separate table.... | 2020-09-07T19:20:32 | 2020-09-07T19:20:32 | {} | MEMBER |
ClickHouse | ClickHouse | 688,490,096 | 14,557 | nikitamikhaylov | > It is OOM while building - completely unrelated to any details of code formatting.
Thanks! | 2020-09-07T19:48:59 | 2020-09-07T19:48:59 | {} | MEMBER |
ClickHouse | ClickHouse | 688,741,197 | 1,981 | kreuzerkrieg | Odd. I have the same error after the config was changed and log level node was left empty. `syslog` has a correct message:
> Sep 8 09:11:00 ip-172-16-8-5 clickhouse-server[3774]: Poco::Exception. Code: 1000, e.code() = 0, e.displayText() = Invalid argument: Not a valid log level, Stack trace (when copying this mess... | 2020-09-08T09:22:54 | 2020-09-08T09:22:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,931,862 | 14,538 | simPod | EDIT: maybe it does, I have ongoing alter query running 10h+. Perhaps it will succeed after it finishes.
-------------
Thank you for the answer. What should `materialize ttl` do? When I run it on the target table nothing happens. Partitions without TTL info stays like that.
v20.6.3.28
When run with ON CLUST... | 2020-09-08T14:53:23 | 2020-09-08T15:07:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,161,586 | 13,961 | azat | >Is there anyway to let the IN works like Global in in a single server
AFAICS right now it is not possible (plus the distributed query will be overwritten anyway).
> Or let the Global in not combine temp table.
I don't understand this (I think that this is because I did not make it clear about the underlying p... | 2020-09-08T22:05:52 | 2020-09-08T22:05:52 | {} | MEMBER |
ClickHouse | ClickHouse | 689,191,352 | 14,532 | CurtizJ | `test_adaptive_granularity/test.py::test_version_update_two_nodes` is known flaky test.
| 2020-09-08T23:37:58 | 2020-09-08T23:37:58 | {} | MEMBER |
ClickHouse | ClickHouse | 688,625,534 | 14,571 | akuzm | The stack trace of a child query failing with `too many parts` looks like this:
```
stack_trace: 0. /build/obj-x86_64-linux-gnu/../contrib/poco/Foundation/src/Exception.cpp:27: Poco::Exception::Exception(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, int) @ 0x11... | 2020-09-08T05:25:11 | 2020-09-08T05:25:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,633,993 | 10,647 | RO-29 | I am facing the same issue, the Kafka engine consumer is just trying to parse the faulty message in the loop despite me having configured `kafka_skip_broken_messages`
It's not processing new messages at all
```
ENGINE = Kafka SETTINGS kafka_broker_list = 'localhost:9092',
kafka_topic_list = 'events',
kafka_gr... | 2020-09-08T05:51:23 | 2020-09-08T05:51:23 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 688,673,276 | 14,572 | chang0391 | below is error.log:
2020.09.08 15:17:34.476418 [ 31922 ] {d8d30b12-f09c-4062-acf5-a0908ed9cef7} <Error> executeQuery: Code: 99, e.displayText() = DB::Exception: Packet payload is not fully read. Stopped after 4518 bytes, while 3678 bytes are in buffer. (version 20.8.1.4513 (official build)) (from [::1]:34988) (in quer... | 2020-09-08T07:22:06 | 2020-09-08T07:22:06 | {} | NONE |
ClickHouse | ClickHouse | 688,683,270 | 14,540 | KochetovNicolai | Test was probably fixed in #14055 | 2020-09-08T07:41:17 | 2020-09-08T07:41:17 | {} | MEMBER |
ClickHouse | ClickHouse | 688,832,527 | 14,583 | alexey-milovidov | LGTM.
Two points:
- we don't support `clang-8` already for a long time and we don't need any fixes;
- the target is not build in "Arcadia" (that is typical pain point due to an old compiler).
But this change is still worth doing. | 2020-09-08T12:25:53 | 2020-09-08T12:25:53 | {} | MEMBER |
ClickHouse | ClickHouse | 688,849,145 | 14,568 | filimonov | > My problems and #13619 Almost the same,system.replication_ There is a lot of data in the queue,I don't know what to do
That sounds completely different. If you can't find the reason / you think it's a bug - please open separate issue. | 2020-09-08T12:59:25 | 2020-09-08T12:59:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,125,506 | 9,375 | alexey-milovidov | @jason-godden
Congrats, @Avogar has implemented this feature! | 2020-09-08T20:43:41 | 2020-09-08T20:43:41 | {} | MEMBER |
ClickHouse | ClickHouse | 689,130,066 | 14,562 | alexey-milovidov | > test_server_initialization/test.py::test_live_view_dependency
Broken integration test @vzakaznikov | 2020-09-08T20:52:33 | 2020-09-08T20:52:33 | {} | MEMBER |
ClickHouse | ClickHouse | 689,187,720 | 14,576 | den-crane | btw, this util works only against not replicated tables and nobody tested it for couple years. | 2020-09-08T23:25:47 | 2020-09-08T23:25:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,508,037 | 14,555 | 4ertus2 | It's strange but `FunctionBinaryArithmetic::castType` does not work with some types anymore. I.e DataTypeInterval or Date - UInt8.
Update: it's cause condition is false for `PlusImpl` when it is wrapped into nemeless namespace
```
std::is_same_v<Op<UInt8, UInt8>, PlusImpl<UInt8, UInt8>>;
``` | 2020-09-07T20:54:08 | 2020-09-07T22:41:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,126,042 | 14,045 | filimonov | > Here is an example that really causes **segmentation fault** everytime:
I've tried it in several scenarios (concurrency / fake load etc). It never crashed the server.
> The issue does not depend on the table scheme. It also repeats for other tables and other **update** mutations.
Currently, hundreds (maybe... | 2020-09-08T20:44:45 | 2020-09-08T20:50:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,141,528 | 14,045 | filimonov | It would be great to exclude some HW/VM failures / or have reproducible example.
BTW - can you share your clickhouse configuration? (/etc/clickhouse-server - you can mask passwords and other secrets) Any non-standard OS settings? | 2020-09-08T21:18:10 | 2020-09-08T21:20:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,851,713 | 14,569 | filimonov | What is $(interface) in your query?
Can you create isolated example?
P.S. 19.4 is not supported anymore | 2020-09-08T13:04:06 | 2020-09-08T21:34:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,189,584 | 3,289 | georgiopandarez | I ran into this today.
JSONEachRow format with input_format_import_nested_json won't JSON looking like {"a": {"b": 123 } which makes it kinda useless with a Nested schema. | 2020-09-08T23:32:37 | 2020-09-08T23:32:37 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 688,526,827 | 14,549 | alexey-milovidov | AST fuzzer: https://github.com/ClickHouse/ClickHouse/issues/14560 | 2020-09-07T22:20:40 | 2020-09-07T22:20:40 | {} | MEMBER |
ClickHouse | ClickHouse | 688,605,655 | 14,568 | yd1261137568 | My problems and https://github.com/ClickHouse/ClickHouse/issues/13619 Almost the same,system.replication_ There is a lot of data in the queue,I don't know what to do | 2020-09-08T04:04:28 | 2020-09-08T04:04:28 | {} | NONE |
ClickHouse | ClickHouse | 688,810,911 | 14,554 | qoega | As I understand it is no more needed. | 2020-09-08T11:38:15 | 2020-09-08T11:38:15 | {} | MEMBER |
ClickHouse | ClickHouse | 688,812,041 | 14,547 | qoega | Does this PR test this file or it will be done separately? | 2020-09-08T11:40:34 | 2020-09-08T11:40:34 | {} | MEMBER |
ClickHouse | ClickHouse | 688,813,543 | 14,553 | qoega | Probably unrelated
test_distributed_over_live_view/test.py::TestLiveViewOverDistributedSuite::test_distributed_over_live_view_order_by_key[lv_over_distributed_table-node0]
| 2020-09-08T11:44:09 | 2020-09-08T11:44:35 | {} | MEMBER |
ClickHouse | ClickHouse | 688,846,956 | 14,566 | filimonov | " even though they are listed as distinct privileges"
Ah, that's strange actually. | 2020-09-08T12:55:18 | 2020-09-08T12:55:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,850,333 | 14,568 | filimonov | > system.processes The table shows a large number of drop partiton jams,They can't be killed.Still stuck after restart
Maybe they are waiting for somethin (check system.mutations & system.merges).
If you want force removeing of those - you need to stop the node, clean the /clickhouse/task_queue in zookeeper, and... | 2020-09-08T13:01:36 | 2020-09-08T13:01:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,128,646 | 14,045 | filimonov | That segfault + trash in data + Cannot mmap 16.00 EiB - looks like the same issue, so that doesn't help too much.
The questions above are still relevant. | 2020-09-08T20:49:40 | 2020-09-08T20:49:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,815,082 | 14,547 | alesapin | > Does this PR test this file or it will be done separately?
Yes, it checks this file, I'm waiting all checks to be started. | 2020-09-08T11:47:49 | 2020-09-08T11:47:49 | {} | MEMBER |
ClickHouse | ClickHouse | 688,886,040 | 14,195 | Slach | Hello @bobrik, you are work on another task ClickHouse as storage backend for jaeger
https://github.com/jaegertracing/jaeger/issues/1438
could you help us review these changes for add support ClickHouse OpenTelemetry client->Jaeger | 2020-09-08T13:46:56 | 2020-09-08T13:46:56 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,979,512 | 2,605 | simPod | @blinkov @alex-zaitsev Hi, any way this could get reevaluated? | 2020-09-08T16:05:59 | 2020-09-08T16:05:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,102,449 | 14,591 | myrrc | Related: #14590 | 2020-09-08T19:54:59 | 2020-09-08T19:54:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,122,123 | 14,513 | alexey-milovidov | @filimonov Partition elimination does not depend on monotonicity at all - it simply evaluates the expressions for every part.
Treat is as a minmax index with granularity 1 across data parts. | 2020-09-08T20:37:13 | 2020-09-08T20:37:13 | {
"+1": 2
} | MEMBER |
ClickHouse | ClickHouse | 689,144,873 | 14,592 | filimonov | If 10 parallel inserts will flush the data every 10 sec (each) the average QPS will be 1. | 2020-09-08T21:26:16 | 2020-09-08T21:26:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,546,918 | 14,558 | bharatnc | With the changes above, beginning to see that the path is getting concatenated properly now (grepping journalctl with newest binary installed):
```bash
➜ journalctl | grep "Saved preprocessed configuration to"
Sep 07 11:12:23 tower clickhouse-server[480975]: Saved preprocessed configuration to '/var/lib/clickhous... | 2020-09-07T23:59:54 | 2020-09-08T08:28:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,757,812 | 14,578 | alesapin | Ok, let's merge without tests. | 2020-09-08T09:53:06 | 2020-09-08T09:53:06 | {} | MEMBER |
ClickHouse | ClickHouse | 688,948,788 | 14,415 | dreamerfable | Detailed description for updates of custom-paritioning-key:
- fix some wrong links
- fix some wrong format like the result of query
- correct the translation of the system table's name.It should not be translated
- correct the hint level according to the en version
- make some sentence more understandable | 2020-09-08T15:18:43 | 2020-09-08T15:18:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,487,525 | 14,557 | alexey-milovidov | This is a bug in error diagnostics in `gcc`. It does not correctly handle the case when assembler tool was terminated by signal. Instead it is processing partially read data and emits another error.
In contrast to `gcc`, `clang` does not have this bug. | 2020-09-07T19:41:00 | 2020-09-07T19:41:09 | {} | MEMBER |
ClickHouse | ClickHouse | 688,570,121 | 11,512 | dengc367 | Great, I am looking forward to merge this pr now, I test it in my environment. who can merge in to the master branch? | 2020-09-08T01:38:44 | 2020-09-08T01:38:44 | {} | NONE |
ClickHouse | ClickHouse | 688,577,226 | 14,201 | amuluowin | > First of all [#14201 (comment)](https://github.com/ClickHouse/ClickHouse/issues/14201#issuecomment-683350463) was just a work around, I'm not insisting, but
>
> > but system.text_log has no save the msg body
>
> There is `message` column.
>
> > and the log will more and more
>
> You can use materialize vi... | 2020-09-08T02:10:02 | 2020-09-08T02:10:02 | {} | NONE |
ClickHouse | ClickHouse | 688,924,536 | 14,494 | alesapin | > As I remember, we wanted to run recompression in low-priority threads to increase CPU utilization, when it's quiet low. Will it be ok to run recompression as regualr ttl merge (despite they are run in separate pool now)?
Two reasons why low thread priority is not implemented here:
1) In most cases, we will recomp... | 2020-09-08T14:42:26 | 2020-09-08T14:43:12 | {} | MEMBER |
ClickHouse | ClickHouse | 688,943,313 | 14,538 | den-crane | >What should materialize ttl do?
It applies new TTL rules and removes expired TTL rows.
>Partitions without TTL info stays like that.
What do you mean by that?
Can you show `show create table`
> is executing longer than distributed_ddl_task_timeout (=180) seconds.
It's the expected behavior. This comma... | 2020-09-08T15:10:29 | 2020-09-08T15:10:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,122,963 | 14,591 | alexey-milovidov | Maybe add a test? | 2020-09-08T20:38:47 | 2020-09-08T20:38:47 | {} | MEMBER |
ClickHouse | ClickHouse | 689,133,884 | 14,045 | radiatorius | CPU supports sse2:
**model name : Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz**
`flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts mmx fxsr sse sse2 ss syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts nopl xtopology tsc_reliable nonstop_tsc cpuid pni pclmulqdq ssse... | 2020-09-08T21:00:45 | 2020-09-08T21:00:45 | {} | NONE |
ClickHouse | ClickHouse | 688,649,906 | 12,761 | alesapin | `test_distributed_over_live_view` flaky | 2020-09-08T06:35:50 | 2020-09-08T06:35:50 | {} | MEMBER |
ClickHouse | ClickHouse | 688,735,185 | 14,045 | radiatorius | The issue does not depend on the table scheme. It also repeats for other tables and other **update** mutations. Here is an example that really causes **segmentation fault** everytime:
```
CREATE TABLE repository.test_table
(
`uuid_field` UUID DEFAULT generateUUIDv4(),
`int32_field` UInt32,
`int8_fie... | 2020-09-08T09:12:21 | 2020-09-08T09:12:21 | {} | NONE |
ClickHouse | ClickHouse | 688,830,492 | 14,496 | alesapin | Only comment updated, merge without tests | 2020-09-08T12:21:46 | 2020-09-08T12:21:46 | {} | MEMBER |
ClickHouse | ClickHouse | 688,866,399 | 14,555 | 4ertus2 | OOM in gcc builds
```
g++-9: fatal error: Killed signal terminated program cc1plus
``` | 2020-09-08T13:29:24 | 2020-09-08T13:29:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,897,392 | 5,182 | filimonov | Please check / help review: https://github.com/ClickHouse/ClickHouse/pull/14195/files#diff-bdc158a0a0a1ffa9fdd2041444cf17abR1
| 2020-09-08T14:03:16 | 2020-09-08T14:03:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,995,587 | 2,605 | alex-zaitsev | @simPod , there are IPv4 and IPv6 datatypes already. Does it work for you or you are looking for PostgreSQL compatibility? | 2020-09-08T16:32:16 | 2020-09-08T16:32:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,130,614 | 14,580 | filimonov | https://github.com/ClickHouse/ClickHouse/blob/c0fef71507b43cc8021bea5a3d96de38505322a8/src/Processors/Formats/Impl/ArrowColumnToCHColumn.cpp#L302-L304 | 2020-09-08T20:53:51 | 2020-09-08T20:53:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,134,560 | 14,576 | filimonov | There is an util to make that convertation, you can build it and try if you want (on own risk) - see
https://github.com/ClickHouse/ClickHouse/blob/master/utils/convert-month-partitioned-parts/main.cpp | 2020-09-08T21:02:08 | 2020-09-08T21:02:08 | {
"eyes": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,195,548 | 13,688 | CurtizJ | Test for Redis fixed in #14532. | 2020-09-08T23:46:47 | 2020-09-08T23:46:47 | {} | MEMBER |
ClickHouse | ClickHouse | 688,486,795 | 14,557 | alexey-milovidov | It is OOM while building - completely unrelated to any details of code formatting. | 2020-09-07T19:38:41 | 2020-09-07T19:38:41 | {} | MEMBER |
ClickHouse | ClickHouse | 688,525,655 | 14,556 | alexey-milovidov | How to edit the files:
```
find . -name ya.make -or -name ya.make.in | xargs grep -l -P 'LIBRARY|PROGRAM' | xargs sed -i -r -e 's/^SRCS\(/CFLAGS(-g0)\n\nSRCS(/'
``` | 2020-09-07T22:14:57 | 2020-09-07T22:14:57 | {} | MEMBER |
ClickHouse | ClickHouse | 688,599,249 | 4,060 | YuYuYan | thank you very much ,please help me | 2020-09-08T03:36:57 | 2020-09-08T03:36:57 | {} | NONE |
ClickHouse | ClickHouse | 688,650,166 | 12,761 | alesapin | OOM in GCC build not related to changes. | 2020-09-08T06:36:31 | 2020-09-08T06:36:31 | {} | MEMBER |
ClickHouse | ClickHouse | 688,651,277 | 14,533 | alesapin | failures not related to changes | 2020-09-08T06:38:57 | 2020-09-08T06:38:57 | {} | MEMBER |
ClickHouse | ClickHouse | 688,710,315 | 14,045 | alesapin | Can you provide your table schema and exact mutation query? | 2020-09-08T08:30:21 | 2020-09-08T08:30:21 | {} | MEMBER |
ClickHouse | ClickHouse | 688,837,699 | 14,583 | nikitamikhaylov | > LGTM.
>
> Two points:
>
> * we don't support `clang-8` already for a long time and we don't need any fixes;
>
> * the target is not build in "Arcadia" (that is typical pain point due to an old compiler).
>
>
> But this change is still worth doing.
I know that this target is not build in Arcad... | 2020-09-08T12:36:48 | 2020-09-08T12:36:48 | {} | MEMBER |
ClickHouse | ClickHouse | 688,526,231 | 14,549 | alexey-milovidov | One build did not finish due to OOM, other builds Ok. | 2020-09-07T22:17:52 | 2020-09-07T22:17:52 | {} | MEMBER |
ClickHouse | ClickHouse | 688,609,136 | 14,287 | Slach | @alexey-milovidov could you explain?
it is mean we won't fix to behavior of metric? or we just can't to cherry pick these changes? | 2020-09-08T04:21:04 | 2020-09-08T04:21:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,650,472 | 14,552 | alesapin | Ok, it works | 2020-09-08T06:37:12 | 2020-09-08T06:37:12 | {} | MEMBER |
ClickHouse | ClickHouse | 688,846,449 | 14,566 | filimonov | > I would expect that a user with only ALTER TTL granted would not also be able to run ALTER MATERIALIZE TTL queries.
Can you explain why do you think so? My expectations are opposite.
And in recent ClickHouse versions materialize ttl happen automatically when you alter ttl.
What do you think materialize do... | 2020-09-08T12:54:17 | 2020-09-08T12:54:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 688,920,252 | 14,573 | den-crane | Currently CH supports `create table R as L` but not `create table R like L` | 2020-09-08T14:36:07 | 2020-09-08T14:36:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 689,112,141 | 4,060 | filimonov | @YuYuYan It can be the same leak only if you use a very old clickhouse version.
Otherwise, it's something different and we need a detailed report and reproducible scenario to be able to help you. | 2020-09-08T20:16:22 | 2020-09-08T20:16:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,378,821 | 14,055 | alexey-milovidov | Fuzzer:
```
SELECT equals(count(), NULL GLOBAL IN
(
SELECT
2 GLOBAL IN (1023),
100
), -1)
FROM null_in
WHERE i IN (1, 3, NULL)
```
```
[linux-ubuntu-14-04-trusty] 2020.08.26 02:03:19.054107 [ 240 ] <Fatal> BaseDaemon: 11. /build/obj-x86_64-linux-gnu/../src/Colum... | 2020-08-26T00:47:42 | 2020-08-26T00:47:42 | {} | MEMBER |
ClickHouse | ClickHouse | 680,647,301 | 13,991 | bohutang | > @zhang2014 thanks, and by the way, do you think gtid-mode = on is a must, since I switch it off and I can see the replication also works.
Yes, other pr will support GTID based replication only, the postition is not a safe way and MySQL not prefered.
PR is here(still in review)
https://github.com/ClickHouse/Click... | 2020-08-26T05:11:46 | 2020-08-26T05:11:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,161,433 | 14,035 | ramazanpolat | > TSVRaw is output only format, it can be used with Kafka to produce the data, not for consume.
[Documentation ](https://clickhouse.tech/docs/en/interfaces/formats/#tabseparatedraw) says that it is both input and output format.
![image](https://user-images.githubusercontent.com/3527695/91206961-af042980-e710-11ea... | 2020-08-25T17:22:39 | 2020-08-25T17:22:39 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.