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 | 696,990,142 | 15,115 | alexey-milovidov | AWS c6gd.metal (Graviton 2):
```
sudo mdadm --create --verbose /dev/md0 --level=0 --raid-devices=2 /dev/nvme1n1 /dev/nvme2n1
sudo mkfs.ext4 -F /dev/md0
sudo mkdir -p /mnt/md0
sudo mount /dev/md0 /mnt/md0
sudo chown ec2-user /mnt/md0
cd /mnt/md0
[0.004, 0.003, 0.001],
[0.085, 0.030, 0.032],
[0.029, 0.028, ... | 2020-09-22T21:23:43 | 2020-09-22T21:24:59 | {} | MEMBER |
ClickHouse | ClickHouse | 697,051,869 | 15,112 | alexey-milovidov | @qoega has prepared SQLancer to almost ready state.
Yes, it's reasonable to run it in CI. | 2020-09-23T00:29:32 | 2020-09-23T00:29:32 | {} | MEMBER |
ClickHouse | ClickHouse | 697,076,509 | 15,021 | zhang2014 | Yep, that's is we plan to do
https://github.com/ClickHouse/ClickHouse/issues/4006#issuecomment-657486437
- [ ] Add gc thread for remove deleted rows
| 2020-09-23T02:00:21 | 2020-09-23T02:00:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,223,315 | 14,711 | alesapin | Can you provide an example of the generated docs page about cmake? | 2020-09-23T08:42:55 | 2020-09-23T08:42:55 | {} | MEMBER |
ClickHouse | ClickHouse | 697,265,873 | 15,180 | ikainazarov | > The outdated parts were removed in the background. The expired parts were replaced by empty parts with commited state, which generated in the merge phase. If the generated empty parts were not commited, and just remove the outdated parts, this problem will be solved.
>
> I have fixed this problem in my test env. C... | 2020-09-23T10:05:13 | 2020-09-23T10:05:13 | {} | NONE |
ClickHouse | ClickHouse | 697,346,275 | 15,191 | fastio | > Ok, so we have to add some tests for this case?
Sure, I'll do it. | 2020-09-23T12:56:19 | 2020-09-23T12:56:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,746,604 | 4,250 | antonmarin | Hi. Is it solved? | 2020-09-22T14:09:06 | 2020-09-22T14:09:06 | {} | NONE |
ClickHouse | ClickHouse | 696,881,343 | 15,122 | alexey-milovidov | @den-crane
> Usually it delays for 2-4 weeks after release.
That's slightly pessimistic overestimation. | 2020-09-22T17:54:09 | 2020-09-22T17:54:09 | {} | MEMBER |
ClickHouse | ClickHouse | 696,924,255 | 15,157 | tavplubix | Thank you for an attempt to fix this annoying issue, but the fix seems to be very ad hoc. `StorageJoin` engine tables are not that special and I don't think that we should load them first as a special case. Also, it doesn't fix cases like the following:
```
create database z;
create table z.j (a Int32, b Int32) Engi... | 2020-09-22T19:12:45 | 2020-09-22T19:49:17 | {} | MEMBER |
ClickHouse | ClickHouse | 697,057,492 | 10,296 | alexey-milovidov | Integration to Phoronix Test Suite is also considered. | 2020-09-23T00:51:14 | 2020-09-23T00:51:14 | {} | MEMBER |
ClickHouse | ClickHouse | 697,078,085 | 15,035 | azat | >Merge with master and check again?
Still a problem, because PipelineExecutor does not call wait, since it does not use ThreadPool, it only check the state, anyway it is not related to this PR, and it will fail even w/o biding to specific CPU, affinity just makes test more reliable
Fix for this problem is here - #1... | 2020-09-23T02:06:08 | 2020-09-23T02:06:08 | {} | MEMBER |
ClickHouse | ClickHouse | 697,170,690 | 15,135 | akuzm | Stress test broken in master. | 2020-09-23T06:51:39 | 2020-09-23T06:51:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,173,990 | 15,161 | John-belt | Yup that makes sense, thanks | 2020-09-23T06:59:45 | 2020-09-23T07:00:06 | {} | NONE |
ClickHouse | ClickHouse | 697,256,530 | 15,181 | simPod | Ok it might have restarted ALTER query that runs now 🤔 I'll get back when it finishes. | 2020-09-23T09:46:26 | 2020-09-23T09:46:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,298,414 | 15,190 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/7677 | 2020-09-23T11:17:08 | 2020-09-23T11:17:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,716,248 | 14,856 | KochetovNicolai | `test_rename_distributed_parallel_insert_and_select` is not connected to this pr. | 2020-09-22T13:19:16 | 2020-09-22T13:19:16 | {} | MEMBER |
ClickHouse | ClickHouse | 696,749,069 | 4,250 | den-crane | @antonmarin it was nothing so solve. It came from Materialized View design.
_MV это insert trigger. Этот триггер следит за инсертами в исходную таблицу, и выполняет над вставленными записями преобразование select и вставляет результат, либо в явно созданную таблицу, либо в неявно созданную inner таблицу._
_Т.е. ис... | 2020-09-22T14:13:06 | 2020-09-22T14:13:06 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,881,794 | 14,588 | alexey-milovidov | 20.9 is released, changelog is available: https://github.com/ClickHouse/ClickHouse/blob/master/CHANGELOG.md | 2020-09-22T17:54:51 | 2020-09-22T17:54:51 | {} | MEMBER |
ClickHouse | ClickHouse | 697,058,253 | 12,332 | alexey-milovidov | #15087 | 2020-09-23T00:53:57 | 2020-09-23T00:53:57 | {} | MEMBER |
ClickHouse | ClickHouse | 697,211,645 | 14,265 | ilejn | Retry strategy did not work well: new presumably unrelated test failures.
Colleagues, could you provide further suggestions or merge this PR, please. | 2020-09-23T08:20:19 | 2020-09-23T08:20:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,232,290 | 14,265 | filimonov | Fuzzer failure is the #13393
Stress test (thread) reported https://github.com/ClickHouse/ClickHouse/issues/14414#issuecomment-697226214
Stress test (UB) `munmap_chunk(): invalid pointer` - have no idea how to get the context. | 2020-09-23T08:59:47 | 2020-09-23T09:02:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,936,333 | 15,168 | filimonov | +1
https://github.com/ClickHouse/ClickHouse/pull/15169/files
| 2020-09-22T19:36:33 | 2020-09-22T19:36:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,010,538 | 14,791 | filimonov | As for me in short / small scope that unfolding of {database} and {table} and probably {UUID} during table creation seems 'good enough' improvement which will unblock that PR - since the zookeeper path is immutable on zookeeper side after created, let's keep it immutable on ClickHouse side too (well, replica / shard wi... | 2020-09-22T22:13:56 | 2020-09-22T22:18:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,077,598 | 15,021 | liangli1213 | @zhang2014 , many thanks. | 2020-09-23T02:04:18 | 2020-09-23T02:04:18 | {} | NONE |
ClickHouse | ClickHouse | 697,084,389 | 7,320 | shwangdev | Really need this function. | 2020-09-23T02:22:38 | 2020-09-23T02:22:38 | {} | NONE |
ClickHouse | ClickHouse | 696,882,189 | 15,122 | simPod | He meant hours for sure 🙃 | 2020-09-22T17:55:32 | 2020-09-22T17:55:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,901,188 | 15,153 | amosbird | ### toDateTime
http://oracle.wentropy.com/c82231bfaacd9a5c
### toString
http://oracle.wentropy.com/115055bb23b3a5b8
### toUInt32
http://oracle.wentropy.com/f94d0d6dab522cee
So toDateTime comparison somehow fallbacks to the slow generic implementation.
P.S. The user experience of https://github.com/... | 2020-09-22T18:29:19 | 2020-09-22T18:29:19 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 697,075,736 | 15,175 | zhang2014 | fixed in https://github.com/ClickHouse/ClickHouse/pull/14852 | 2020-09-23T01:57:47 | 2020-09-23T01:57:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,076,843 | 15,176 | liangli1213 | @BohuTANG | 2020-09-23T02:01:37 | 2020-09-23T02:01:37 | {} | NONE |
ClickHouse | ClickHouse | 697,223,824 | 15,181 | simPod | Should dumping traces have any effect on it?
I'll ask admins to do continuous restart of the cluster. | 2020-09-23T08:43:49 | 2020-09-23T08:43:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,298,766 | 15,190 | den-crane | check this https://github.com/ClickHouse/ClickHouse/issues/12586 | 2020-09-23T11:17:55 | 2020-09-23T11:17:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,328,146 | 15,181 | simPod | It is processing, very slowly though 🤔 | 2020-09-23T12:22:45 | 2020-09-23T12:22:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,743,164 | 15,122 | den-crane | work in progress https://github.com/ClickHouse/ClickHouse/pull/15154/files | 2020-09-22T14:03:37 | 2020-09-22T14:03:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,885,298 | 14,627 | 4ertus2 | There's implicit requirement for RIGHT|FULL JOIN that it's an equi-join. But it's possible to make not euqi-join condition in ON section without exception. It should be forbidden till #8802 | 2020-09-22T18:00:45 | 2020-09-22T18:00:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,899,313 | 14,962 | alexey-milovidov | Easy. | 2020-09-22T18:25:53 | 2020-09-22T18:25:53 | {} | MEMBER |
ClickHouse | ClickHouse | 696,903,887 | 15,161 | alexey-milovidov | It is enabled by default for all queries since version 20.5.
In older version you can enable it manually with the `log_queries` setting.
In new version, you can disable it with the same setting...
That's consistent with your report :)
> Does it depend on disk space also?? i mean the table will grow over time and ... | 2020-09-22T18:34:20 | 2020-09-22T18:35:12 | {} | MEMBER |
ClickHouse | ClickHouse | 697,074,632 | 15,021 | liangli1213 | @BohuTANG for your A2, do you have plan to let ClickHouse to retrieve the removed row? so space can be released on disk. | 2020-09-23T01:53:28 | 2020-09-23T01:53:28 | {} | NONE |
ClickHouse | ClickHouse | 696,864,958 | 14,414 | filimonov | Stress tests ThreadSanitizer: data race /build/obj-x86_64-linux-gnu/../src/Core/SettingsFields.h:43:52 in DB::SettingFieldNumber<unsigned long>::operator=(unsigned long) https://clickhouse-test-reports.s3.yandex.net/14265/494e59e7c12e0887a32b2564463a84aea500118b/stress_test_(thread)/stderr.log | 2020-09-22T17:25:55 | 2020-09-23T09:00:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,276,059 | 15,181 | simPod | Yes, though the queue is still full. Is there any way to clear it? | 2020-09-23T10:26:52 | 2020-09-23T10:26:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,391,535 | 15,180 | ikainazarov | thank you guys | 2020-09-23T13:54:27 | 2020-09-23T13:54:27 | {} | NONE |
ClickHouse | ClickHouse | 696,898,431 | 14,962 | alexey-milovidov | @akuzm The new test case is a little bit flawed, could you please make some edits? | 2020-09-22T18:24:06 | 2020-09-22T18:24:06 | {} | MEMBER |
ClickHouse | ClickHouse | 696,905,872 | 15,153 | alexey-milovidov | > The user experience of https://github.com/Slach/clickhouse-flamegraph is very appealing.
Also try https://github.com/laplab/clickhouse-speedscope, it's even more rich.
> So toDateTime comparison somehow fallbacks to the slow generic implementation.
It shouldn't... | 2020-09-22T18:38:03 | 2020-09-22T18:38:03 | {} | MEMBER |
ClickHouse | ClickHouse | 696,988,474 | 15,173 | den-crane | >20.4.4.18
"Possible deadlock avoided." has been fixed in 20.6
You need to upgrade at least to 20.6 (though it's not supported anymore).
Supported versions are 20.7 , 20.8 , 20.9 | 2020-09-22T21:20:14 | 2020-09-22T21:20:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,027,304 | 15,145 | alexey-milovidov | > test_rename_column
- recently broken, maybe here: #11913
> Stress test (thread)
- reported | 2020-09-22T23:03:47 | 2020-09-22T23:03:47 | {} | MEMBER |
ClickHouse | ClickHouse | 697,056,856 | 10,296 | alexey-milovidov | The first step: https://github.com/ClickHouse/ClickHouse/pull/15115 | 2020-09-23T00:48:50 | 2020-09-23T00:48:50 | {} | MEMBER |
ClickHouse | ClickHouse | 697,151,514 | 15,118 | chang0391 | @BohuTANG @zhang2014
Thanks | 2020-09-23T06:00:04 | 2020-09-23T06:00:04 | {} | NONE |
ClickHouse | ClickHouse | 697,213,922 | 15,144 | akuzm | https://clickhouse-test-reports.s3.yandex.net/15144/ab19bb25fd8c286713580649ad1c183d493ea5dc/performance_comparison/report.html#fail1
disabling percpu arenas didn't help, it still segfaults | 2020-09-23T08:24:48 | 2020-09-23T08:24:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,214,929 | 15,144 | akuzm | > https://clickhouse-test-reports.s3.yandex.net/15144/ab19bb25fd8c286713580649ad1c183d493ea5dc/performance_comparison/report.html#fail1
>
> disabling percpu arenas didn't help, it still segfaults
Oh wait, it did help -- only the old server segfaults, not the new one. Now to find some way to disable it in runtime.... | 2020-09-23T08:26:49 | 2020-09-23T08:26:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,255,835 | 15,186 | haochencheng | > What does `git status` show?
> The error is in some test, you can skip it by using `ninja clickhouse-bundle` -- it will only build the server, client and other essential parts, skipping all non-essential things like unit tests.
thanks . | 2020-09-23T09:45:05 | 2020-09-23T09:45:05 | {} | NONE |
ClickHouse | ClickHouse | 697,333,457 | 15,190 | den-crane | @worenga you did not understand.
>Reason being is that clickhouse ensures that the length
because of the **dot** !!!!! in a column name.
Nested it's a syntax sugar for "CREATE TABLE" only. It means nothing. It does nothing.
| 2020-09-23T12:32:56 | 2020-09-23T12:33:38 | {
"+1": 1,
"confused": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,895,150 | 10,913 | 4ertus2 | Probably related #7593 | 2020-09-22T18:18:10 | 2020-09-22T18:18:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,933,043 | 14,686 | filimonov | Restart is not needed, it rereads the schema every time. | 2020-09-22T19:29:54 | 2020-09-22T19:29:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,084,145 | 4,914 | h12567 | May I ask is there any docs on this? Also what's the design considerations for changing from BlockStream processing to this Processor logic? | 2020-09-23T02:22:29 | 2020-09-23T02:59:16 | {} | NONE |
ClickHouse | ClickHouse | 697,250,411 | 15,180 | fastio | The outdated parts were removed in the background. The expired parts were replaced by empty parts with commited state, which generated in the merge phase. If the generated empty parts were not commited, and just remove the outdated parts, this problem will be solved.
I have fixed this problem in my test env. Can I m... | 2020-09-23T09:34:17 | 2020-09-23T09:37:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,290,655 | 15,185 | den-crane | related https://github.com/ClickHouse/ClickHouse/issues/4390#issuecomment-463623903 | 2020-09-23T11:00:03 | 2020-09-23T11:00:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,882,890 | 15,163 | alexey-milovidov | There is intention to enable:
\- compact parts;
\- atomic database;
by default in this release but it's not necessarily.
| 2020-09-22T17:56:49 | 2020-09-22T17:56:58 | {} | MEMBER |
ClickHouse | ClickHouse | 697,023,547 | 14,791 | alexey-milovidov | > As for me in short / small scope that unfolding of {database} and {table} and probably {UUID} during table creation seems 'good enough' improvement which will unblock that PR
Agree. | 2020-09-22T22:51:46 | 2020-09-22T22:51:46 | {} | MEMBER |
ClickHouse | ClickHouse | 697,077,231 | 15,007 | azat | >Integration tests (asan) — fail: 88, passed: 676, error: 77
>Integration tests (release) — fail: 86, passed: 676, error: 79
At least now it does not timedout | 2020-09-23T02:03:01 | 2020-09-23T02:03:01 | {} | MEMBER |
ClickHouse | ClickHouse | 697,247,765 | 15,181 | simPod | Though restarting the replicas had no effect on items in the queue. | 2020-09-23T09:29:05 | 2020-09-23T09:29:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,254,942 | 15,186 | akuzm | What does `git status` show?
The error is in some test, you can skip it by using `ninja clickhouse-bundle` -- it will only build the server, client and other essential parts, skipping all non-essential things like unit tests. | 2020-09-23T09:43:17 | 2020-09-23T09:43:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,953,734 | 15,170 | filimonov | topological sorting/dependencies analysis is a very good option, but relatively big/complicated.
That will allow solving issues like https://github.com/ClickHouse/ClickHouse/issues/7944 (needed for example to backup recovery) and will require to detect all possible dependencies between tables (there are a lot of non... | 2020-09-22T20:12:52 | 2020-09-22T20:12:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,003,644 | 12,510 | filimonov | Exit codes for application is 1 byte integer. It can't go higher that 255.
Real exception codes can be any number.
So it can't be fixed.
We can think of mapping group of exceptions codes to single exit code (we do smth similar for http) | 2020-09-22T21:56:28 | 2020-09-22T21:56:28 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,270,134 | 15,181 | alesapin | DDL worker seems to be alive:
```
Thread 211 (Thread 0x7f1e80ead700 (LWP 30444)):
#0 0x00007f1f2229b17f in pthread_cond_wait@@GLIBC_2.3.2 () from /lib/x86_64-linux-gnu/libpthread.so.0
#1 0x0000000018dfb36b in Poco::EventImpl::waitImpl() ()
#2 0x000000001588f277 in DB::DDLWorker::runMainThread() ()
#3 0x0000... | 2020-09-23T10:13:55 | 2020-09-23T10:13:55 | {} | MEMBER |
ClickHouse | ClickHouse | 697,270,902 | 15,162 | 4ertus2 | The problem with PR is distributed queries expect ON with `_shard_num` working
```
SELECT * FROM a JOIN b ON _shard_num = b.shard_num
``` | 2020-09-23T10:15:39 | 2020-09-23T10:15:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,716,124 | 15,089 | alesapin | > Doesn't work for compact parts :(
fixed | 2020-09-22T13:19:04 | 2020-09-22T13:19:04 | {} | MEMBER |
ClickHouse | ClickHouse | 697,058,788 | 14,692 | alexey-milovidov | It means that some read operations are slow. There are many read operations and under high system load it's expected that some of them may be slower than the threshold. If you see these messages only occasionally it's normal.
Otherwise it may indicate that you have too slow IO subsystem or too high load. | 2020-09-23T00:56:07 | 2020-09-23T00:56:26 | {} | MEMBER |
ClickHouse | ClickHouse | 697,217,363 | 15,181 | alesapin | Maybe DDLWorker thread is dead for some reason? Can you dump traces with gdb using: `gdb -ex 'set pagination off' -ex "set logging file gdb.log" -ex 'set logging on' -ex 'backtrace' -ex 'thread apply all backtrace' -ex 'backtrace' --pid your_clickhouse_pid`. After that try to restart the server, DDLWorker should start ... | 2020-09-23T08:31:26 | 2020-09-23T08:31:26 | {} | MEMBER |
ClickHouse | ClickHouse | 697,311,537 | 15,065 | ucasfl | > Implementation of a table engine to consume application log files in ClickHouse
Hi, can I do this task? @alexey-milovidov | 2020-09-23T11:48:02 | 2020-09-23T11:48:02 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 697,314,348 | 15,180 | fastio | @ikainazarov As mentioned above, it's unexpected behaviour. I try to make a pr to improvement it. | 2020-09-23T11:54:26 | 2020-09-23T11:54:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,773,475 | 15,123 | lzy1g1225 | Thanks, for convenience I am using clickhouse-copier now, however while copying I saw it is creating newtable_piece_0, newtable_piece_1 ...., but in taks.xml I only specified one 'newtable' following the documentation, what are these xxx_piece_x tables? Am I missing something? | 2020-09-22T14:52:05 | 2020-09-22T14:53:11 | {} | NONE |
ClickHouse | ClickHouse | 696,851,851 | 14,265 | ilejn | It seems that stress test failure caused by [mis]configuration: memory limit exceeded.
Should I rebase/push to trigger test run? | 2020-09-22T17:01:37 | 2020-09-22T17:01:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,061,992 | 15,125 | l1t1 | thanks
```
[root@centos ~]# systemctl start clickhouse-server
[root@centos ~]# systemctl status clickhouse-server
● clickhouse-server.service - ClickHouse Server (analytic DBMS for big data)
Loaded: loaded (/etc/systemd/system/clickhouse-server.service; enabled; vendor preset: disabled)
Active: active (runn... | 2020-09-23T01:07:38 | 2020-09-23T01:07:38 | {} | NONE |
ClickHouse | ClickHouse | 697,151,242 | 15,179 | chang0391 | @BohuTANG @zhang2014
Thanks | 2020-09-23T05:59:06 | 2020-09-23T06:01:16 | {} | NONE |
ClickHouse | ClickHouse | 697,226,214 | 14,414 | filimonov | Stress tests DB::LazyPipeFDs::close() data race: https://clickhouse-test-reports.s3.yandex.net/14265/e2ee97202bf7b81cae3ab05458b6c4d8609c2617/stress_test_(thread)/stderr.log | 2020-09-23T08:48:20 | 2020-09-23T08:48:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,263,995 | 15,177 | KochetovNicolai | I am a little bit out of context here.
Generally, `ThreadPool` doesn't seem to be needed here because every thread executes only one task, no scheduling is needed (it is done by executor itself).
In #15035 @azat said that `wait()` call us needed, however I do not know why.
Also I remember using `ThreadPool` here ... | 2020-09-23T10:01:22 | 2020-09-23T10:01:22 | {} | MEMBER |
ClickHouse | ClickHouse | 697,328,628 | 15,191 | alesapin | Seems like it will not work for replicated tables because we commit info in zookeeper separately? | 2020-09-23T12:23:47 | 2020-09-23T12:23:47 | {} | MEMBER |
ClickHouse | ClickHouse | 697,494,720 | 15,198 | emli | It works. | 2020-09-23T14:58:21 | 2020-09-23T14:58:21 | {} | NONE |
ClickHouse | ClickHouse | 696,725,974 | 14,526 | alexey-milovidov | The answer is: open the commit like here: https://github.com/ClickHouse/ClickHouse/pull/13394/commits/15be6a0dd59075099ef35112ec94b153f80ace26 and look at the list of tags.
(in this example, we see that the minimum version with this commit is 20.7)
It works if commit was merged, not cherry-picked or rebased.
Ano... | 2020-09-22T13:35:27 | 2020-09-22T13:36:18 | {} | MEMBER |
ClickHouse | ClickHouse | 696,518,253 | 13,613 | grmorozov | I have the same problem with service cannot restart if some table has DEFAULT dictGetUInt8(...). Also it worked for me on older version (19.13) with flag _dictionaries_lazy_load_ set to false. Found related discussion at [#4717](https://github.com/ClickHouse/ClickHouse/issues/4717). | 2020-09-22T05:39:51 | 2020-09-22T17:14:30 | {} | NONE |
ClickHouse | ClickHouse | 696,866,414 | 12,567 | simPod | I read tests but don't very understand what is this good for 🤔 | 2020-09-22T17:28:32 | 2020-09-22T17:28:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 696,881,004 | 15,122 | alexey-milovidov | Changelog has been published. | 2020-09-22T17:53:34 | 2020-09-22T17:53:34 | {} | MEMBER |
ClickHouse | ClickHouse | 696,941,280 | 14,686 | yonesko | Every time when what?
I have Kafka engine that consumes protobuf topic and mat view on it.
To my experience I had to restart to reread the schema | 2020-09-22T19:46:42 | 2020-09-22T19:46:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,133,065 | 13,953 | hodgesrm | This is a great topic. It's also important to relate the requirements to specific use cases that we know users value. Here's a short list of some of the cases we've seen with our users.
- Single instance protection. ClickHouse runs in many embedded as well as standalone applications. Backups are the only viable ... | 2020-09-23T04:53:21 | 2020-09-23T04:53:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,180,934 | 14,534 | andrea-s | My two cents: in an analytics setting, it is *extremely* useful to be able to redefine a column on the fly - as it allows to build cleaning logic in place without altering the overarching data model.
Example (oversimplified):
1) customer is consuming the column "country" as part of their analysis
2) customer comp... | 2020-09-23T07:15:33 | 2020-09-23T07:17:08 | {} | NONE |
ClickHouse | ClickHouse | 697,244,596 | 15,181 | simPod | Traces dump https://gist.github.com/simPod/578ef00a3874a199df1cd4dbd2fb13b9 | 2020-09-23T09:23:13 | 2020-09-23T09:23:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,169,759 | 15,158 | akuzm | Yandex check broken in master.
Integration test reported https://github.com/ClickHouse/ClickHouse/issues/14414
AST Fuzzer -- cannot reproduce, looks similar to https://github.com/ClickHouse/ClickHouse/issues/14974 | 2020-09-23T06:49:40 | 2020-09-23T06:49:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,172,883 | 14,962 | akuzm | Integration test failures reported https://github.com/ClickHouse/ClickHouse/issues/14414
Stress test broken in master.
Yandex check doesn't like distributed test, need to disable it somehow. | 2020-09-23T06:57:02 | 2020-09-23T06:57:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,306,197 | 15,152 | maqroll | @tavplubix failing checks seem unrelated. | 2020-09-23T11:35:14 | 2020-09-23T11:47:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,335,698 | 15,190 | den-crane | ```
create table foo (`bar` Nested(id String,outcome UInt8) ) Engine=MergeTree order by tuple();
show create table foo;
┌─statement─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│ CREATE TABLE dw.foo
(
`bar.id` Arr... | 2020-09-23T12:37:05 | 2020-09-23T12:37:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,343,510 | 15,191 | alesapin | Ok, so we have to add some tests for this case? | 2020-09-23T12:51:17 | 2020-09-23T12:51:17 | {} | MEMBER |
ClickHouse | ClickHouse | 696,739,188 | 15,119 | alesapin | This is a known mutation bug in the early 20.3 versions. Just update to the latest 20.3 version. | 2020-09-22T13:57:00 | 2020-09-22T13:57:00 | {} | MEMBER |
ClickHouse | ClickHouse | 696,993,319 | 15,169 | alexey-milovidov | LGTM. No test needed. | 2020-09-22T21:31:12 | 2020-09-22T21:31:12 | {} | MEMBER |
ClickHouse | ClickHouse | 697,066,305 | 15,021 | bohutang | A1:
The code for the PARTITION clause conversion strategy from MySQL to ClickHouse is here:
https://github.com/ClickHouse/ClickHouse/blob/master/src/Interpreters/MySQL/InterpretersMySQLDDLQuery.cpp#L254-L282
4294967/18446744073709551/18446744073709551 is the max value of UInt[*] divided by 1000.
A2:
The curren... | 2020-09-23T01:21:37 | 2020-09-23T01:31:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 697,104,468 | 15,134 | amosbird | It's useful because of nested subqueries in views. | 2020-09-23T03:05:48 | 2020-09-23T03:05:48 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 697,271,687 | 15,180 | alesapin | Currently, ClickHouse doesn't delete empty parts from MergeTree for several reasons. It's expected behavior. However, they can be merged with non-empty parts in the background. We already have Draft PR where we try to delete empty parts https://github.com/ClickHouse/ClickHouse/pull/13252. | 2020-09-23T10:17:21 | 2020-09-23T10:17:21 | {} | MEMBER |
ClickHouse | ClickHouse | 697,272,173 | 15,180 | alesapin | > The outdated parts were removed in the background. The expired parts were replaced by empty parts with commited state, which generated in the merge phase. If the generated empty parts were not commited, and just remove the outdated parts, this problem will be solved.
>
> I have fixed this problem in my test env. C... | 2020-09-23T10:18:29 | 2020-09-23T10:18:29 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 697,323,578 | 15,190 | worenga | @den-crane I do not agree that
`bar` Nested( id String, outcome UInt8),
is syntactic sugar and could be replaced by
`bar.id` Array(String),
`bar.outcome` Array(UInt8),
Reason being is that clickhouse ensures that the length of the two arrays are the same at insertion with nested wh... | 2020-09-23T12:13:16 | 2020-09-23T12:15:33 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 674,365,546 | 13,735 | alexey-milovidov | 01060_shutdown_table_after_detach - #13746 | 2020-08-15T08:01:10 | 2020-08-15T08:01:10 | {} | MEMBER |
ClickHouse | ClickHouse | 674,365,854 | 13,723 | alexey-milovidov | Fuzzer: https://github.com/ClickHouse/ClickHouse/issues/13750 | 2020-08-15T08:03:43 | 2020-08-15T08:03:43 | {} | MEMBER |
ClickHouse | ClickHouse | 674,393,084 | 13,746 | alexey-milovidov | > test_storage_rabbitmq
Known bad test. | 2020-08-15T12:52:42 | 2020-08-15T12:52:42 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.