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 | 643,088,144 | 6,549 | blinkov | @thyn, do you have any further questions? | 2020-06-12T06:19:04 | 2020-06-12T06:19:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,267,650 | 11,625 | den-crane | >when the shard needs to visit other shards,it can decrypt firstly
Decrypt how? Using another password stored in /etc/secret.aes ?
You can configure passwordless schema and restrict access by ip addresses
For example it opens access to a user `default` from only these 3 ip / network.
```
cat /etc/clickh... | 2020-06-12T13:21:59 | 2020-06-12T13:21:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,654,913 | 11,590 | alexey-milovidov | Please note that the original tcmalloc also was not working well without some tuning:
https://github.com/ClickHouse/ClickHouse/commit/ca9c4bbdbb4c79e6cc945d53282cd1a4acdfb22c
Actually it was working well but degraded after some release.
This is very similar to what happened with jemalloc:
https://github.com/jemal... | 2020-06-13T17:41:36 | 2020-06-13T17:47:37 | {} | MEMBER |
ClickHouse | ClickHouse | 643,707,442 | 7,746 | alexey-milovidov | @KochetovNicolai It's your TODO in code. | 2020-06-14T02:08:56 | 2020-06-14T02:08:56 | {} | MEMBER |
ClickHouse | ClickHouse | 642,618,638 | 11,581 | alexey-milovidov | @vzakaznikov The only issue is that they can possibly overflow to 0. We should not return 0 exit code on error, it's a bug. | 2020-06-11T12:43:39 | 2020-06-11T12:43:39 | {} | MEMBER |
ClickHouse | ClickHouse | 643,286,530 | 11,630 | den-crane | Usual issue with other databases: in case col_string = 1 need to cast col_string to Int and this breaks index access. | 2020-06-12T14:01:59 | 2020-06-12T14:01:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,650,224 | 11,640 | alexey-milovidov | Short answer is that OPTIMIZE and SELECT FINAL may behave in a different way if data is incomplete.
For more detailed answer, let's ask @KochetovNicolai | 2020-06-13T17:03:20 | 2020-06-13T17:03:20 | {} | MEMBER |
ClickHouse | ClickHouse | 642,605,182 | 11,590 | alexey-milovidov | By the way, your flame graph is not from perf test, because it is:
- reversed;
- contains inline functions...
Let's look at flame graphs that are automatically built as a result of perftest.
(The link to `.7z` file at bottom) | 2020-06-11T12:13:33 | 2020-06-11T12:13:55 | {} | MEMBER |
ClickHouse | ClickHouse | 642,607,685 | 11,590 | alexey-milovidov | [svg.zip](https://github.com/ClickHouse/ClickHouse/files/4764675/svg.zip)
| 2020-06-11T12:19:25 | 2020-06-11T12:19:25 | {} | MEMBER |
ClickHouse | ClickHouse | 642,815,377 | 7,572 | alexey-milovidov | @filimonov There is a use case - to build limited API without access to default handlers.
We can:
- provide example directly in config.xml on how to configure default handlers;
- add support for `<default/>` specification on the same level as `<rule>` to automatically enable all default handlers. | 2020-06-11T17:07:11 | 2020-06-11T17:07:11 | {} | MEMBER |
ClickHouse | ClickHouse | 643,674,741 | 11,565 | azat | Looks too tricky, I'm guess simple round-robin will be enough? | 2020-06-13T20:33:21 | 2020-06-13T20:33:21 | {} | MEMBER |
ClickHouse | ClickHouse | 642,607,147 | 11,590 | azat | >Maybe it is accidentially compiled with wrong flags?
Yes this definitely should be verified
>By the way, your flame graph is not from perf test, because it is:
It was built based on the output.7z from the performance test, using [script](https://gist.github.com/akuzm/3f06cb15c7093d823554a14f1a9a1fcd) by @akuz... | 2020-06-11T12:18:09 | 2020-06-11T12:18:09 | {} | MEMBER |
ClickHouse | ClickHouse | 642,624,388 | 11,401 | akuzm | > * >
>
>
> Indeed interesting, and this is because when force purging uses mmap, it is [no-op if overcommit is enabled](https://github.com/jemalloc/jemalloc/blob/3e19ebd2ea5372c2f5932af6bb268ae8cb5df354/src/pages.c#L268-L269) (and I guess you have defaults, i.e. heuristic overcommit)
Oh, I missed this i... | 2020-06-11T12:49:16 | 2020-06-11T12:49:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,653,807 | 11,576 | filimonov | #10128 p.2 | 2020-06-11T13:36:21 | 2020-06-11T13:36:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,690,995 | 10,212 | den-crane | hm,
tuple is a WA ?
create table xt(A Date) engine=MergeTree partition by toYYYYMM(A) order by A;
insert into xt values(today());
alter table xt drop partition **tuple**(toYYYYMM(today()));
0 rows in set. Elapsed: 0.001 sec.
select * from xt
0 rows in set. Elapsed: 0.001 sec.
| 2020-06-11T14:21:57 | 2020-06-11T14:22:08 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,716,815 | 7,572 | tavplubix | I will add more configuration examples and ability to configure `ReplicasStatusHandler` and `PrometheusRequestHandler`. | 2020-06-11T14:59:55 | 2020-06-11T14:59:55 | {} | MEMBER |
ClickHouse | ClickHouse | 642,931,164 | 11,612 | akuzm | Stateless tests have passed, will merge now. | 2020-06-11T21:14:10 | 2020-06-11T21:14:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,279,819 | 11,626 | den-crane | By default the replication is password-less so malicious replica can steal your parts (data).
You can use `interserver_http_credential` or restrict access using `networks` for default user
```
cat /etc/clickhouse-server/conf.d/user_default.xml
<?xml version="1.0" ?>
<yandex>
<users>
<default>
... | 2020-06-12T13:48:17 | 2020-06-12T15:16:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,695,060 | 11,590 | azat | >There is a feature request to add CMake to tcmalloc: google/tcmalloc#4
>As you have added it in this PR, you can also contribute it to upstream.
Yep, I'm going to submit this (and also some patches against upstream should be submitted too), but first this should be done in a cleaner way (and I need to dig into abs... | 2020-06-13T23:51:49 | 2020-06-13T23:51:49 | {} | MEMBER |
ClickHouse | ClickHouse | 642,604,301 | 11,590 | alexey-milovidov | Why do we see something like https://github.com/abseil/abseil-cpp/blob/master/absl/base/internal/errno_saver.h in flame graph? This function should be 100% inlined. Maybe it is accidentially compiled with wrong flags? | 2020-06-11T12:11:40 | 2020-06-11T12:11:40 | {} | MEMBER |
ClickHouse | ClickHouse | 642,727,879 | 7,572 | filimonov | OK.
Actually there is one more option - add standard handers to the handers from config always at the end.
So user will be able to overwrite them by his own handlers which will have higher priority.
For example, if the user will decide to remove `/ping` handler he will just configure.
```
<rule>
... | 2020-06-11T15:14:09 | 2020-06-11T15:15:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,044,273 | 5,183 | adadfa2 | A large number of the following logs are printed on our logs, what are the solutions to solve this
```
2020.06.12` 11:12:13.865814 [ 138025 ] {} <Error> ServerErrorHandler: Poco::Exception. Code: 1000, e.code() = 104, e.displayText() = Connection reset by peer, Stack trace (when copying this message, always include... | 2020-06-12T03:26:35 | 2020-06-12T03:35:40 | {} | NONE |
ClickHouse | ClickHouse | 643,110,382 | 11,569 | AlexanderChiuluvB | > This .yml file is generated on fly. Please, show full logs for pytest and your env variables.
Thanks, I have solve the problem. After the compilation, I change the config dirs and binary dirs in cluster.py. And it works. | 2020-06-12T07:15:14 | 2020-06-12T07:15:14 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 643,232,011 | 11,629 | jackpgao | It seems releated with `columns.txt`
```sql
DB::Exception: File not found: /data1/clickhouse/data/xxxxx/20181224_1101_3227_17/columns.txt.tmp. Stack trace:
0. Poco::FileNotFoundException::FileNotFoundException(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, int) @ ... | 2020-06-12T11:57:28 | 2020-06-12T11:58:14 | {} | NONE |
ClickHouse | ClickHouse | 643,516,595 | 11,638 | azat | Cc: @zhang2014 | 2020-06-12T23:09:04 | 2020-06-12T23:09:04 | {} | MEMBER |
ClickHouse | ClickHouse | 643,704,893 | 9,899 | alexey-milovidov | It does not reproduce in master with trivial example:
```
milovidov-desktop :) SET prefer_localhost_replica = 0
SET prefer_localhost_replica = 0
Ok.
0 rows in set. Elapsed: 0.001 sec.
milovidov-desktop :) SELECT hostName() AS h, count() FROM cluster(test_cluster_two_shards, system.one) GROUP BY h;
S... | 2020-06-14T01:30:53 | 2020-06-14T01:30:53 | {} | MEMBER |
ClickHouse | ClickHouse | 643,705,172 | 8,665 | alexey-milovidov | Reproduced on master. | 2020-06-14T01:34:51 | 2020-06-14T01:34:51 | {} | MEMBER |
ClickHouse | ClickHouse | 642,627,749 | 11,588 | akuzm | * PVS studio doesn't like the `getLock` in context for some reason -- a false positive
* Performance test is broken in master | 2020-06-11T12:55:22 | 2020-06-11T12:55:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,637,462 | 11,588 | akuzm | * Doc check -- Ivan Blinkov says that he fixed it in master. | 2020-06-11T13:10:00 | 2020-06-11T13:10:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,662,479 | 11,597 | SaltTan | Duplicate https://github.com/ClickHouse/ClickHouse/issues/10655 | 2020-06-11T13:49:19 | 2020-06-11T13:49:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,684,510 | 7,572 | filimonov | > > @azat - but standard paths are still there:
>
> I see, but it is under `else`, so with enabled `http_handlers` they do not work
Indeed. That's crap.
We need to push standard handers (ping / replica_status / prometheus) before user-defined handlers always. | 2020-06-11T14:15:03 | 2020-06-11T14:15:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,029,958 | 11,618 | 1818hero | In addition, using LIMIT BY sometimes can change the case, but why? | 2020-06-12T02:29:24 | 2020-06-12T02:29:24 | {} | NONE |
ClickHouse | ClickHouse | 642,645,480 | 11,401 | alexey-milovidov | > test_mysql_protocol/test.py::test_mysql_federated
Known flaky test, reported. | 2020-06-11T13:24:05 | 2020-06-11T13:24:05 | {} | MEMBER |
ClickHouse | ClickHouse | 642,679,579 | 7,572 | azat | >@azat - but standard paths are still there:
I see, but it is under `else`, so with enabled `http_handlers` they do not work | 2020-06-11T14:07:59 | 2020-06-11T14:07:59 | {} | MEMBER |
ClickHouse | ClickHouse | 642,699,525 | 11,594 | den-crane | Using cast to Nullable.
```
SELECT
cast(toString(number % 2), 'Nullable(String)') AS a,
cast(toString(number % 3), 'Nullable(String)') AS b,
count()
FROM numbers(10000)
GROUP BY
a,
b
WITH CUBE
┌─a─┬─b─┬─count()─┐
│ 0 │ 2 │ 1667 │
│ 1 │ 2 │ 1666 │
│ 1 │ 1 │ 1667 │
│ ... | 2020-06-11T14:36:07 | 2020-06-11T14:36:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,703,407 | 7,572 | tavplubix | Yes, user may want to configure default handlers other way and it can be done with custom handlers, so current behaviour seems ok to me. | 2020-06-11T14:42:00 | 2020-06-11T14:42:00 | {} | MEMBER |
ClickHouse | ClickHouse | 642,846,491 | 11,484 | alexey-milovidov | @azat `01300_client_save_history_when_terminated`
No idea why it fails. It is my recently added test, if we will not find any solution, I will simply remove it. | 2020-06-11T18:07:57 | 2020-06-11T18:08:11 | {} | MEMBER |
ClickHouse | ClickHouse | 643,287,737 | 11,288 | alexey-milovidov | \+ Undiagnosed error with secondary index from one of Yandex Cloud clients (from private chat). | 2020-06-12T14:04:31 | 2020-06-12T14:04:37 | {} | MEMBER |
ClickHouse | ClickHouse | 643,653,200 | 11,631 | alexey-milovidov | If you think I should continue working on this, add "thumbs up" reaction. | 2020-06-13T17:27:57 | 2020-06-13T17:27:57 | {} | MEMBER |
ClickHouse | ClickHouse | 642,811,292 | 11,401 | azat | >Oh, I missed this if... But how does it free the memory then? It does free it somehow, I checked :) (both virt and rss).
Yes it does :)
Freeing pages is another process (should be this https://github.com/jemalloc/jemalloc/blob/5.2.1/src/extent.c#L1865), force purge just makes pages cleaning, clean page can be reus... | 2020-06-11T17:02:18 | 2020-06-11T17:02:18 | {} | MEMBER |
ClickHouse | ClickHouse | 642,811,798 | 10,624 | alexey-milovidov | It's also easy to read CSV with backslash escaping in ClickHouse with `CustomSeparated` format.
It is controlled by the settings:
format_custom_escaping_rule
format_custom_field_delimiter
format_custom_row_before_delimiter
format_custom_row_after_delimiter
format_custom_row_between_delimiter
format_custom_resu... | 2020-06-11T17:03:22 | 2020-06-11T17:03:22 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 642,817,093 | 11,484 | azat | 01300_client_save_history_when_terminated - https://clickhouse-test-reports.s3.yandex.net/11498/5c72b02f86a1e483c5220aac7d572809f53f4b78/functional_stateless_tests_(unbundled).html#fail1 | 2020-06-11T17:10:49 | 2020-06-11T17:10:49 | {} | MEMBER |
ClickHouse | ClickHouse | 642,881,042 | 11,484 | alexey-milovidov | 00721_force_by_identical_result_after_merge_zookeeper:
https://clickhouse-test-reports.s3.yandex.net/0/e354bbc78c86caed22feb052699ee896484face6/functional_stateless_tests_(debug).html#fail1
| 2020-06-11T19:21:18 | 2020-06-11T19:21:18 | {} | MEMBER |
ClickHouse | ClickHouse | 642,885,403 | 11,613 | azat | Looks like the following should be added ci fest? It does not looks like PR is related to me.
01305_replica_create_drop_zookeeper MSAN https://clickhouse-test-reports.s3.yandex.net/11613/176ebb26e0db3555efbfbf6554b327be5f352263/functional_stateless_tests_(memory).html#fail1
| 2020-06-11T19:30:48 | 2020-06-11T19:30:48 | {} | MEMBER |
ClickHouse | ClickHouse | 643,478,697 | 11,639 | alexey-milovidov | TODO: cleanupThread. | 2020-06-12T20:52:58 | 2020-06-12T20:52:58 | {} | MEMBER |
ClickHouse | ClickHouse | 643,520,404 | 11,590 | azat | Simple test with threads shows that it is slower (2x), but before making conclusions the results should be analyzed more first, so I'm still digging. | 2020-06-12T23:25:57 | 2020-06-12T23:26:43 | {} | MEMBER |
ClickHouse | ClickHouse | 643,653,084 | 11,631 | alexey-milovidov | Not worth doing. | 2020-06-13T17:26:46 | 2020-06-13T17:26:46 | {} | MEMBER |
ClickHouse | ClickHouse | 642,981,087 | 11,480 | 4ertus2 | It looks like I've broken CROSS to INNER JOIN rewrite here when disabled column match. In case when we have not qualified names. It should be fixed in #11513 | 2020-06-11T23:31:12 | 2020-06-11T23:31:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,060,707 | 11,618 | den-crane | What CH version do you use?
Can you show `show create table tablename` ?
It was an issue https://github.com/ClickHouse/ClickHouse/issues/10566
| 2020-06-12T04:40:01 | 2020-06-12T04:40:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,341,306 | 11,630 | filimonov | Yep, that why I'm saying only of allowing cast of literal (i.e. constants used in the query) to column type, and never opposite (that sounds relatively safe, but maybe there are some other issues). | 2020-06-12T15:40:40 | 2020-06-12T15:40:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,561,963 | 11,628 | zhang2014 | > Build failed without build log. It's inner CI problem. Maybe something wrong with docker or host. Just wait for build restart. | 2020-06-13T03:35:32 | 2020-06-13T03:35:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,653,025 | 11,151 | alexey-milovidov | There were two issues with the test:
1. `couner` - typo.
2. Using the substring "fail" in test name makes looking at the CI results inconvenient.
Fixed in d8312d0f00 | 2020-06-13T17:26:10 | 2020-06-13T17:26:10 | {} | MEMBER |
ClickHouse | ClickHouse | 643,747,448 | 11,565 | nvartolomei | This gets complicated, but a more flexible solution might be nested replica lists/groups.
```xml
<shard>
<internal_replication>true</internal_replication>
<replica_list load_balancing="in_order">
<!-- pick first AZ if it is available, otherwise failback to second AZ
the config for nodes in the ... | 2020-06-14T10:26:50 | 2020-06-14T10:29:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,663,973 | 11,569 | alesapin | This `.yml` file is generated on fly. Please, show full logs for pytest and your env variables. | 2020-06-11T13:50:27 | 2020-06-11T13:50:27 | {} | MEMBER |
ClickHouse | ClickHouse | 642,671,003 | 7,572 | filimonov | @azat - but standard paths are still there:
https://github.com/ClickHouse/ClickHouse/pull/10547/files#diff-5981515e3ab857757e2ccc763c560d2cR121-R128
https://github.com/ClickHouse/ClickHouse/blob/aeac8cb6212ce3658a5ffb99116ddedf3910ac65/programs/server/HTTPHandlerFactory.cpp#L121-L135
or you mean that it can be ov... | 2020-06-11T13:56:50 | 2020-06-11T13:59:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,702,686 | 11,432 | alexey-milovidov | > Or remove after checks, but only if there is no data in table.
Ok. This variant is implemented.
> I suppose the correct solution is to not create any metadata on disk before checks.
There is some strange complicated logic - we have to do some checks after initialization of some part of the metadata... and th... | 2020-06-11T14:40:44 | 2020-06-11T14:40:44 | {} | MEMBER |
ClickHouse | ClickHouse | 642,768,055 | 11,605 | alesapin | The most important part is to add processing hook for `push` event. | 2020-06-11T15:57:48 | 2020-06-11T15:57:48 | {} | MEMBER |
ClickHouse | ClickHouse | 642,852,050 | 11,484 | azat | @alexey-milovidov the first though that pops up in my mind when I first saw was "why not just check the history file, instead of using expect?" this will make test simpler (and set `CLICKHOUSE_HISTORY_FILE` to force the file location)
*Besides this test will not work in parallel (due to one common history file).* | 2020-06-11T18:19:46 | 2020-06-11T18:19:46 | {} | MEMBER |
ClickHouse | ClickHouse | 642,966,763 | 2,045 | jondeandres | @bufuchangfeng any update on this? Thanks | 2020-06-11T22:40:16 | 2020-06-11T22:40:16 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 643,676,815 | 11,352 | alexey-milovidov | @vitlibar It's a simple task - just disable optimization if row-level security present. | 2020-06-13T20:52:29 | 2020-06-13T20:52:29 | {} | MEMBER |
ClickHouse | ClickHouse | 643,711,124 | 11,656 | Yuntao-Wu | @qoega
There are a lot of errors in the document of Chinese version. I can fix it in my spare time. | 2020-06-14T03:06:02 | 2020-06-14T03:08:00 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,590,147 | 6,533 | filimonov | > detach/attach MV does not solve a problem of not missing/loosing messages in case of multiple MV.
Detach/attach table with engine=Kafka is the best workaround / without described side effects. | 2020-06-11T11:45:05 | 2020-06-11T11:45:05 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,708,633 | 11,409 | den-crane | Unclear how to reproduce the issue. It's needed clarification. | 2020-06-11T14:48:24 | 2020-06-11T14:48:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,038,811 | 5,183 | winter7 | @filimonov can you please tell me what does `Code: 1000, e.code() = 104` means? I found my server keep printing this error and stack trace like this (btw, I'm in eks)
```
2020.06.12 02:59:11.229712 [ 498 ] {} <Error> ServerErrorHandler: Poco::Exception. Code: 1000, e.code() = 104, e.displayText() = Connection reset b... | 2020-06-12T03:04:36 | 2020-06-12T04:10:45 | {} | NONE |
ClickHouse | ClickHouse | 643,393,307 | 11,247 | den-crane | Check inodes
df -i | 2020-06-12T17:20:40 | 2020-06-12T17:20:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,623,536 | 11,590 | alexey-milovidov | Maybe we can create an isolated test case with a huge number of randomly distributed middle sized allocations from miltiple threads? | 2020-06-11T12:47:43 | 2020-06-11T12:47:43 | {} | MEMBER |
ClickHouse | ClickHouse | 642,729,437 | 7,572 | filimonov | And i think the more common use case is to extend standatd handlers, not replace all of them. | 2020-06-11T15:16:35 | 2020-06-11T15:16:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,736,300 | 10,624 | kokes | @shivakumarss You can tell Spark to escape quotes using a second quote, instead of the non-standard backslash. That way you'll get correctly exported data, which you can then import to CH as well as other tools. I [wrote about this](https://kokes.github.io/blog/2018/05/19/spark-sane-csv-processing.html) and recommended... | 2020-06-11T15:28:01 | 2020-06-11T15:28:01 | {} | NONE |
ClickHouse | ClickHouse | 642,784,846 | 11,606 | CurtizJ | 3 different tests instead of 1, because setting `min_insert_block_size_rows` can't be set in scope of query and in performnace tests session settings can be set only for whole test. It's worth to test part formats' dependence of amount of data to insert. | 2020-06-11T16:19:01 | 2020-06-11T16:19:01 | {} | MEMBER |
ClickHouse | ClickHouse | 642,821,967 | 11,484 | alexey-milovidov | NULL pointer in
2020.06.11 18:12:51.946411 [ 14042 ] {} BaseDaemon: 5. /build/obj-x86_64-linux-gnu/../src/Storages/MergeTree/MergeTreeData.cpp:1640: DB::MergeTreeData::PartsTemporaryRename::tryRenameAll() @ 0x10f0f2bb in /usr/bin/clickhouse
https://github.com/ClickHouse/ClickHouse/pull/11601 | 2020-06-11T17:21:04 | 2020-06-11T17:21:04 | {} | MEMBER |
ClickHouse | ClickHouse | 643,359,155 | 11,634 | filimonov | 19.14-20.4 is also affected, didn't test further. | 2020-06-12T16:17:07 | 2020-06-12T16:25:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,728,483 | 11,624 | wallace-clickhouse | Thank you very much!
For my situation, I expect to call clickhouse-client by java code,so I think it's difficult for me to use prompted way to input the password,is there any other way for me to achieve it? | 2020-06-14T07:07:55 | 2020-06-14T07:07:55 | {} | NONE |
ClickHouse | ClickHouse | 642,819,061 | 11,601 | alexey-milovidov | PVS Ok. | 2020-06-11T17:14:47 | 2020-06-11T17:14:47 | {} | MEMBER |
ClickHouse | ClickHouse | 642,918,567 | 11,585 | alesapin | Ok, tests passed | 2020-06-11T20:43:23 | 2020-06-11T20:43:23 | {} | MEMBER |
ClickHouse | ClickHouse | 643,108,437 | 11,347 | TCeason | Oh, I wanna try to do it. | 2020-06-12T07:10:22 | 2020-06-12T07:10:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,265,552 | 11,624 | den-crane | You can use prompted password input
```
cl -u test --password -q "select 1"
Password for user (test):
```
And you can hide a password in `ps` using shell variables.
And cli-client probably should hide password in ps by itself https://unix.stackexchange.com/questions/88665/how-does-ps-know-to-hide-passwo... | 2020-06-12T13:17:25 | 2020-06-12T13:17:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,707,419 | 9,899 | sundy-li | Maybe you should try it with storage `ReplicatedMergeTree`
```
dn592.jja :) SELECT hostName() AS h, count() FROM cluster(ck_cluster_preonline, default, hdfs_router_audit) GROUP BY h
SELECT
hostName() AS h,
count()
FROM cluster(ck_cluster_preonline, default, hdfs_router_audit)
GROUP BY h
Re... | 2020-06-14T02:08:32 | 2020-06-14T02:08:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,608,453 | 11,590 | alexey-milovidov | Maybe it is just slower on middle-sized allocations (from a few KB to a few MB)? | 2020-06-11T12:21:21 | 2020-06-11T12:21:47 | {} | MEMBER |
ClickHouse | ClickHouse | 642,618,183 | 11,590 | akuzm | > Why do we see something like https://github.com/abseil/abseil-cpp/blob/master/absl/base/internal/errno_saver.h in flame graph? This function should be 100% inlined. Maybe it is accidentially compiled with wrong flags?
Looks like it is inlined -- the file is errno_saver, but the function is AbslInternalSpinLockDela... | 2020-06-11T12:42:40 | 2020-06-11T12:42:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,621,060 | 11,401 | akuzm | Experiment no. 5: no MADV_FREE/DONTNEED at all.
Surprisingly, it's the fastest: [results](https://clickhouse-test-reports.s3.yandex.net/11401/e2301f9ca38e914484fe91d871821647737e496c/performance_comparison/report.html#fail1) | 2020-06-11T12:45:48 | 2020-06-11T12:45:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,697,476 | 11,597 | den-crane | @kainever7 try to upgrade to 20.3.9.70 | 2020-06-11T14:32:23 | 2020-06-11T14:32:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,091,920 | 11,409 | gubinjie | @den-crane
Very sorry, did not see your reply.
Because this process is also a long-term process, it feels slower in long-term use, and there is no specific performance in the log.
You can pay attention to it for a long time, because we use quartz to run the same script every day, and the calculated time is differen... | 2020-06-12T06:29:54 | 2020-06-12T06:29:54 | {} | NONE |
ClickHouse | ClickHouse | 643,392,703 | 11,409 | den-crane | No idea. It could be anything. Maybe number of partitions -> parts grows over time. | 2020-06-12T17:19:13 | 2020-06-12T17:19:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,561,875 | 11,638 | zhang2014 | LGTM | 2020-06-13T03:34:29 | 2020-06-13T03:34:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,692,639 | 11,590 | alexey-milovidov | There is a feature request to add CMake to tcmalloc: https://github.com/google/tcmalloc/issues/4
As you have added it in this PR, you can also contribute it to upstream. | 2020-06-13T23:31:45 | 2020-06-13T23:31:45 | {} | MEMBER |
ClickHouse | ClickHouse | 643,729,373 | 11,626 | wallace-clickhouse | Great! I like password-less communication beetween replications,haha...
However why we don't use passdword-less way also for the data exchange among shards,it leads to a problem that I have to configure a plaintext password in metrika.xml,it's really not secure.
Thank you! | 2020-06-14T07:18:28 | 2020-06-14T07:18:28 | {} | NONE |
ClickHouse | ClickHouse | 642,681,059 | 11,505 | den-crane | No info. | 2020-06-11T14:10:20 | 2020-06-11T14:10:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,700,828 | 11,594 | den-crane | it will be addressed in https://github.com/ClickHouse/ClickHouse/issues/10540 | 2020-06-11T14:38:28 | 2020-06-11T14:38:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 642,848,349 | 11,563 | azat | >@azat Deadlock introduced in this PR: https://gist.github.com/alesapin/d915d1deaa27d49aa31223daded02be2
@alesapin Thanks! Will get back with a fix | 2020-06-11T18:12:04 | 2020-06-11T18:12:04 | {} | MEMBER |
ClickHouse | ClickHouse | 642,898,883 | 11,613 | alexey-milovidov | @azat #11614 | 2020-06-11T20:00:58 | 2020-06-11T20:00:58 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 642,901,224 | 11,614 | alexey-milovidov | Need that:
https://clang.llvm.org/extra/clang-tidy/checks/readability-implicit-bool-conversion.html
| 2020-06-11T20:06:37 | 2020-06-11T20:06:37 | {} | MEMBER |
ClickHouse | ClickHouse | 643,122,870 | 11,623 | sundy-li | LGTM | 2020-06-12T07:39:35 | 2020-06-12T07:39:35 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,389,611 | 11,071 | r-zenine | Hi @Akazz,
I made all the changes :).
Thanks for your feedback.
Have a good day. | 2020-06-12T17:11:33 | 2020-06-12T17:11:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,676,658 | 11,630 | alexey-milovidov | We have to implement it for SQL compatibility.
But it is not in development.
It will be more handy if we can postpone it after implementation of #3775 but that task is abandoned, and we should find simpler solution.
Example: a setting to enable cast of constants in comparison operators.
Note that it was already... | 2020-06-13T20:51:07 | 2020-06-13T20:51:28 | {} | MEMBER |
ClickHouse | ClickHouse | 642,697,072 | 7,572 | alexey-milovidov | If I remember correctly, custom HTTP handlers can be also used to disable or replace default HTTP handlers. It makes sense because it allows to provide limited access to the server. And to keep default HTTP handlers along with custom, you have to add them to configuration.
@tavplubix am I right? | 2020-06-11T14:31:39 | 2020-06-11T14:31:39 | {} | MEMBER |
ClickHouse | ClickHouse | 642,703,534 | 7,572 | filimonov | I though default configuration lists a way how to configure `/ping` & `/replica_status` to return them back, but it's not. For `/replicas_status` it seems not possible, same for /prometeus | 2020-06-11T14:42:10 | 2020-06-11T14:42:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 643,292,140 | 11,475 | rtarasov-mg | We found why is the problem with dictionaries update
New variable in mysql 8
This is introduced in 8.0 and not present in 5.7
```
mysql> show variables like 'information_schema_stats_expiry';
+---------------------------------+-------+
| Variable_name | Value |
+------------------------------... | 2020-06-12T14:13:35 | 2020-06-12T14:13:35 | {} | NONE |
ClickHouse | ClickHouse | 643,492,278 | 11,639 | alexey-milovidov | TODO: support downgrade scenario. | 2020-06-12T21:33:33 | 2020-06-12T21:33:33 | {} | MEMBER |
ClickHouse | ClickHouse | 646,688,699 | 8,639 | alesapin | 1. `02000_optimize_skip_unused_shards_type_mismatch` -- `Unknown setting force_optimize_skip_unused_shards`
2. `01254_array_of_unnamed_tuples` -- `Unknown table function generateRandom`
3. `01249_bad_arguments_for_bloom_filter` -- formatting of create query has changed - ok
4. `01245_limit_infinite_sources` -- `Ther... | 2020-06-19T15:09:15 | 2020-06-19T15:09:15 | {} | MEMBER |
ClickHouse | ClickHouse | 646,788,421 | 11,805 | alexey-milovidov | Results on the diagram are not sorted by the value. | 2020-06-19T17:46:25 | 2020-06-19T17:46:25 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.