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 | 692,478,986 | 14,830 | filimonov | > How can i grant privileges about SOURCES?
Update clickhouse version. 20.3 should not be used for RBAC | 2020-09-15T05:45:38 | 2020-09-15T05:45:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,500,535 | 14,524 | bohutang | Already on it , but there are some CI issues to deal, https://github.com/ClickHouse/ClickHouse/pull/14564
| 2020-09-15T06:38:56 | 2020-09-15T06:38:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,362,535 | 14,827 | vitlibar | > Looks like some tests are not checking that the view is actually auto-deleted.
It seems to me that the test `00980_create_temporary_live_view` checks exactly that a live view is deleted automatically after the timeout is passed. | 2020-09-14T22:59:43 | 2020-09-15T08:40:05 | {} | MEMBER |
ClickHouse | ClickHouse | 692,784,369 | 14,814 | bobrik | > In principle, we keep the history in master in a shape that allows to bisect on `--first-parent` to skip all these commits, but the git support for this is abysmal.
TIL, will keep that in mind for the first time.
So I made this change to figure out where the first initialization happens:
```diff
$ git diff ... | 2020-09-15T15:14:50 | 2020-09-15T15:14:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,344,990 | 13,156 | vitlibar | > rebased history is ready, waiting for perf tests to push.
You can push it right now, pushing will restart the tests anyway.
> Or I can submit it via another PR (so that you can diff them), what you will prefer?
Let's stay in this PR. | 2020-09-14T22:22:43 | 2020-09-14T22:22:54 | {} | MEMBER |
ClickHouse | ClickHouse | 692,505,056 | 14,609 | alesapin | 00652_replicated_mutations_zookeeper flaky, reported in CI Fest https://github.com/ClickHouse/ClickHouse/issues/14414. | 2020-09-15T06:49:38 | 2020-09-15T06:49:38 | {} | MEMBER |
ClickHouse | ClickHouse | 692,598,717 | 14,781 | bohutang | From the underlying, this issue duplicated https://github.com/ClickHouse/ClickHouse/issues/14786
it's payload.available() caused, and on the fixing | 2020-09-15T09:38:58 | 2020-09-15T09:38:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,621,222 | 14,749 | hexiaoting | new PR for this : https://github.com/ClickHouse/ClickHouse/pull/14842 | 2020-09-15T10:18:39 | 2020-09-15T10:19:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,637,306 | 14,841 | filimonov | wrong repo (you can open issue in https://github.com/ClickHouse/ClickHouse-odbc )
| 2020-09-15T10:51:47 | 2020-09-15T10:51:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,837,824 | 14,814 | bharatnc | @bobrik thank you for the traces, I remember struggling with the ThreadPoolConfig quite a bit when I was implementing this feature with respect to initialization. But like akuzm mentions, I'm surprised that the tests did not catch this. | 2020-09-15T16:45:18 | 2020-09-15T16:49:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,379,618 | 14,782 | l1t1 | putty is xterm ,securecrt was vt100, change to xterm too, show command ok, thanks | 2020-09-14T23:58:13 | 2020-09-14T23:58:13 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 692,436,273 | 14,807 | whelan9453 | Hi
I delete the cluster and rebuild it with the same config. This time it works.
I guess for some reason my previous cluster didn't use the new config. | 2020-09-15T03:12:25 | 2020-09-15T03:12:25 | {} | NONE |
ClickHouse | ClickHouse | 692,458,123 | 4,006 | zhang2014 | > This is a very nice feature to connect Mysql(OLTP) and Clickhouse(OLAP). There is a question that is it possible to ensure that multiple rows are atomically updated to Clickhouse?
Of course. but there are some bugs that we're trying to fix. | 2020-09-15T04:35:47 | 2020-09-15T04:35:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,501,532 | 14,814 | bobrik | Computer says it's 96d2e9c997e39d38f3c520090d3659b87fda775f, but there are "bad" commits that produce "The global thread pool is not initalized" error.
If you look at my initial report, you can see:
```
Thread 1 (LWP 40705):
#0 0x00007fa2b1c56495 in __pthread_timedjoin_ex () from target:/lib/x86_64-linux-gnu/l... | 2020-09-15T06:41:18 | 2020-09-15T06:41:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,505,839 | 1,952 | Amixp | and continue
`
2020.09.15 07:57:38.241059 [ 9376 ] {} <Warning> radius.radacct: Part 202002-1_394_394_0 is covered by 202002-1_394_398_1 but should be merged into 202002-1_394_400_1. This shouldn't happen often.
2020.09.15 07:57:38.307718 [ 9416 ] {} <Warning> radius.radacct (ReplicatedMergeTreePartCheckThread): Che... | 2020-09-15T06:51:32 | 2020-09-15T06:51:32 | {} | NONE |
ClickHouse | ClickHouse | 692,612,042 | 14,840 | akuzm | We do have a SIGSEGV handler already, probably what's left is to use `sigaltstack` so that it works even on stack overflow. | 2020-09-15T10:01:29 | 2020-09-15T10:01:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,639,459 | 14,827 | vzakaznikov | Yes, the test `00980_create_temporary_live_view.sql` does check it. However, it does not do any `WATCH` queries after which the table is not being deleted as it should.
```bash
SET allow_experimental_live_view = 1;
DROP TABLE IF EXISTS lv;
DROP TABLE IF EXISTS mt;
SELECT name, value from system.settings WHER... | 2020-09-15T10:56:03 | 2020-09-15T10:57:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,927,005 | 14,839 | alexey-milovidov | The change is obvious and should not introduce this error. | 2020-09-15T19:23:41 | 2020-09-15T19:23:41 | {} | MEMBER |
ClickHouse | ClickHouse | 693,000,423 | 14,845 | alexey-milovidov | > test_materialize_mysql_database
Most likely already fixed by @zhang2014 | 2020-09-15T21:54:56 | 2020-09-15T21:54:56 | {} | MEMBER |
ClickHouse | ClickHouse | 692,278,401 | 14,821 | alexey-milovidov | Let's see what builds will be too long to have this test.
Then I'll add more examples of data usage into the test. | 2020-09-14T19:54:40 | 2020-09-14T19:54:40 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 691,787,997 | 14,749 | hexiaoting | Ok, I will try to fix this problem | 2020-09-14T03:19:47 | 2020-09-15T01:47:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,418,523 | 4,910 | sszc |
Our existing case requires that the external dictionary field can be NULL. | 2020-09-15T02:11:41 | 2020-09-15T02:11:41 | {} | NONE |
ClickHouse | ClickHouse | 692,438,312 | 14,436 | peaksnail | hi
what operation will cause this? I have do nothing on the table ,and it's locked. | 2020-09-15T03:19:38 | 2020-09-15T03:19:38 | {} | NONE |
ClickHouse | ClickHouse | 692,482,962 | 14,814 | bobrik | @bharatnc, @akuzm, any ideas? Our configs do not have an explicit configuration for `max_thread_pool_size`, since they are the same since v20.3 times. | 2020-09-15T05:56:30 | 2020-09-15T05:56:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,514,903 | 14,834 | alesapin | It was not released, so nothing affected. | 2020-09-15T07:10:25 | 2020-09-15T07:10:25 | {} | MEMBER |
ClickHouse | ClickHouse | 692,600,303 | 14,835 | leoones | > duplicated #14786
Are you talking about data duplication ? | 2020-09-15T09:41:38 | 2020-09-15T09:41:38 | {} | NONE |
ClickHouse | ClickHouse | 692,760,867 | 14,809 | zhang2014 | @alexey-milovidov I will fix it. | 2020-09-15T14:41:14 | 2020-09-15T14:41:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,351,233 | 11,913 | CurtizJ | The same about `00804_test_alter_compression_codecs`. | 2020-09-14T22:40:14 | 2020-09-14T22:40:14 | {} | MEMBER |
ClickHouse | ClickHouse | 692,386,354 | 14,814 | bobrik | An attempt at running GDB via `nsenter -m` to read debug info properly failed miserably:
```
(gdb) thread apply all bt full
Thread 6 (Thread 0x7fb68827c700 (LWP 11)):
#0 0x00007fb6d4fd67ef in epoll_wait () from /lib/x86_64-linux-gnu/libc.so.6
No symbol table info available.
#1 0x00000000130afe47 in Poco::Ne... | 2020-09-15T00:20:52 | 2020-09-15T00:20:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,424,894 | 14,829 | vegastar002 | ck version: clickhouse-common-static-20.8.2.3-2.x86_64 | 2020-09-15T02:32:19 | 2020-09-15T02:32:19 | {} | NONE |
ClickHouse | ClickHouse | 692,427,509 | 11,596 | chang0391 | @4ertus2 ,
When the function "Grace hash join" can be implemented? | 2020-09-15T02:41:07 | 2020-09-15T02:41:07 | {} | NONE |
ClickHouse | ClickHouse | 692,416,786 | 14,794 | compasses | Big thanks, seems more clear now.
In consideration of resource cost, the 'partial_merge' algorithm will have more less memory footprint, and I think usually more memory cost always cost more time. And especially under some resource limit environment.
Start from partial merge algorithm, there are many ways to opt... | 2020-09-15T02:05:48 | 2020-09-15T03:20:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,498,925 | 14,524 | leoones | @chang0391
How is this problem solved ? Is MySQL5.6 not supported ? | 2020-09-15T06:35:02 | 2020-09-15T06:35:02 | {} | NONE |
ClickHouse | ClickHouse | 692,518,580 | 14,830 | spihiker | When update 20.6 ,work well .thank you . | 2020-09-15T07:18:17 | 2020-09-15T07:18:17 | {} | NONE |
ClickHouse | ClickHouse | 692,551,499 | 14,297 | alesapin | `00600_replace_running_query` in flaky in debug? Test log https://gist.github.com/alesapin/db850cf068eb64f6a66e4603ee090b9b. | 2020-09-15T08:20:43 | 2020-09-15T08:20:43 | {} | MEMBER |
ClickHouse | ClickHouse | 692,674,274 | 14,840 | filimonov | > We do have a SIGSEGV handler already, probably what's left is to use `sigaltstack` so that it works even on stack overflow.
Yes, it seems so, the tricky part is that should happen for every thread (according to that article). | 2020-09-15T12:12:04 | 2020-09-15T12:12:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,717,842 | 14,844 | den-crane | SummingMergeTree -- does automatic summing of numeric columns, AggregatingMergeTree does not.
AggregatingMT require AggregateFunction or SimpleAggregateFunction .
eval **SimpleAggregateFunction**(sum,UInt64)
```sql
create table exp_mergetree(
id UInt64,
eval SimpleAggregateFunction(sum,UInt64), ... | 2020-09-15T13:33:53 | 2020-09-15T13:35:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,338,939 | 10,759 | yonzhang | @cw9 I have similar "query hangs" issue. Do you have any progress on solving your issue? | 2020-09-14T22:05:53 | 2020-09-14T22:05:53 | {} | NONE |
ClickHouse | ClickHouse | 692,600,879 | 14,835 | bohutang | no, issue related, all of them are same root cause, and we are fixing | 2020-09-15T09:42:39 | 2020-09-15T09:42:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,906,325 | 14,684 | azat | >This is a great feature and I would love to be able to use it. but we have some DDL generated by program(may be generated and executed simultaneously on different nodes)
And you are worrying about parallel access to one table from the DDLWorker? But this is possible via multiple clients, and protecting from this wi... | 2020-09-15T18:50:35 | 2020-09-15T18:50:35 | {} | MEMBER |
ClickHouse | ClickHouse | 692,750,389 | 14,809 | bohutang | From the out.log it seems that ClickHouse does not sync the MySQL binlog event but not sure the root cause.
By the way, I can't reproduce it on my pc.
| 2020-09-15T14:25:13 | 2020-09-15T14:25:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,751,792 | 14,414 | bohutang | test_materialize_mysql_database/test.py::test_materialize_database_ddl_with_empty_transaction_8_0
https://clickhouse-test-reports.s3.yandex.net/14809/2b113f7553f7f73c9c21a15e11999de10dfe5953/integration_tests_(asan).html#fail1 | 2020-09-15T14:27:25 | 2020-09-15T14:27:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,810,514 | 14,414 | vitlibar | test_materialize_mysql_database/test.py::test_materialize_database_ddl_with_empty_transaction_5_7
https://clickhouse-test-reports.s3.yandex.net/14809/2b113f7553f7f73c9c21a15e11999de10dfe5953/integration_tests_(asan)/integration_run.3.txt.out.log | 2020-09-15T15:57:06 | 2020-09-15T15:57:06 | {} | MEMBER |
ClickHouse | ClickHouse | 692,914,393 | 14,839 | alexey-milovidov | We have not observed failures of this test before. | 2020-09-15T19:01:24 | 2020-09-15T19:01:24 | {} | MEMBER |
ClickHouse | ClickHouse | 692,917,851 | 14,609 | alexey-milovidov | It's not a bugfix.
Even more, there is no performance improvements in gcc-10 compared to gcc-9 on our codebase. | 2020-09-15T19:05:32 | 2020-09-15T19:05:32 | {} | MEMBER |
ClickHouse | ClickHouse | 692,957,068 | 14,846 | alexey-milovidov | Would it help if we disable ThinLTO for `libglibc-compatibility`? | 2020-09-15T20:20:19 | 2020-09-15T20:20:19 | {} | MEMBER |
ClickHouse | ClickHouse | 692,269,450 | 14,747 | den-crane | >since it started.
Since what started?
The root cause is unclear. No guarantee that it's fixed in the latest versions.
Label `obsolete-version` means that an issue is not relevant (not important) until it reproduced at stable/master releases.
The current stable 20.6+ (or 20.3 LTS). | 2020-09-14T19:36:57 | 2020-09-14T19:37:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,339,098 | 6,081 | abraithwaite | We haven't observed this since, so going to close this out. | 2020-09-14T22:06:23 | 2020-09-14T22:06:23 | {} | NONE |
ClickHouse | ClickHouse | 692,552,576 | 14,806 | alesapin | I'll try to fix it in branch. | 2020-09-15T08:22:43 | 2020-09-15T08:22:43 | {} | MEMBER |
ClickHouse | ClickHouse | 692,581,120 | 14,837 | nikitamikhaylov | Thanks! But I already know about it) | 2020-09-15T09:09:21 | 2020-09-15T09:09:21 | {} | MEMBER |
ClickHouse | ClickHouse | 692,681,559 | 14,764 | mjodeh | Thanx @alexey-milovidov
9- What about using Buffer tables on cluster? is it possible to do so?
10- Also I tested the `Compact format` on the staging server and `rows=90000` and the insertion rate 2k/sec I noticed that the size it takes is far more the `Wide format`, is it correct or am I missing something here? | 2020-09-15T12:27:21 | 2020-09-15T12:27:21 | {} | NONE |
ClickHouse | ClickHouse | 692,842,747 | 14,856 | KochetovNicolai | ```sql
EXPLAIN
SELECT number
FROM numbers(10) AS f
ANY LEFT JOIN
(
SELECT number
FROM numbers(10)
) AS s USING (number)
WHERE number IN
(
SELECT number * 2
FROM numbers(10)
)
FORMAT TSV
Expression (Projection)
CreatingSets (Create sets before main query execution)
Expression (... | 2020-09-15T16:54:11 | 2020-09-15T16:54:11 | {} | MEMBER |
ClickHouse | ClickHouse | 692,266,690 | 14,684 | alexey-milovidov | Need to consider the cases when client execute ON CLUSTER queries one after the other without waiting for the previous query... is it possible? | 2020-09-14T19:31:17 | 2020-09-14T19:31:17 | {} | MEMBER |
ClickHouse | ClickHouse | 692,320,797 | 14,747 | Slind14 | Thank you. I should have done that before submitting an issue. Turns out the issue is an unexpected behavior where two records submitted with the same timestamp result in them being summed. Interestingly this wasn't the case before and is likely related to an API behavior change.
I didn't realize that we were only ... | 2020-09-14T21:20:23 | 2020-09-14T21:20:23 | {} | NONE |
ClickHouse | ClickHouse | 692,325,190 | 13,156 | vitlibar | > P.S. if you are ready to merge, please let me know, I want to make the history cleaner (or at least squash it).
I have marked this PR as approved. It does mean I'm ready to merge it :)
But I have to ensure before merging that the PR doesn't break any tests or there is a good explanation why some tests are broken.... | 2020-09-14T21:30:44 | 2020-09-14T21:31:34 | {} | MEMBER |
ClickHouse | ClickHouse | 692,375,633 | 14,807 | whelan9453 | > Check errors in CH logs /var/log/clickhouse-server/
>
> Try to remove/ comment empty tags `<root></root> <identity></identity>`
I removed empty tags but still got the same error.
Check the error log and get this. It's weird that ClickHouse still try to connect ZooKeeper with the previous config.
ClickHous... | 2020-09-14T23:43:50 | 2020-09-14T23:43:50 | {} | NONE |
ClickHouse | ClickHouse | 692,787,792 | 14,834 | alesapin | Performance report https://clickhouse-test-reports.s3.yandex.net/14834/6f8da7dfdf3c0f8a9af9c973dcd5b5c87a0479ac/performance_comparison/report.html (for some reason no status) | 2020-09-15T15:20:07 | 2020-09-15T15:20:07 | {} | MEMBER |
ClickHouse | ClickHouse | 692,821,267 | 14,839 | Jokser | `test_multiple_disks/test.py::test_jbod_overflow[replicated_mt_with_overflow-ReplicatedMergeTree('/clickhouse/replicated_mt_with_overflow', '1')] | FAIL | 27.81`
The same test passed in release mode. Seems it's just flaky. Moreover, no changes related to ReplicatedMergeTree have made.
| 2020-09-15T16:15:27 | 2020-09-15T16:16:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,453,320 | 4,006 | EmmyMiao87 | This is a very nice feature to connect Mysql(OLTP) and Clickhouse(OLAP). There is a question that is it possible to ensure that multiple rows are atomically updated to Clickhouse? | 2020-09-15T04:18:14 | 2020-09-15T04:18:14 | {} | NONE |
ClickHouse | ClickHouse | 692,480,803 | 14,814 | bobrik | It's definitely something to do with the global pool:
* Good: 18eb141ea15456ecc93d5f47948c54723e461ba1
* Bad: 2da7b356ab016ce08589bd80e8ceb6748c52b42e
* Middle: `Application: DB::Exception: The global thread pool is not initalized` (typo here as well)
Good to bad git log:
```
commit 2da7b356ab016ce08589bd80... | 2020-09-15T05:50:58 | 2020-09-15T05:50:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 691,918,157 | 13,956 | alex-zaitsev | > The whole idea looks dangerous and very strange. Merges know nothing about volumes, replicas know nothing about volumes. So let's look at some cases:
>
> 1. Imagine someone forbids merges for some volume. All merges including at least one part from this volume cannot be assigned. Probably you'll have suboptimal me... | 2020-09-14T08:54:37 | 2020-09-15T06:38:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,592,244 | 14,835 | bohutang | duplicated https://github.com/ClickHouse/ClickHouse/issues/14786 | 2020-09-15T09:27:58 | 2020-09-15T09:28:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,615,965 | 14,814 | akuzm | > In principle, we keep the history in master in a shape that allows to bisect on `--first-parent` to skip all these commits, but the git support for this is abysmal.
Oh wow, they added it at last https://stackoverflow.com/questions/5638211/how-do-you-get-git-bisect-to-ignore-merged-branches/63568395#63568395 | 2020-09-15T10:08:30 | 2020-09-15T10:08:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,265,531 | 14,743 | alexey-milovidov | I've increased the priority of perf test. Let's wait for it... | 2020-09-14T19:29:11 | 2020-09-14T19:29:11 | {} | MEMBER |
ClickHouse | ClickHouse | 692,290,801 | 14,684 | azat | >Need to consider the cases when client execute ON CLUSTER queries one after the other without waiting for the previous query... is it possible?
No, this is handled in the DDLQueryStatusInputStream, and this hasn't been changed, only the DDLWorker, who executes queries not wait for execution
```sql
n1 :) SYSTEM ... | 2020-09-14T20:18:51 | 2020-09-14T20:18:51 | {} | MEMBER |
ClickHouse | ClickHouse | 692,343,367 | 11,913 | CurtizJ | ```00926_adaptive_index_granularity_pk: [ FAIL ] 670.21 sec. - Timeout!```
In this test a table with `index_granularity_bytes=40` is created, which leads to `index_granularity=1`. With compact parts this configuration is extremely inefficient, because of writing compressed block fo... | 2020-09-14T22:18:18 | 2020-09-14T22:18:18 | {} | MEMBER |
ClickHouse | ClickHouse | 692,506,275 | 1,952 | Amixp | **version 20.7.2.30 (official build))** | 2020-09-15T06:52:34 | 2020-09-15T06:52:34 | {} | NONE |
ClickHouse | ClickHouse | 692,596,106 | 14,795 | bohutang | From the root cause, this issue duplicated with https://github.com/ClickHouse/ClickHouse/issues/14786 | 2020-09-15T09:34:30 | 2020-09-15T09:34:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,596,807 | 14,829 | tavplubix | Probably related #14781 | 2020-09-15T09:35:43 | 2020-09-15T09:35:43 | {} | MEMBER |
ClickHouse | ClickHouse | 692,686,317 | 14,809 | alexey-milovidov | @zhang2014 @BohuTANG about MaterializeMySQL test ^ | 2020-09-15T12:37:26 | 2020-09-15T12:37:26 | {} | MEMBER |
ClickHouse | ClickHouse | 692,797,914 | 14,814 | bobrik | So I added this change to get stacks of `GlobalThreadPool::initialize()`:
```diff
diff --git a/src/Common/ThreadPool.cpp b/src/Common/ThreadPool.cpp
index 49516d777f..41cc744ced 100644
--- a/src/Common/ThreadPool.cpp
+++ b/src/Common/ThreadPool.cpp
@@ -278,6 +278,15 @@ void GlobalThreadPool::initialize(size_t m... | 2020-09-15T15:36:20 | 2020-09-15T15:36:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,925,432 | 14,846 | alesapin | Seems like clang-11 can't build clickhouse if ThinLTO enabled.
```
2020-09-15 13:13:07 [3969/8300] Linking CXX executable contrib/llvm/llvm/bin/llvm-tblgen
2020-09-15 13:13:07 FAILED: contrib/llvm/llvm/bin/llvm-tblgen
2020-09-15 13:13:07 : && /usr/bin/ccache /usr/bin/clang++-11 -fdiagnostics-color=always -std=c++... | 2020-09-15T19:20:21 | 2020-09-15T19:20:21 | {} | MEMBER |
ClickHouse | ClickHouse | 692,928,250 | 13,383 | alexey-milovidov | @Enmk did not answer, will look by myself. | 2020-09-15T19:26:07 | 2020-09-15T19:26:07 | {} | MEMBER |
ClickHouse | ClickHouse | 692,464,819 | 1,952 | Amixp | I have this:
`
2020.09.15 07:57:37.817047 [ 9379 ] {} <Warning> radius.radacct: Part 202002-1_394_394_0 is covered by 202002-1_394_398_1 but should be merged into 202002-1_394_400_1. This shouldn't happen often.
2020.09.15 07:57:37.883105 [ 9399 ] {} <Warning> radius.radacct (ReplicatedMergeTreePartCheckThread): Che... | 2020-09-15T04:59:51 | 2020-09-15T04:59:51 | {} | NONE |
ClickHouse | ClickHouse | 692,504,979 | 14,414 | alesapin | `00652_replicated_mutations_zookeeper` need to check test https://clickhouse-test-reports.s3.yandex.net/14609/393c213866ff56ef98ae2f66ee50afebc9459319/functional_stateless_tests_(thread).html#fail1 | 2020-09-15T06:49:29 | 2020-09-15T06:49:29 | {} | MEMBER |
ClickHouse | ClickHouse | 692,609,727 | 14,814 | akuzm | Thanks for investigating. Do you have the stack trace for the "The global thread pool is not initalized" error?
It's weird that our tests don't catch this, probably we are missing something important... | 2020-09-15T09:57:17 | 2020-09-15T09:57:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,279,249 | 14,747 | Slind14 | I looked at the records of what it returned in the past. One month ago, it never occurred that it returned twice the amount. Since last month it has been occurring very regularly. Sometimes every second time we ran the query.
```
CREATE TABLE default.qLZPsR3 (`impressionUUID` FixedString(16), [...]
ENGINE = Repli... | 2020-09-14T19:56:28 | 2020-09-14T19:57:04 | {} | NONE |
ClickHouse | ClickHouse | 692,361,025 | 11,913 | CurtizJ | Other fixed tests checked some implementation defined things like checksums or number of files or hardcoded paths to files. Also some compatibility tests, which run cluster with different versions older than 20.3, were failed, because replication protocol is incompatible with such old versions, when polymorphic parts a... | 2020-09-14T22:54:44 | 2020-09-14T22:54:44 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 640,537,780 | 3,311 | PhantomPhreak | A new fancy RBAC approach has been implemented in a newest clickhouse versions, so i had a hope that `SHOW CREATE TABLE` clauses may be explicitly granted to the user, that may help to mask a problem decribed in this issue. Unfortunately, with any grant on a table or a database, even with a simple `SELECT` grant, user ... | 2020-06-08T11:11:55 | 2020-09-15T08:10:52 | {} | NONE |
ClickHouse | ClickHouse | 692,553,370 | 14,587 | alesapin | `test_materialize_mysql_database/test.py::test_materialize_database_ddl_with_empty_transaction_8_0` Flaky test, failing in master. | 2020-09-15T08:24:08 | 2020-09-15T08:24:08 | {} | MEMBER |
ClickHouse | ClickHouse | 692,608,908 | 14,809 | 4ertus2 | Fail in test_materialize_mysql_database/test.py::test_materialize_database_ddl_with_empty_transaction_8_0 is not ralated to bigints | 2020-09-15T09:55:46 | 2020-09-15T09:55:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,695,980 | 13,859 | mergar | same here:
```
% uname -a
FreeBSD **.org 12.1-RELEASE-p7 FreeBSD 12.1-RELEASE-p7 r363767 GENERIC amd64
```
Clickhouse version: **20.10.1.4650 (official build)**
Patch solve the issue. @KochetovNicolai any plan to merge #13869 ? | 2020-09-15T12:56:47 | 2020-09-15T12:56:47 | {} | NONE |
ClickHouse | ClickHouse | 692,266,715 | 14,747 | Slind14 | We don't. In fact, we didn't make any changes to how we work with tables since it started.
Seeing that it is based on the parts data, we might be running into the same issue with select count() if we upgrade, no? | 2020-09-14T19:31:19 | 2020-09-14T19:31:19 | {} | NONE |
ClickHouse | ClickHouse | 692,288,136 | 14,750 | filimonov | Mist probably where vs prewhere issue (different columns are moved to prewhere),
or may be by the way how the condition are written (`a and (b or c)` )
Try using prewhere instead of where | 2020-09-14T20:13:36 | 2020-09-14T20:13:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,407,129 | 14,774 | Awezome | Thx for reply . I read some pages from web and now understand how it works. | 2020-09-15T01:33:29 | 2020-09-15T01:33:29 | {} | NONE |
ClickHouse | ClickHouse | 692,445,242 | 14,684 | zhang2014 | > P.S. with pool_size=1, everything will be executed sequentially
That's great.
> So you suggesting to serialize access to the (database, table), yes it will disallow parallel queries via ON CLUSTER for it, but w/o ON CLUSTER parallel access via multiple clients is still possible right now, so does it worth it? (... | 2020-09-15T03:46:46 | 2020-09-15T03:46:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,474,887 | 14,814 | filimonov | Is zookeeper healthy, and network ok? Is there some relevant info in ZooKeeper logs? | 2020-09-15T05:33:02 | 2020-09-15T05:33:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,614,450 | 14,814 | akuzm | > Thanks for investigating. Do you have the stack trace for the "The global thread pool is not initalized" error?
>
> It's weird that our tests don't catch this, probably we are missing something important...
On a second thought, probably I don't need it if it doesn't happen in master -- these are some intermedia... | 2020-09-15T10:05:36 | 2020-09-15T10:05:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,804,862 | 14,609 | bobrik | Is it expected to be backported into v20.9? | 2020-09-15T15:47:31 | 2020-09-15T15:47:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,838,101 | 14,809 | bobrik | Is this going to be backported to v20.9? | 2020-09-15T16:45:49 | 2020-09-15T16:45:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,912,985 | 14,809 | alexey-milovidov | Most likely not, as it is not a bugfix. | 2020-09-15T19:00:32 | 2020-09-15T19:00:32 | {} | MEMBER |
ClickHouse | ClickHouse | 692,261,997 | 14,747 | den-crane | @Slind14 do you swap tables using `rename table xxx to xxx_old, xxx_new to xxx` ?
Starting from 19.17 `select count()` is optimized by metadata (the same as system.parts) https://github.com/ClickHouse/ClickHouse/pull/7510 | 2020-09-14T19:22:27 | 2020-09-14T19:25:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,288,628 | 14,747 | den-crane | Let's isolate the issue from your code.
Can you show an output of this command:
```
for i in `seq 1 3600`; do echo -n `date`" "; clickhouse-client -q "select sum(rows) nrows from system.parts where active and table in ('qLZPsR3')"; sleep 1; done;
```
executed during one hour or less if it shows nrows jumping. | 2020-09-14T20:14:33 | 2020-09-14T20:44:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,341,006 | 13,156 | azat | >I have marked this PR as approved. It does mean I'm ready to merge it :)
Ok, great!
rebased history is ready, waiting for perf tests to push.
Or I can submit it via another PR (so that you can diff them), what you will prefer? | 2020-09-14T22:11:33 | 2020-09-14T22:11:33 | {} | MEMBER |
ClickHouse | ClickHouse | 692,360,584 | 14,827 | vzakaznikov | @vitlibar, the https://github.com/ClickHouse/ClickHouse/commit/70b0fe88b1c6394f4415c020c39d28ac2fc5704a caused regression :(.
Looks like some tests are not checking that the view is actually auto-deleted.
```
./00991_temporary_live_view_watch_live.sh.disabled
./00991_temporary_live_view_watch_live.reference
.... | 2020-09-14T22:53:18 | 2020-09-14T22:53:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,443,587 | 14,814 | bobrik | Seems like bisecting is the best way forward here, I'm on it.
So far at `v20.6.1.3965-testing-138-geefb498517-dev` we're still good, ~10 steps to go:
```
$ git bisect good
Bisecting: 1011 revisions left to test after this (roughly 10 steps)
[eefb498517a16c346061c07b5cc5280b21c0c68b] [docs] refactor Domains ove... | 2020-09-15T03:40:07 | 2020-09-15T03:40:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,439,417 | 14,831 | bharatnc | cc: @tavplubix - while changes in this PR are partly related to PR #14760 - made quick changes to construct `query_start_time` and `query_start_time_microseconds` from the same timespec with tests. | 2020-09-15T03:24:03 | 2020-09-15T07:01:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 692,582,467 | 14,628 | tavplubix | `Stress test (memory) — Killed by signal` - reported | 2020-09-15T09:11:29 | 2020-09-15T09:11:29 | {} | MEMBER |
ClickHouse | ClickHouse | 700,243,710 | 15,383 | filimonov | > I/O is falling during insertion
That is expected. You filled page cache, and/or you you create lot of parts (see parts_to_delay_insert).
For the page cache - you can't write faster than your disks can write the data in long term, and than clickhouse can merge new parts.
You can only decrease the number of... | 2020-09-28T19:46:23 | 2020-09-28T19:46:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 700,368,396 | 15,383 | qinzl1 | @filimonov
thx，i will tried the new version | 2020-09-29T01:22:26 | 2020-09-29T01:22:26 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.