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 | 816,441,752 | 18,856 | alexey-milovidov | No info provided. | 2021-04-09T06:23:12 | 2021-04-09T06:23:12 | {} | MEMBER |
ClickHouse | ClickHouse | 816,463,761 | 22,831 | godliness | @alexey-milovidov @filimonov
@den-crane
Any help will be appreciated. | 2021-04-09T07:06:14 | 2021-04-09T08:34:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,577,027 | 22,893 | heneyin | > Upgrade to version 21.3 is recommended.
The current version is the latest version of 20.3, 20.3.21 | 2021-04-09T10:09:33 | 2021-04-09T10:09:33 | {} | NONE |
ClickHouse | ClickHouse | 816,660,247 | 21,942 | akuzm | Flaky check complains about the test being long, probably because generation of stack traces is slow. I think we can keep it as is because the run time is almost acceptable and I'm not sure how to speed it up. | 2021-04-09T12:52:48 | 2021-04-09T12:52:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,782,598 | 22,885 | abyss7 | What is a purpose for GCC support on MacOS? We don't check it in our CI and it will break eventually.
Also @alexey-milovidov I don't think that it's a good idea to make changes to submodules just for compliance with GCC+Darwin compilability - we'll have to maintain those changes in future when doing merges with upstre... | 2021-04-09T15:58:06 | 2021-04-09T15:58:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,824,529 | 22,885 | traceon | @abyss7 any macos build breaks eventually as far as compilation is considered, because cross-compiling for macos is far from reproducing the real environment - you can see that by looking at the trend of the reported issues for mac compilation. The purpose for this PR is mainly to make sure that the code is portable ag... | 2021-04-09T17:06:05 | 2021-04-09T17:06:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,828,576 | 22,885 | traceon | > Yes, I know, that's why we suggest to use LLVM based version in our instruction.
By "vanilla" I mean exactly LLVM, usually installed through Homebrew package manager. In macOS, the native considered to be AppleClang. | 2021-04-09T17:13:00 | 2021-04-09T17:13:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,923,745 | 22,831 | filimonov | https://github.com/ClickHouse/ClickHouse/issues/10461#issuecomment-623006030 | 2021-04-09T19:36:39 | 2021-04-09T19:36:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,040,266 | 22,436 | sasha-id | Got it, I think I read your article regarding this, using LIVE VIEW as caching layer. | 2021-04-10T00:29:11 | 2021-04-10T00:29:11 | {} | NONE |
ClickHouse | ClickHouse | 817,154,185 | 22,697 | TCeason | Integration tests (thread) — Timeout, fail: 0, passed: 1248, flaky: 0
The timeout means? @tavplubix | 2021-04-10T15:31:32 | 2021-04-10T15:32:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,447,253 | 22,855 | alexey-milovidov | Our tasks are started on servers with 64 GB RAM (msan) and 48 GB RAM (tsan). | 2021-04-09T06:34:41 | 2021-04-09T06:34:41 | {} | MEMBER |
ClickHouse | ClickHouse | 816,564,464 | 22,894 | alesapin | Ok, stress passed | 2021-04-09T09:48:09 | 2021-04-09T09:48:09 | {} | MEMBER |
ClickHouse | ClickHouse | 816,574,794 | 22,910 | kitaisreal | In this version right size is used for bytes_allocated sparse_hashed dictionaries. | 2021-04-09T10:05:38 | 2021-04-09T10:05:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,977,569 | 22,945 | alesapin | > @alesapin Is it possible to leave only actual download operation in pool and move ZooKeeper interaction outside of the pool?
But where it should be? If we will perform all ZK operations in a single (task assigning) thread it will be even slower. What is the idea? | 2021-04-09T21:19:37 | 2021-04-09T21:19:37 | {} | MEMBER |
ClickHouse | ClickHouse | 817,040,304 | 22,436 | vzakaznikov | You are most welcome! | 2021-04-10T00:29:22 | 2021-04-10T00:29:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,124,770 | 20,237 | alexey-milovidov | @gratcliff On Apple M1, instead of running x86_64 build under Rosetta (it does not work as shown in this issue)
AArch64 build of ClickHouse should be run.
@filimonov Could you please tell how to run AArch64 build of ClickHouse on Apple M1? | 2021-04-10T11:52:55 | 2021-04-10T11:52:55 | {} | MEMBER |
ClickHouse | ClickHouse | 817,191,815 | 22,960 | kitaisreal | @azat I removed test 01802_dict_sparse_hashed_bytes_allocated.sql because it does check that we use right method to calculate size of sparse hash table, but check overall size of hashed dictionary, and it will be inconvenient to update it every time internal implementation will change. What do you think ? | 2021-04-10T19:33:14 | 2021-04-10T19:33:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,365,295 | 22,696 | amosbird | I've successfully built a working hash function in java that aligns perfectly with clickhouse (murmur32). The critical part is to check integer promotion rules during bit manipulation, which is different between java and c++. | 2021-04-09T02:56:43 | 2021-04-09T02:56:43 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 816,382,072 | 12,656 | zhang2014 | Ok, let's temporarily close it. | 2021-04-09T03:52:38 | 2021-04-09T03:52:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,575,089 | 22,927 | alexey-milovidov | Please fix the documentation accordingly. | 2021-04-09T10:06:07 | 2021-04-09T10:06:07 | {} | MEMBER |
ClickHouse | ClickHouse | 816,575,649 | 22,913 | kitaisreal | In this version right size is used for bytes_allocated sparse_hashed dictionaries. | 2021-04-09T10:07:05 | 2021-04-09T10:07:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,601,052 | 22,928 | alexey-milovidov | Could you please add the documentation? | 2021-04-09T10:57:11 | 2021-04-09T10:57:11 | {} | MEMBER |
ClickHouse | ClickHouse | 816,646,343 | 22,609 | Avogar | We should add the needed submodule in this list:
https://github.com/ClickHouse/ClickHouse/blob/30f2d839c34208db69132e063ba669aacebbcd2c/docker/test/fasttest/run.sh#L160
I will do it. | 2021-04-09T12:26:51 | 2021-04-09T13:32:49 | {
"heart": 1
} | MEMBER |
ClickHouse | ClickHouse | 816,735,560 | 12,990 | qoega | similar issue with test_storage_kerberized_kafka
https://github.com/ClickHouse/ClickHouse/pull/12991#issuecomment-816734602 | 2021-04-09T14:45:55 | 2021-04-09T14:45:55 | {} | MEMBER |
ClickHouse | ClickHouse | 816,770,054 | 22,762 | abyss7 | Can't see in this PR where and how do you use new creator variant, can you show please? | 2021-04-09T15:38:08 | 2021-04-09T15:38:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,888,811 | 22,824 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/16260 | 2021-04-09T18:50:16 | 2021-04-09T18:50:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,039,612 | 22,436 | sasha-id | Thank you for detail explanation! | 2021-04-10T00:25:58 | 2021-04-10T00:25:58 | {} | NONE |
ClickHouse | ClickHouse | 817,191,116 | 21,639 | traceon | Thanks, @changvvb . Could you please also merge this: https://github.com/changvvb/ClickHouse/pull/3 . This is my final change, hopefully. | 2021-04-10T19:27:23 | 2021-04-10T19:27:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,051,368 | 22,679 | ianton-ru | I hope this PR fixes the bug.
https://github.com/ClickHouse/ClickHouse/pull/22864 | 2021-04-08T18:38:28 | 2021-04-08T18:38:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,065,101 | 22,726 | azat | @CurtizJ can you take a look? | 2021-04-08T18:56:33 | 2021-04-08T18:56:33 | {} | MEMBER |
ClickHouse | ClickHouse | 816,446,041 | 22,855 | alexey-milovidov | OOMs in stress test:
```
milovidov@milovidov-desktop:~/Downloads$ grep 'Fatal' clickhouse-server\(3\).log
2021.04.08 16:56:25.238665 [ 298 ] {} <Fatal> Application: Child process was terminated by signal 9 (KILL). If it is not done by 'forcestop' command or manually, the possible cause is OOM Killer (see 'dmesg' ... | 2021-04-09T06:32:08 | 2021-04-09T06:32:08 | {} | MEMBER |
ClickHouse | ClickHouse | 816,785,932 | 6,555 | alesapin | Implemented in `ALTER` on top of mutations. | 2021-04-09T16:02:51 | 2021-04-09T16:02:51 | {} | MEMBER |
ClickHouse | ClickHouse | 817,133,816 | 21,639 | traceon | @changvvb could you please sync with `master` one more time? | 2021-04-10T13:05:18 | 2021-04-10T13:05:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,176,798 | 22,852 | alexey-milovidov | CC @Enmk | 2021-04-08T20:44:21 | 2021-04-08T20:44:21 | {} | MEMBER |
ClickHouse | ClickHouse | 816,449,920 | 22,855 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/pull/22894 | 2021-04-09T06:40:13 | 2021-04-09T06:40:13 | {} | MEMBER |
ClickHouse | ClickHouse | 816,536,699 | 22,833 | kitaisreal | AST fuzzer https://github.com/ClickHouse/ClickHouse/issues/22857.
Performance is problem with flat_dictionary tests. | 2021-04-09T09:02:13 | 2021-04-09T09:02:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,625,266 | 22,709 | gyuton | Internal documentation ticket: DOCSUP-8406. | 2021-04-09T11:46:10 | 2021-04-09T11:46:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,646,229 | 22,886 | vzakaznikov | Another fix was merged. Closing. | 2021-04-09T12:26:37 | 2021-04-09T12:26:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,771,366 | 22,760 | abyss7 | It would be great if you add your own integration test for this improvement. | 2021-04-09T15:40:17 | 2021-04-09T15:40:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,924,727 | 22,831 | filimonov | See also https://github.com/ClickHouse/ClickHouse/pull/22673 | 2021-04-09T19:38:41 | 2021-04-09T19:38:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,610,187 | 21,888 | Avogar | @rf Now I see the problem deltaSumTimestamp should solve. I didn't get it before, sorry. If `deltaSumTimestamp` can really help you, I will review the code and merge it. But I have and idea how to solve your problem without `deltaSumTimestamp`. Try such query:
```sql
select bucket, name, deltaSumMerge(s_deltasum), de... | 2021-04-09T11:14:56 | 2021-04-09T11:14:56 | {} | MEMBER |
ClickHouse | ClickHouse | 816,620,137 | 22,899 | vitlibar | This fix depends on fix https://github.com/ClickHouse/ClickHouse/pull/19020 which is not in 20.8 | 2021-04-09T11:35:03 | 2021-04-09T11:35:03 | {} | MEMBER |
ClickHouse | ClickHouse | 816,734,602 | 12,991 | qoega | Still getting an error on master
https://clickhouse-test-reports.s3.yandex.net/0/f7edcdf7c847f7d45ceede4a6073400ec2c985b7/integration_tests_(memory).html
```
Uninitialized bytes in __interceptor_getaddrinfo at offset 20 inside [0x7febe03d1ec0, 48)
==7==WARNING: MemorySanitizer: use-of-uninitialized-value
#0 0x... | 2021-04-09T14:44:31 | 2021-04-09T14:44:31 | {} | MEMBER |
ClickHouse | ClickHouse | 817,184,573 | 21,942 | gyuton | Internal documentation ticket: DOCSUP-8425. | 2021-04-10T18:34:21 | 2021-04-10T18:34:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,458,186 | 22,609 | pingyu | The failure of build seems to be caused by DataSketches submodule is not checked out.
How can I fix it ? @Avogar | 2021-04-09T06:55:58 | 2021-04-09T06:55:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,698,031 | 22,853 | vdimir | Backported to 21.3+ | 2021-04-09T13:52:16 | 2021-04-09T13:52:16 | {} | MEMBER |
ClickHouse | ClickHouse | 817,040,040 | 17,623 | zhicwu | May also consider `clickhouse-data-bridge` proposed by @ramazanpolat at ClickHouse/clickhouse-jdbc-bridge#97. Would be nice to start with a new table function and engine with enhanced protocol to support key features like: 1) pushdown hints; and 2) context passing(current user for access control etc.). | 2021-04-10T00:28:05 | 2021-04-10T00:28:05 | {
"heart": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,174,737 | 20,102 | rrelenschi | > @rrelenschi I tried to reproduce your example but it works fine:
>
> ```
> $ clickhouse-client
> ClickHouse client version 21.5.1.1.
> Connecting to localhost:9000 as user default.
> Connected to ClickHouse server version 21.5.1 revision 54447.
>
> milovidov-desktop :) CREATE DATABASE mydatabase
>
> CRE... | 2021-04-10T17:26:15 | 2021-04-10T17:26:15 | {} | NONE |
ClickHouse | ClickHouse | 816,600,831 | 22,928 | alexey-milovidov | Not supposed to work for now.
`toStartOf*` from `DateTime64` argument returns `DateTime` or `Date`.
The `DateTime` data type does not support extended range. | 2021-04-09T10:56:44 | 2021-04-09T10:56:44 | {} | MEMBER |
ClickHouse | ClickHouse | 817,038,499 | 20,237 | gratcliff | Is there any plan to provide this? Apple M1s are becoming more prevalent and this would be fantastic to support! | 2021-04-10T00:20:25 | 2021-04-10T00:20:25 | {
"+1": 9
} | NONE |
ClickHouse | ClickHouse | 817,126,139 | 22,945 | alexey-milovidov | Ok.
In previous versions task assignment was also performed from multiple threads.
But it was more complicated and now it is better. | 2021-04-10T12:04:23 | 2021-04-10T12:05:09 | {} | MEMBER |
ClickHouse | ClickHouse | 817,137,624 | 21,639 | changvvb | > @changvvb could you please sync with `master` one more time?
Done
| 2021-04-10T13:33:40 | 2021-04-10T13:33:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,153,386 | 20,019 | alexey-milovidov | It looks very easy to fix. | 2021-04-10T15:25:47 | 2021-04-10T15:25:47 | {} | MEMBER |
ClickHouse | ClickHouse | 816,620,041 | 22,898 | vitlibar | This fix depends on fix https://github.com/ClickHouse/ClickHouse/pull/19020 which is not in 21.1 | 2021-04-09T11:34:50 | 2021-04-09T11:34:50 | {} | MEMBER |
ClickHouse | ClickHouse | 816,962,900 | 22,945 | alexey-milovidov | @alesapin Is it possible to leave only actual download operation in pool and move ZooKeeper interaction outside of the pool? | 2021-04-09T20:47:22 | 2021-04-09T20:47:22 | {} | MEMBER |
ClickHouse | ClickHouse | 817,150,551 | 22,609 | pingyu | Comments addressed. PTAL, thanks~ @Avogar | 2021-04-10T15:05:34 | 2021-04-10T15:05:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,153,289 | 20,019 | alexey-milovidov | @akuzm I get different error:
```
Illegal type Nullable(Nothing) of column for filter. Must be UInt8 or Nullable(UInt8)..
``` | 2021-04-10T15:25:09 | 2021-04-10T15:25:09 | {} | MEMBER |
ClickHouse | ClickHouse | 816,825,231 | 22,885 | traceon | Also, vanilla clang in macos is just equally unpopular and marginal as gcc. | 2021-04-09T17:07:14 | 2021-04-09T17:07:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,858,209 | 22,885 | abyss7 | > There are many of them, one that comes to my mind first is `mig` (Mach Interface Generator) not available in Linux, required for Kerberos compilation under mac. There are many discrepancies in things that can be formally compiled too. It is strange that you recommend (if you do) such cross-compiled binary for product... | 2021-04-09T18:01:27 | 2021-04-09T18:01:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,038,875 | 22,436 | vzakaznikov | Hi @alex-klepa, efficiency will depend on how fast it takes to execute the query itself and the overall insert rate. If the result can be calculated fast you can use polling method and resend the query every second. Now, if you have 100's of clients sending the exact same query then a LIVE VIEW might be much more effic... | 2021-04-10T00:22:12 | 2021-04-10T00:22:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,039,801 | 22,436 | vzakaznikov | If don't need real-time updates then you can also try a solution using LIVE VIEW table but not using WATCH query. Instead you can use plain select on the LIVE VIEW. In this case what will happened is that you will still use polling method and resend your query to the LIVE VIEW every second. Now if the insert rate is no... | 2021-04-10T00:26:55 | 2021-04-10T00:26:55 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,124,377 | 22,947 | alexey-milovidov | (changed to Improvement, because of this is a fix to marginal feature) | 2021-04-10T11:50:16 | 2021-04-10T11:50:16 | {} | MEMBER |
ClickHouse | ClickHouse | 817,151,536 | 20,102 | alexey-milovidov | @rrelenschi I tried to reproduce your example but it works fine:
```
$ clickhouse-client
ClickHouse client version 21.5.1.1.
Connecting to localhost:9000 as user default.
Connected to ClickHouse server version 21.5.1 revision 54447.
milovidov-desktop :) CREATE DATABASE mydatabase
CREATE DATABASE mydatabas... | 2021-04-10T15:12:14 | 2021-04-10T15:12:14 | {} | MEMBER |
ClickHouse | ClickHouse | 817,152,180 | 12,559 | alexey-milovidov | All issues have been fixed in recent releases. | 2021-04-10T15:17:05 | 2021-04-10T15:17:05 | {} | MEMBER |
ClickHouse | ClickHouse | 816,575,161 | 22,912 | kitaisreal | In this version right size is used for bytes_allocated sparse_hashed dictionaries. | 2021-04-09T10:06:14 | 2021-04-09T10:06:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,964,928 | 22,885 | abyss7 | > I don't think anyone will ever try cross-compiling for macOS on their own, if any other way of building it natively exist. People will rather report a problem as an issue in github, than resort to cross-compiling. In this sense, the coverage that is in the CI right now doesn't really cover most of the use cases.
S... | 2021-04-09T20:51:51 | 2021-04-09T20:51:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,041,356 | 22,273 | vzakaznikov | > @vzakaznikov May I ask, do you have any clue about this issue?
That version is "quite old". Do you have ability to try the same action on a newer release as mentioned above? | 2021-04-10T00:34:54 | 2021-04-10T00:34:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,145,656 | 22,926 | alexey-milovidov | For some reason, our generated HTML documentation goes through Jinja. Looks like a mistake. | 2021-04-10T14:31:31 | 2021-04-10T14:31:31 | {} | MEMBER |
ClickHouse | ClickHouse | 816,697,647 | 22,849 | vdimir | Backported to 21.3+ | 2021-04-09T13:51:41 | 2021-04-09T13:51:41 | {} | MEMBER |
ClickHouse | ClickHouse | 816,920,975 | 22,933 | filimonov | https://github.com/ClickHouse/ClickHouse/issues/13826 ? | 2021-04-09T19:31:03 | 2021-04-09T19:31:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,966,830 | 22,885 | alexey-milovidov | I agree - these fixes will eventually break and we cannot support them.
But this PR is still good to merge. At least to make more developers happy.
The amount of harm (unsupported modifications in libraries) is minimal and it's easy to revert if needed. | 2021-04-09T20:55:48 | 2021-04-09T20:55:48 | {} | MEMBER |
ClickHouse | ClickHouse | 816,625,457 | 22,866 | gyuton | Internal documentation ticket: DOCSUP-8406. | 2021-04-09T11:46:35 | 2021-04-09T11:46:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,826,293 | 22,885 | abyss7 | > Also, vanilla clang in macos is just equally unpopular and marginal as gcc.
Yes, I know, that's why we suggest to use LLVM based version in our instruction. | 2021-04-09T17:09:01 | 2021-04-09T17:09:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,832,904 | 22,885 | traceon | > Can you point out any example of difference in compilation on real Darwin vs cross-build using LLVM Clang?
There are many of them, one that comes to my mind first is `mig` (Mach Interface Generator) not available in Linux, required for Kerberos compilation under mac. There are many discrepancies in things that can... | 2021-04-09T17:19:59 | 2021-04-09T17:19:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,867,598 | 21,191 | abyss7 | It would be great if you provide exact logs and errors about inability to build MySQL in your particular configuration. | 2021-04-09T18:14:50 | 2021-04-09T18:14:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,953,309 | 22,885 | traceon | > The point is - to stick with a single supported way of building Darwin version of ClickHouse, which we can afford to test with our CI.
I don't think anyone will ever try cross-compiling for macOS on their own, if any other way of building it natively exist. People will rather report a problem as an issue in github... | 2021-04-09T20:26:41 | 2021-04-09T20:27:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,119,904 | 22,885 | alexey-milovidov | > AST fuzzer (UBSan) — ../src/Common/FieldVisitors.h:193:22: runtime error: 3.40282e+38 is outside the range of representable values of type 'unsigned long'
#22943 | 2021-04-10T11:16:16 | 2021-04-10T11:16:16 | {} | MEMBER |
ClickHouse | ClickHouse | 817,132,686 | 3,889 | alexey-milovidov | @den-crane
SQLite has IS and IS NOT operators that can compare not only with NULL:
```
x IS 1
x IS y
x IS NOT y
``` | 2021-04-10T12:56:50 | 2021-04-10T12:56:50 | {} | MEMBER |
ClickHouse | ClickHouse | 817,152,781 | 15,420 | alexey-milovidov | CREATE TABLE statement is flushing data to disk with `fsync`. It can trigger flushing of other dirty pages from OS page cache to disk. Which can be arbitrary slow (multiple munites and even hours) if you are using:
- USB sticks;
- network volumes (EBS and similar);
- malfunctioning HDD;
and similar. | 2021-04-10T15:21:23 | 2021-04-10T15:22:02 | {} | MEMBER |
ClickHouse | ClickHouse | 816,327,984 | 22,868 | CurtizJ | Example of test report https://clickhouse-test-reports.s3.yandex.net/22868/add449750283f066449da78afd4f8f358d8f5a45/integration_tests_(release).html. | 2021-04-09T00:55:39 | 2021-04-09T00:55:39 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 816,536,423 | 22,857 | kitaisreal | ```
CREATE TABLE table_map (a Map(String, String), b String) ENGINE = MergeTree() ORDER BY a;
INSERT INTO table_map values ({'name':'zhangsan', 'gender':'male'}, 'name'), ({'name':'lisi', 'gender':'female'}, 'gender');
SELECT toDecimal32OrZero(NULL, '0.03', NULL), b FROM table_map WHERE a = map('name', NULL, '', NUL... | 2021-04-09T09:01:40 | 2021-04-09T09:01:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,570,019 | 22,914 | kitaisreal | In this version right size is used for bytes_allocated sparse_hashed dictionaries. | 2021-04-09T09:57:34 | 2021-04-09T09:57:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,697,871 | 22,851 | vdimir | Backported to 21.3+ | 2021-04-09T13:52:01 | 2021-04-09T13:52:01 | {} | MEMBER |
ClickHouse | ClickHouse | 816,798,527 | 22,609 | Avogar | You also should support case when ENABLE_DATASKETCHES=0 in code. Now it just cannot build with it. Put all logic with submodule under `if USE_DATASKETCHES` and throw an exception when it's not enabled and uniqThetaSketch is called.
I enabled it by default in fasttest to see other tests. Maybe we should disable it in t... | 2021-04-09T16:22:36 | 2021-04-09T16:23:35 | {} | MEMBER |
ClickHouse | ClickHouse | 816,934,376 | 21,850 | akuzm | fuzzer https://github.com/ClickHouse/ClickHouse/issues/22943
functional stateless tests https://github.com/ClickHouse/ClickHouse/pull/22944 | 2021-04-09T19:58:39 | 2021-04-09T19:58:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 817,150,326 | 22,893 | alexey-milovidov | We no longer provide support for version 20.3.
The support period for version 20.3 was one year and it has ended.
Upgrading is the only option. | 2021-04-10T15:04:24 | 2021-04-10T15:04:24 | {} | MEMBER |
ClickHouse | ClickHouse | 816,535,915 | 22,893 | alexey-milovidov | Upgrade to version 2**1**.3 is recommended. | 2021-04-09T09:00:54 | 2021-04-10T15:04:58 | {} | MEMBER |
ClickHouse | ClickHouse | 816,327,723 | 22,868 | CurtizJ | Internal part of this task https://nda.ya.ru/t/gCZpsCoR3rkTRF. | 2021-04-09T00:54:47 | 2021-04-09T00:54:47 | {} | MEMBER |
ClickHouse | ClickHouse | 816,471,519 | 18,215 | alesapin | `test_keeper_snapshots_multinode/test.py::test_restart_multinode` flaky :( Will try to fix. | 2021-04-09T07:19:43 | 2021-04-09T07:19:43 | {} | MEMBER |
ClickHouse | ClickHouse | 816,573,933 | 22,911 | kitaisreal | In this version right size is used for bytes_allocated sparse_hashed dictionaries. | 2021-04-09T10:04:17 | 2021-04-09T10:04:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,625,377 | 22,687 | gyuton | Internal documentation ticket: DOCSUP-8406. | 2021-04-09T11:46:26 | 2021-04-09T11:46:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,770,692 | 22,928 | zvonand | > Could you please add the documentation?
https://github.com/ClickHouse/ClickHouse/pull/22935 | 2021-04-09T15:39:08 | 2021-04-09T15:39:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,835,669 | 22,885 | traceon | So a solution to inability to cover all platforms seems to be sticking to writing portable code, and building it in a portable way. This PR is improvement on that front. | 2021-04-09T17:24:19 | 2021-04-09T17:24:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,866,785 | 21,867 | abyss7 | What is your `ld` version is? | 2021-04-09T18:13:28 | 2021-04-09T18:13:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 816,960,045 | 22,806 | alexey-milovidov | As soon as there are no drivers for Dart, you can use ClickHouse GRPC interface.
Here is an example of the usage from Python (hope it will help):
https://github.com/ClickHouse/ClickHouse/tree/master/utils/grpc-client
HTTP interface can also be used. | 2021-04-09T20:40:59 | 2021-04-09T20:41:24 | {} | MEMBER |
ClickHouse | ClickHouse | 817,017,360 | 16,798 | zhicwu | @zhanglistar, [this](https://github.com/ClickHouse/clickhouse-jdbc/pull/596) is the PR. You may use the feature in jdbc driver 0.3.0. | 2021-04-09T23:03:55 | 2021-04-09T23:03:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 819,246,388 | 23,027 | azat | >[Functional stateless tests flaky check (address) — fail: 2, passed: 198](https://clickhouse-test-reports.s3.yandex.net/23027/ed789fc4d945819d8eed8950ee1a0009b4a2253e/functional_stateless_tests_flaky_check_(address).html#fail1)
The problem is that query can still be cancelled already:
```
2021-04-13 23:22:38 --... | 2021-04-14T05:48:12 | 2021-04-14T05:48:12 | {} | MEMBER |
ClickHouse | ClickHouse | 819,440,254 | 23,037 | alexey-milovidov | > Is it useful to integrate clickhouse with ldap in a production cluster?
https://altinity.com/blog/integrating-clickhouse-with-ldap-part-one
https://zhuanlan.zhihu.com/p/355554683
I do have WeChat but did not use it for over a year - most likely it's frozen. | 2021-04-14T11:17:03 | 2021-04-14T11:17:03 | {} | MEMBER |
ClickHouse | ClickHouse | 819,522,574 | 23,062 | UnamedRus | Does it make sense to expand `CLUSTER PAUSE NODE` command, so such node would not participate in new Distributed queries?
It would make easier to gracefully remove nodes without affecting any running query. | 2021-04-14T13:34:44 | 2021-04-14T13:34:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 820,114,085 | 23,088 | azat | >LGTM if there will be no performance regression.
I need to collect additional metrics first, will mark it as ready for review once I will finish. | 2021-04-15T05:22:51 | 2021-04-15T05:22:51 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.