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 | 623,631,618 | 7,887 | alexey-milovidov | Fixed. | 2020-05-04T18:32:14 | 2020-05-04T18:32:14 | {} | MEMBER |
ClickHouse | ClickHouse | 623,643,039 | 9,926 | alexey-milovidov | Thank you! Let we apply this patch and look what will happen... | 2020-05-04T18:55:52 | 2020-05-04T19:30:24 | {} | MEMBER |
ClickHouse | ClickHouse | 623,794,964 | 10,654 | vzakaznikov | @alesapin thanks, the `test_rename_parallel_same_node` test now passes with your fix applied to the current master. However, when I try to run all the tests inside the `test_rename_column` the server crashes with the exception below.
```
2020.05.05 04:41:12.874286 [ 73 ] {} <Fatal> BaseDaemon: #####################... | 2020-05-05T01:47:31 | 2020-05-05T01:47:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,951,540 | 10,654 | alesapin | Trace is not complete and I cannot reproduce this segfault:
```
collected 4 items
test_rename_c... | 2020-05-05T09:28:16 | 2020-05-05T09:28:16 | {} | MEMBER |
ClickHouse | ClickHouse | 623,097,746 | 10,486 | azat | Rebased against upstream/master to fix conflicts (previous HEAD was 4db67291ce40ac710bdb0a4c1984562f8c22c027 - for the test report) | 2020-05-03T11:49:34 | 2020-05-03T11:49:50 | {} | MEMBER |
ClickHouse | ClickHouse | 623,170,483 | 10,636 | alexey-milovidov | By the way, you did not provide the structure of table `test`. | 2020-05-03T19:50:10 | 2020-05-03T19:50:10 | {} | MEMBER |
ClickHouse | ClickHouse | 623,182,676 | 10,638 | alexey-milovidov | Functional stateful tests (ubsan) - Ok. | 2020-05-03T21:09:51 | 2020-05-03T21:09:51 | {} | MEMBER |
ClickHouse | ClickHouse | 623,945,177 | 8,703 | filimonov | One more use case
```
CREATE VIEW view as
SELECT
dim1,dim2,dim3,dim4,dim5,dim6,sum(metric1),sum(metric2),sum(metric2)
from summt
GROUP BY dim1,dim2,dim3,dim4,dim5,dim6;
```
And the query
```
SELECT
dim1,dim6,sum(metric1),sum(metric2),sum(metric2)
from view
GROUP BY dim1,dim6;
``` | 2020-05-05T09:13:36 | 2020-05-05T09:14:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,101,181 | 10,598 | alexey-milovidov | Performance Ok. | 2020-05-03T12:17:00 | 2020-05-03T12:17:00 | {} | MEMBER |
ClickHouse | ClickHouse | 623,062,957 | 10,339 | tarampampam | I can create an PR (if you like this idea) | 2020-05-03T06:39:11 | 2020-05-03T18:22:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,196,676 | 10,604 | alexey-milovidov | Looks Ok. | 2020-05-03T22:55:31 | 2020-05-03T22:55:31 | {} | MEMBER |
ClickHouse | ClickHouse | 623,672,498 | 10,661 | mdelvecchio-wasabi | Is it possible that this could change?
I guess I am asking if a DateTime64(9) is stored on disk in the same way as my UInt64 "nanoseconds since the epoch" column is. I believe they are both 64-bit counts of nanoseconds since the epoch.
Is the above necessary because of the bug described in https://github.com/Cli... | 2020-05-04T19:55:20 | 2020-05-04T19:55:20 | {} | NONE |
ClickHouse | ClickHouse | 623,959,487 | 10,671 | filimonov | 1) Please try `set experimental_use_processors=0;` and check if performance is proper again.
2) please collect the logs of query execution (with experimental_use_processors 0 and 1), by running your query in clickhouse-client with `set send_logs_level='trace'` before it.
3) can you share the schema of the table you a... | 2020-05-05T09:47:58 | 2020-05-05T09:47:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 624,057,803 | 10,661 | den-crane | >Is it possible that this could change?
it seems it will not https://github.com/ClickHouse/ClickHouse/issues/10672 | 2020-05-05T13:33:19 | 2020-05-05T13:33:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 624,077,928 | 10,673 | Crazylearner30 | Yes!! we were using a wrong date 1900-01-01.
We changed it and its working fine. | 2020-05-05T14:08:28 | 2020-05-05T14:08:28 | {} | NONE |
ClickHouse | ClickHouse | 623,183,984 | 10,638 | alexey-milovidov | Having one remaining error:
```
2020-05-03 23:57:15 00900_orc_load: [ FAIL ] 0.30 sec. - having stderror:
2020-05-03 23:57:15 ../contrib/FastMemcpy/FastMemcpy.h:91:56: runtime error: applying zero offset to null pointer
2020-05-03 23:57:15 SUMMARY: Undefined... | 2020-05-03T21:19:15 | 2020-05-03T21:19:15 | {} | MEMBER |
ClickHouse | ClickHouse | 623,224,668 | 10,638 | alexey-milovidov | Performance Ok. | 2020-05-04T01:50:21 | 2020-05-04T01:50:21 | {} | MEMBER |
ClickHouse | ClickHouse | 623,399,805 | 9,522 | BayoNet | Fixed in #10649 | 2020-05-04T11:05:46 | 2020-05-04T11:05:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,480,163 | 10,646 | alexey-milovidov | Yes, it was fixed. | 2020-05-04T13:57:37 | 2020-05-04T13:57:37 | {} | MEMBER |
ClickHouse | ClickHouse | 623,695,931 | 9,189 | blind-oracle | @alexey-milovidov
Hmm, this probably should not cause this effect.
Let's consider a case:
1. OS boots, time is wrong and in future, some existing parts on disk with time in past (relative to server time), if they're inactive they will be immediately purged, that's expected.
2. CH starts, clients connect and in... | 2020-05-04T20:45:23 | 2020-05-04T20:45:23 | {} | NONE |
ClickHouse | ClickHouse | 624,002,997 | 10,654 | alexey-milovidov | Maybe it is related to https://github.com/ClickHouse/ClickHouse/pull/10625 | 2020-05-05T11:37:38 | 2020-05-05T11:37:38 | {} | MEMBER |
ClickHouse | ClickHouse | 624,026,353 | 2,045 | shsab | Any update on encryption? | 2020-05-05T12:29:58 | 2020-05-05T12:29:58 | {} | NONE |
ClickHouse | ClickHouse | 623,225,855 | 10,360 | alexey-milovidov | @BayoNet ? | 2020-05-04T01:58:15 | 2020-05-04T01:58:15 | {} | MEMBER |
ClickHouse | ClickHouse | 623,303,629 | 10,637 | tarampampam | PR description updated, and since `CLICKHOUSE_DB` is optional - no backward compatibility notes is required | 2020-05-04T07:30:10 | 2020-05-04T07:30:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,568,831 | 10,625 | alexey-milovidov | I have published these benchmarks on the website: https://presentations.clickhouse.tech/original_website/benchmark_gcc_clang.html
Note that this is just some noisy results, not like real performance tests. | 2020-05-04T16:30:41 | 2020-05-04T16:30:41 | {} | MEMBER |
ClickHouse | ClickHouse | 623,691,953 | 10,665 | ilugid | Got it working with specifying explicit Parition BY statement, rather than as a parameter to ENGINE. | 2020-05-04T20:36:57 | 2020-05-04T20:36:57 | {} | NONE |
ClickHouse | ClickHouse | 624,056,473 | 9,786 | den-crane | Yes, it covers. | 2020-05-05T13:30:48 | 2020-05-05T13:30:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,481,224 | 10,647 | filimonov | JSONEachRow support only some special type of the JSONs. When whole JSON object is in the single line. | 2020-05-04T13:59:24 | 2020-05-04T13:59:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,566,218 | 10,657 | alexey-milovidov | Ok.
PS. When any library is writing to stderr it is rather ignorant. Maybe rdkafka provides some way to plug in our own logger? | 2020-05-04T16:26:03 | 2020-05-04T16:26:03 | {} | MEMBER |
ClickHouse | ClickHouse | 623,636,419 | 3,039 | alexey-milovidov | > When the query is heavy, you must load the sub-query result into memory, which costs a lot.
Not necessarily. Subqueries can be processed in a streaming fashion.
Even for queries with GROUP BY.
> So, is there any other way to avoid query from a sub-query in my situation?
No. | 2020-05-04T18:42:01 | 2020-05-04T18:42:09 | {} | MEMBER |
ClickHouse | ClickHouse | 624,002,534 | 10,625 | alexey-milovidov | We have to revert again because it reproduces in some other places:
```
$ clickhouse-client --host 1.1.1.1
ClickHouse client version 20.5.1.1.
Connecting to 1.1.1.1:9000 as user default.
terminating with uncaught exception of type DB::NetException: Timeout: connect timed out: 1.1.1.1:9000 (1.1.1.1:9000)
Aborted... | 2020-05-05T11:36:24 | 2020-05-05T11:36:24 | {} | MEMBER |
ClickHouse | ClickHouse | 624,007,092 | 10,654 | vzakaznikov | > Maybe it is related to #10625
Yes, maybe. I will re-build using GCC instead of clang. | 2020-05-05T11:47:54 | 2020-05-05T11:47:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 624,056,905 | 10,310 | den-crane | it does. | 2020-05-05T13:31:37 | 2020-05-05T13:31:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 624,211,173 | 9,095 | alexey-milovidov | @artbeglaryan Yes, sending data to every shard is single-treaded when you use asynchronous INSERTS (that send data in background) and asynchronous mode is by default.
You can enable synchronous INSERTS with setting `insert_distributed_sync` set to 1.
With synchronous INSERTS, when you execute multiple queries concu... | 2020-05-05T17:55:07 | 2020-05-05T17:55:07 | {} | MEMBER |
ClickHouse | ClickHouse | 624,216,998 | 6,111 | alexey-milovidov | This issue is fixed after introducing user control with SQL queries.
It is available since version 20.4-prestable in experimental stage.
Documentation is added here: https://github.com/ClickHouse/ClickHouse/pull/9983 | 2020-05-05T18:06:07 | 2020-05-05T18:06:07 | {} | MEMBER |
ClickHouse | ClickHouse | 623,242,169 | 7,084 | filimonov | https://github.com/ClickHouse/clickhouse-odbc/issues/202 | 2020-05-04T03:37:18 | 2020-05-04T03:37:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,317,232 | 8,853 | AppKiv | no, I think it's not related, #9363 - about data, this issue about names - table definition/headers, maybe http-get restriction. | 2020-05-04T08:00:18 | 2020-05-04T08:00:18 | {} | NONE |
ClickHouse | ClickHouse | 623,442,666 | 10,647 | preved911 | I can't understand how configure consumer right now, because after JSONEachRow is causing exception, it is not continue getting messages from kafka. It is a bug? Should I create issue? | 2020-05-04T12:47:13 | 2020-05-04T12:47:13 | {} | NONE |
ClickHouse | ClickHouse | 623,654,257 | 6,741 | filimonov | It looks like an acceptable solution, do we need smth more here? Or we should introduce some stricter SQL compatibility?
```
SELECT
avgResample(10, 20, 1)(value, number),
avgOrDefaultResample(10, 20, 1)(value, number),
avgOrNullResample(10, 20, 1)(value, number),
avgResample(10, 20, 1)(toNul... | 2020-05-04T19:18:46 | 2020-05-04T19:22:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,736,137 | 4,966 | alexey-milovidov | But I recommend to always add LZ4 just in case:
```
CREATE TABLE codec_example
(
timestamp DateTime CODEC(DoubleDelta, LZ4),
slow_values Float32 CODEC(Gorilla, LZ4)
)
``` | 2020-05-04T22:14:16 | 2020-05-04T22:14:16 | {} | MEMBER |
ClickHouse | ClickHouse | 623,943,234 | 10,267 | filimonov | @hczhcz maybe you can check that? | 2020-05-05T09:09:11 | 2020-05-05T09:09:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,040,230 | 10,631 | alexey-milovidov | See also:
https://bugzilla.mozilla.org/show_bug.cgi?id=1614375
https://github.com/llvm/llvm-project/blob/master/compiler-rt/test/ubsan/TestCases/Pointer/nullptr-and-nonzero-offset-constants.cpp | 2020-05-03T01:40:16 | 2020-05-03T01:40:16 | {} | MEMBER |
ClickHouse | ClickHouse | 623,225,082 | 10,579 | alexey-milovidov | @4ertus2 it was merged. | 2020-05-04T01:53:01 | 2020-05-04T01:53:01 | {} | MEMBER |
ClickHouse | ClickHouse | 623,410,240 | 9,334 | filimonov | Hi @radist101, clickhouse-side log may have the answers - check that: https://github.com/ClickHouse/ClickHouse/issues/9334#issuecomment-590281028
From my perspective, the fixed IP in `KAFKA_ADVERTISED_HOST_NAME` may be potential reason.
Check also that compose file: https://github.com/filimonov/ClickHouse/blob/ka... | 2020-05-04T11:32:06 | 2020-05-04T11:32:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,612,889 | 10,657 | azat | >PS. When any library is writing to stderr it is rather ignorant. Maybe rdkafka provides some way to plug in our own logger?
Indeed, but I want to verify the lifetime of log and rdkafka threads before | 2020-05-04T17:56:32 | 2020-05-04T17:56:32 | {} | MEMBER |
ClickHouse | ClickHouse | 623,657,642 | 7,736 | silviucpp | @champtar Yes this was my finding as well. | 2020-05-04T19:25:51 | 2020-05-04T19:25:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,960,757 | 10,647 | filimonov | Check that: https://www.altinity.com/blog/clickhouse-kafka-engine-faq | 2020-05-05T09:51:07 | 2020-05-05T09:51:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 624,010,875 | 10,625 | amosbird | I think I've figured out some pattern of this unwinding failure:
It's related to exceptions thrown in ctor and created by std::make_unique or std::make_shared. However I cannot create a minimal test case to reproduce... | 2020-05-05T11:55:54 | 2020-05-05T11:55:54 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 624,216,932 | 10,288 | cw9 | Hi, I saw the same issue in 20.3.5, is this confirmed to be fixed in 20.3.6? | 2020-05-05T18:06:00 | 2020-05-05T18:06:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,633,363 | 3,708 | alexey-milovidov | This is closely related to the task https://clickhouse.tech/docs/ru/whats-new/extended-roadmap/#modifikatory-distinct-order-by-dlia-agregatnykh-funktsii | 2020-05-04T18:35:40 | 2020-05-04T18:35:40 | {} | MEMBER |
ClickHouse | ClickHouse | 623,735,575 | 4,966 | alexey-milovidov | @adrian17 Both `DoubleDelta` and `Gorilla` compresses data.
PS. `DoubleDelta` codec has bad name (the origin of this name is from some paper). It is not just about applying Delta transformation two times, it also encodes small integers in less number of bits. | 2020-05-04T22:12:49 | 2020-05-04T22:12:49 | {} | MEMBER |
ClickHouse | ClickHouse | 623,755,159 | 10,661 | den-crane | Yes, it's related to #9786. But now I have doubts that #9786 is a bug. It seems that is expected behaviour the same as https://github.com/ClickHouse/ClickHouse/issues/10310. | 2020-05-04T23:10:19 | 2020-05-04T23:11:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,096,823 | 10,339 | alexey-milovidov | Yes, it will be appreciated! | 2020-05-03T11:41:40 | 2020-05-03T11:41:40 | {} | MEMBER |
ClickHouse | ClickHouse | 623,097,171 | 10,516 | azat | Any feedback on this? This is bug fix pr for #10341 | 2020-05-03T11:44:31 | 2020-05-03T11:44:31 | {} | MEMBER |
ClickHouse | ClickHouse | 623,163,312 | 10,637 | filimonov | My expectation are
1) we don't need CLICKHOUSE_DEFAULT_ variables
2) it controlled by existing vars
```
CLICKHOUSE_DB: default
CLICKHOUSE_USER: default
CLICKHOUSE_PASSWORD: <empty>
```
3) If one of those vars is defined another way - it should change the behavior. (But that that may introduc... | 2020-05-03T19:03:12 | 2020-05-03T19:03:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,593,177 | 9,843 | den-crane | @alexey-milovidov what about https://github.com/ClickHouse/ClickHouse/issues/10566 ? | 2020-05-04T17:18:13 | 2020-05-04T17:18:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,601,089 | 9,843 | alexey-milovidov | It also looks like a critical issue. I see that a pull request with fix is already here.
A bug that was fixed quickly is not a bug at all. | 2020-05-04T17:34:01 | 2020-05-04T17:34:01 | {} | MEMBER |
ClickHouse | ClickHouse | 623,632,463 | 10,661 | den-crane | the only way is
```
ALTER TABLE db.table add COLUMN `EndTimeNS1` DateTime64(9) default cast(substr(toString(EndTimeNS), 1, 10) ||'.'|| substr(toString(EndTimeNS), 11, 9), 'DateTime64(9)');
set mutations_sync = 1;
alter table db.table update EndTimeNS1 = EndTimeNS1;
alter table db.table rename column EndTimeNS t... | 2020-05-04T18:33:55 | 2020-05-04T18:35:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,946,600 | 7,612 | Crazylearner30 | > Not sure, probably it works in test releases
> like this
> insert into mytable
> select * from s3('http://s3auth:s3auth@host/bucket/test.csv', 'CSV', 'column1 UInt32, column2 UInt32, column3 UInt32')
>
>
> That is not correct. One shall use this syntax:
> select * from s3('http://host/bucket/test.csv', 'acce... | 2020-05-05T09:16:38 | 2020-05-05T09:16:38 | {} | NONE |
ClickHouse | ClickHouse | 623,951,375 | 9,786 | filimonov | I think that FR will cover the usecase: https://github.com/ClickHouse/ClickHouse/issues/10672 ? | 2020-05-05T09:27:50 | 2020-05-05T09:27:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 624,195,355 | 10,602 | blinkov | @ceyhunkerti afaik this issue has already been fixed a while ago. Could you try again to see if it still affects you? | 2020-05-05T17:24:52 | 2020-05-05T17:24:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,238,312 | 8,853 | filimonov | Related #9363 ? | 2020-05-04T03:16:22 | 2020-05-04T03:16:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,242,032 | 1,826 | filimonov | https://github.com/ClickHouse/clickhouse-odbc/issues/202 | 2020-05-04T03:36:30 | 2020-05-04T03:36:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,497,277 | 10,647 | preved911 | So, I should use another `kafka_format` for to avoid this error? Which format's you can to recommend for my case? | 2020-05-04T14:27:17 | 2020-05-04T14:27:17 | {} | NONE |
ClickHouse | ClickHouse | 623,995,079 | 10,654 | vzakaznikov | Interesting. For me the crash is reproduced every other run. That short trace is all I get in the log.
For example, here is another one. Let's see what CI shows.
```
2020.05.05 14:12:36.659326 [ 8 ] {} <Trace> BaseDaemon: Received signal 11
2020.05.05 14:12:36.659673 [ 71 ] {} <Fatal> BaseDaemon: ##################... | 2020-05-05T11:17:34 | 2020-05-05T11:17:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,174,464 | 10,625 | alexey-milovidov | Just download and open directly. | 2020-05-03T20:17:21 | 2020-05-03T20:17:21 | {} | MEMBER |
ClickHouse | ClickHouse | 623,305,621 | 10,637 | tarampampam | Local tests passed:
```shell script
$ docker build --tag clickhouse:local --build-arg "version=19.16.*" .
$ docker run --rm -e "CLICKHOUSE_PASSWORD=bar" -e "CLICKHOUSE_DB=baz" clickhouse:local
Create user 'default' instead default
...
/entrypoint.sh: create database 'baz'
/entrypoint.sh: ignoring /docker-ent... | 2020-05-04T07:34:45 | 2020-05-04T07:36:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,332,229 | 10,637 | tarampampam | @filimonov PR description updated | 2020-05-04T08:35:00 | 2020-05-04T08:35:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,404,036 | 10,647 | filimonov | Currently, it is not very intuitive. It is not an on/off flag, but maximum number of broken messages per block. Size of the block depends on kafka_max_block_size, poll_interval, and number of messages currently available in the topic.
That feature will be redesigned (i hope soon). It will just have an `_error` vir... | 2020-05-04T11:16:54 | 2020-05-04T11:16:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,572,860 | 10,659 | alesapin | Tests for a fix in this PR https://github.com/ClickHouse/ClickHouse/pull/10654 | 2020-05-04T16:38:23 | 2020-05-04T16:38:23 | {} | MEMBER |
ClickHouse | ClickHouse | 623,576,813 | 10,628 | alexey-milovidov | Sadly enough, GitHub.com does not support IPv6.
It means, that the query
```
SELECT * FROM url('https://raw.githubusercontent.com/ClickHouse/ClickHouse/master/utils/list-versions/version_date.tsv', TSV, 'version String, date Date')
```
won't work on servers of most (good) companies (like Yandex), because they are ... | 2020-05-04T16:46:20 | 2020-05-04T16:46:20 | {} | MEMBER |
ClickHouse | ClickHouse | 624,176,736 | 4,500 | BayoNet | Fixed in #5332 and #5487 | 2020-05-05T16:58:54 | 2020-05-05T16:59:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,060,563 | 8,842 | zhang2014 | I forgot about it when trying to resolve the conflict. let me check again. | 2020-05-03T06:11:08 | 2020-05-03T06:11:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,157,026 | 10,339 | tarampampam | @alexey-milovidov @filimonov PR #10637 created | 2020-05-03T18:22:39 | 2020-05-03T18:22:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,181,585 | 10,641 | alexey-milovidov | It looks like this:
```
2020.05.03 23:40:33.821985 [ 3396 ] {} <Trace> BaseDaemon: Received signal 11
2020.05.03 23:40:33.822160 [ 3525 ] {} <Fatal> BaseDaemon: ########################################
2020.05.03 23:40:33.822199 [ 3525 ] {} <Fatal> BaseDaemon: (version 20.5.1.1) (from thread 3393) (no query) Rece... | 2020-05-03T21:01:40 | 2020-05-03T21:01:40 | {} | MEMBER |
ClickHouse | ClickHouse | 623,183,128 | 10,642 | alexey-milovidov | #9249 | 2020-05-03T21:13:10 | 2020-05-03T21:13:10 | {} | MEMBER |
ClickHouse | ClickHouse | 623,191,555 | 10,604 | vzakaznikov | Last CI run shows that `Functional stateless tests (debug)` had no fails. | 2020-05-03T22:19:54 | 2020-05-03T22:19:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,652,625 | 2,494 | filimonov | It is still there:
```
SELECT countDistinct(toNullable(number))
FROM numbers(1)
WHERE 1 = 0
┌─uniqExact(toNullable(number))─┐
│ ᴺᵁᴸᴸ │
└───────────────────────────────┘
1 rows in set. Elapsed: 0.001 sec.
SELECT uniqExactOrDefault(toNullable(number))
FROM numbers(1)
WHERE 1 = 0... | 2020-05-04T19:15:10 | 2020-05-04T19:15:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,951,455 | 10,310 | filimonov | I think that FR will cover the usecase: #10672 ? | 2020-05-05T09:28:02 | 2020-05-05T09:28:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 624,171,416 | 10,360 | BayoNet | I've missed this message somewhere. I'll do. | 2020-05-05T16:49:08 | 2020-05-05T16:49:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,184,107 | 10,637 | filimonov | Please update also PR description, and add backward compatibility note | 2020-05-03T21:20:10 | 2020-05-03T21:20:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,198,426 | 10,486 | azat | Check failures looks unrelated | 2020-05-03T23:08:14 | 2020-05-03T23:08:14 | {} | MEMBER |
ClickHouse | ClickHouse | 623,224,468 | 10,583 | alexey-milovidov | You mean:
```
curl http://localhost:8123/ -d "SELECT * FROM generateRandom('\"ParsedParams.Key1\" Array(String), \"ParsedParams.Key2\" Array(Float64), x String', 1, 10, 2) LIMIT 10"
curl: (52) Empty reply from server
``` | 2020-05-04T01:49:06 | 2020-05-04T01:49:06 | {} | MEMBER |
ClickHouse | ClickHouse | 623,542,709 | 9,843 | alexey-milovidov | One critical performance issue has been found: https://github.com/ClickHouse/ClickHouse/issues/10655 | 2020-05-04T15:45:09 | 2020-05-04T15:45:09 | {} | MEMBER |
ClickHouse | ClickHouse | 623,755,240 | 9,412 | oik741 | Could you guys release 19.17 branch with this fix?
Can't proceed with upgrade to 20.x because of broken replication =( | 2020-05-04T23:10:34 | 2020-05-04T23:10:34 | {} | NONE |
ClickHouse | ClickHouse | 623,884,797 | 10,594 | wutchzone | Nothing in error log except issue #10158 , I will try to more experiment with this weird behavior and keep this thread updated. | 2020-05-05T06:51:21 | 2020-05-05T06:51:21 | {} | NONE |
ClickHouse | ClickHouse | 624,015,565 | 10,673 | alexey-milovidov | It looks like you are importing some column from Parquet into Date type in ClickHouse.
Date type is represented by UInt16 containing the number of days since 1970-01-01.
But in Parquet file you have some big number (4294941729) or negative number (-25566) instead. | 2020-05-05T12:06:05 | 2020-05-05T12:06:05 | {} | MEMBER |
ClickHouse | ClickHouse | 624,018,477 | 7,612 | alexey-milovidov | It means, literally, that ClickHouse server cannot establish connection to S3.
To have more context on this error, you may:
- check clickhouse-server.log for more detailed error message;
- try to connect to s3 from the same machine using another tools;
(expert level below)
- use tcpdump and check what packets ... | 2020-05-05T12:12:37 | 2020-05-05T12:12:37 | {} | MEMBER |
ClickHouse | ClickHouse | 624,182,250 | 10,602 | BayoNet | @blinkov In production I don't see such behavior, but I see it when building locally. Left TOC bar and lang switcher doesn't work. | 2020-05-05T17:06:12 | 2020-05-05T17:06:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,173,459 | 10,637 | tarampampam | I left `CLICKHOUSE_DB` empty by default for backward compatibility (and variable `CLICKHOUSE_DB` is optional) | 2020-05-03T20:10:45 | 2020-05-03T20:10:45 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,630,106 | 7,736 | champtar | @alesapin @akuzm it seems `POCO_BROKEN_TIMEOUTS` (https://github.com/ClickHouse-Extras/poco/commits/clickhouse/Net/include/Poco/Net/SocketDefs.h) breaks `cancel_http_readonly_queries_on_client_close` | 2020-05-04T18:29:13 | 2020-05-04T18:29:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,723,064 | 4,966 | adrian17 | > a codec that does some transformations but does not compress data without compressing codec later in the list (example: CODEC(Delta));
That means that the example in documentation should be updated, right?
```
CREATE TABLE codec_example
(
timestamp DateTime CODEC(DoubleDelta),
slow_values Float32 CO... | 2020-05-04T21:41:35 | 2020-05-04T21:41:35 | {} | NONE |
ClickHouse | ClickHouse | 624,072,731 | 10,654 | vzakaznikov | Confirmed that there is no issue when using GCC 9.
| 2020-05-05T13:59:24 | 2020-05-05T13:59:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,097,241 | 10,631 | alexey-milovidov | Yes, it helps. | 2020-05-03T11:45:12 | 2020-05-03T11:45:12 | {} | MEMBER |
ClickHouse | ClickHouse | 623,401,748 | 10,651 | hamsterready | Now I can see it is dup of #10613 | 2020-05-04T11:10:48 | 2020-05-04T11:10:48 | {} | NONE |
ClickHouse | ClickHouse | 623,546,552 | 10,654 | vzakaznikov | @alesapin, could you please look at these tests? I've tried to simplify but even a simple `test_rename_parallel_same_node` does not work for me. Seems like we have a problem with renames done in parallel. | 2020-05-04T15:51:29 | 2020-05-04T15:51:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 623,572,678 | 10,654 | alesapin | Fix for the found bug is here https://github.com/ClickHouse/ClickHouse/pull/10659. You can cherry-pick this commit into your branch. | 2020-05-04T16:38:05 | 2020-05-04T16:38:05 | {} | MEMBER |
ClickHouse | ClickHouse | 623,684,991 | 9,189 | alexey-milovidov | We are using part modification time to remove old parts. For existing parts it is initialized with directory modification time on filesystem and for parts created in runtime it is set to wall clock time. So, this is expected to happen. Let's think what we can change... | 2020-05-04T20:22:02 | 2020-05-04T20:22:02 | {} | MEMBER |
ClickHouse | ClickHouse | 630,477,080 | 10,268 | alexey-milovidov | I'm too greedy. | 2020-05-18T22:55:28 | 2020-05-18T22:55:54 | {} | MEMBER |
ClickHouse | ClickHouse | 630,921,400 | 10,908 | qoega | @alesapin You can check new version. | 2020-05-19T16:08:28 | 2020-05-19T16:08:28 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.