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 | 457,874,304 | 1,200 | alexey-milovidov | We have support for Aarch64 and PPC64, but we aren't going to support 32 bit architectures. | 2019-01-26T23:13:13 | 2019-01-26T23:13:13 | {} | MEMBER |
ClickHouse | ClickHouse | 457,875,550 | 1,631 | alexey-milovidov | `std::terminate` after `std::system_error, e.what() = Resource temporarily unavailable`
has been fixed in 19.1. | 2019-01-26T23:36:35 | 2019-01-26T23:36:35 | {} | MEMBER |
ClickHouse | ClickHouse | 457,973,047 | 4,047 | eagle1981 | > what's the affection of this error? Is it safe to ignore?
No visual effect on functions, only disk space wasting in log-files. | 2019-01-28T01:23:29 | 2019-01-28T01:23:29 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 458,156,978 | 4,156 | alexey-milovidov | @proller The issue should be obvious without any additional info. | 2019-01-28T14:40:04 | 2019-01-28T14:40:04 | {} | MEMBER |
ClickHouse | ClickHouse | 458,526,426 | 4,171 | alesapin | Timeout on `stateless.00411_long_accurate_number_comparison`. Not related to changes. | 2019-01-29T12:44:27 | 2019-01-29T12:44:27 | {} | MEMBER |
ClickHouse | ClickHouse | 458,540,075 | 4,171 | alesapin | Timeout `Make check` -- gcc8 + old machine. | 2019-01-29T13:31:09 | 2019-01-29T13:31:19 | {} | MEMBER |
ClickHouse | ClickHouse | 458,695,419 | 1,510 | mcarbonneaux | i encounter this error on large request (900M lines on table2) like that :
```
INSERT INTO <table1> SELECT * FROM <table2>
```
| 2019-01-29T20:30:26 | 2019-01-29T20:30:52 | {} | NONE |
ClickHouse | ClickHouse | 458,858,208 | 4,178 | alesapin | Fix merge conflicts and you will see some test results!
Also it would be better to squash commits https://stackoverflow.com/a/5189600 or merge this PR with squash. | 2019-01-30T08:38:24 | 2019-01-30T08:38:24 | {} | MEMBER |
ClickHouse | ClickHouse | 457,867,954 | 508 | alexey-milovidov | You can get rare "Cannot read all data" errors when fetching data from a replica due to network issues.
All other reasons are considered to be fixed as of now.
Please reopen if you experience recurring "Cannot read all data" errors in the latest stable version. | 2019-01-26T21:35:29 | 2019-01-26T21:35:29 | {} | MEMBER |
ClickHouse | ClickHouse | 457,871,583 | 588 | alexey-milovidov | @proller This is NOT fixed.
```
root@ca26ee8df9c2:/# ls -l /etc/clickhouse-server/
total 28
-rw-r--r-- 1 root root 16993 Jan 23 22:21 config.xml
lrwxrwxrwx 1 root root 40 Jan 26 23:31 preprocessed -> /var/lib/clickhouse/preprocessed_configs
-rw-r--r-- 1 root root 4744 Jan 23 22:21 users.xml
``` | 2019-01-26T22:32:45 | 2019-01-26T22:32:45 | {} | MEMBER |
ClickHouse | ClickHouse | 457,872,383 | 658 | alexey-milovidov | > The -1 = <ClientID> part is used for selecting the total number of events for the given EventDate. <ClientID> is substituted by -1 in this case.
> Unfortunately, this part disables (EventDate, ClientID) key usage, so the clickhouse scans all the rows for the given EventDate even if ClientID isn't equal to -1.
T... | 2019-01-26T22:45:54 | 2019-01-26T22:46:40 | {} | MEMBER |
ClickHouse | ClickHouse | 457,874,994 | 1,445 | alexey-milovidov | Do you experience this issue with the latest ClickHouse release? | 2019-01-26T23:27:03 | 2019-01-26T23:27:03 | {} | MEMBER |
ClickHouse | ClickHouse | 457,933,623 | 3,567 | alexey-milovidov | Duplicate of https://github.com/yandex/ClickHouse/issues/4108 | 2019-01-27T16:45:02 | 2019-01-27T16:45:02 | {} | MEMBER |
ClickHouse | ClickHouse | 458,250,981 | 3,291 | nvartolomei | Test suggestion:
`SELECT -b as b FROM test.merge1 ORDER BY merge1.b DESC;`
And get results before and after this patch. | 2019-01-28T18:40:46 | 2019-01-28T18:40:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,934,160 | 538 | alexey-milovidov | Thank you for the follow up!
We are aware that it's very easy to create too many partitions. We are thinking about how we can avoid this issue from the usability perspective. (For example, the user can by mistake create partitioning schema with one partition per second of input data - then the filesystem quickly bec... | 2019-01-27T16:52:08 | 2019-01-27T16:52:08 | {} | MEMBER |
ClickHouse | ClickHouse | 458,595,282 | 4,108 | 4ertus2 | fix for inner and left joins #4184
right & full are in progress | 2019-01-29T15:58:25 | 2019-01-29T15:58:25 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,633,581 | 3,891 | ztlpn | @andywow By the same problem you mean ALTER timeout or the `No message received` message? | 2019-01-29T17:35:36 | 2019-01-29T17:35:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,684,140 | 3,891 | ztlpn | These are two unrelated issues. I'll close this one as presumably it is fixed in 19.1.6. Let us continue the investigation in the separate issue. | 2019-01-29T19:55:41 | 2019-01-29T19:55:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,879,119 | 2,269 | alexey-milovidov | You should check your overcommit settings: https://clickhouse.yandex/docs/en/operations/tips/#ram | 2019-01-27T00:48:26 | 2019-01-27T00:48:26 | {} | MEMBER |
ClickHouse | ClickHouse | 457,881,506 | 4,162 | den-crane | two blocks vs one block
```
create table fooL (a Int64, v String) engine = Memory;
insert into fooL values (0, 'L');
insert into fooL values (0, 'L');
select distinct v from fooL any left outer join (select toInt64(0) a) using(a)
┌─v─┐
│ L │
└───┘
┌─v─┐
│ L │
└───┘
```
VS
```
create table fooL (a In... | 2019-01-27T01:43:59 | 2019-01-27T01:43:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,124,020 | 4,164 | alesapin | @blinkov Do you have any other remarks? | 2019-01-28T13:07:35 | 2019-01-28T13:07:35 | {} | MEMBER |
ClickHouse | ClickHouse | 458,500,579 | 4,180 | filimonov | Пожалуйста, поищите в /var/log/clickhouse-server/clickhouse-server.log (или в log.X.gz) какие именно запросы обрабатывали эти 2 потока (32038 и 168 ) примерно в это время.
Примерно так:
```
zgrep '2019.01.22 08:' /var/log/clickhouse-server/clickhouse-server.log.* | grep '\[ 32038 \]'
zgrep '2019.01.22 08:' /var/lo... | 2019-01-29T11:08:14 | 2019-01-29T11:08:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,734,660 | 4,146 | den-crane | >It's on a dedicated set of 4 spindles in raid-0
raid0 still gives you latency of a single HDD 10 ms -> 200 IOPS.
any SSD gives 10000-100000 IOPS, the best gives 3-10mil IOPS
migrate to SSD , reduce number of zoo transactions -> bigger batches, reduce number partitions touched by insert.
| 2019-01-25T21:25:53 | 2019-01-25T21:26:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,933,265 | 4,160 | alexey-milovidov | Duplicate of https://github.com/yandex/ClickHouse/issues/4108 | 2019-01-27T16:40:25 | 2019-01-27T16:40:25 | {} | MEMBER |
ClickHouse | ClickHouse | 458,427,203 | 4,177 | alexey-milovidov | Duplicate #4156 | 2019-01-29T06:50:32 | 2019-01-29T06:50:32 | {} | MEMBER |
ClickHouse | ClickHouse | 458,509,331 | 4,168 | zhang2014 | @blinkov Done.
`sumburConsistentHash`, `toLowCardinality`, `lowCardinalityIndices`, `lowCardinalityKeys`, `trunc`, `truncate`, `yandexConsistentHash` functions may require you to add. | 2019-01-29T11:39:51 | 2019-01-29T11:39:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,874,705 | 1,486 | alexey-milovidov | Outdated.
The allocator was changed.
`bad_alloc` exceptions should not cause shutdown. | 2019-01-26T23:21:21 | 2019-01-26T23:21:21 | {} | MEMBER |
ClickHouse | ClickHouse | 457,876,373 | 2,210 | alexey-milovidov | Support for `MODIFY PRIMARY KEY` undocumented feature was removed. | 2019-01-26T23:52:35 | 2019-01-26T23:52:35 | {
"-1": 3,
"laugh": 2
} | MEMBER |
ClickHouse | ClickHouse | 457,879,131 | 538 | pondix | @alexey-milovidov - apologies, how very rude of me to not follow up on my previous comment after resolving the issue 😄 For the sake of completeness, I'll write it down here so perhaps it can help someone else in future.
The problem I was facing was actually due to the files having RANDOM dates... this was because ... | 2019-01-27T00:48:56 | 2019-01-27T00:48:56 | {} | NONE |
ClickHouse | ClickHouse | 458,126,129 | 2,355 | alexey-milovidov | Both questions have been resolved by @ercolanelli-leo (in master, scheduled for the next release). | 2019-01-28T13:13:13 | 2019-01-28T13:13:13 | {} | MEMBER |
ClickHouse | ClickHouse | 458,126,924 | 1,975 | alexey-milovidov | Resolved by @ercolanelli-leo (in master, scheduled for the next release). | 2019-01-28T13:15:29 | 2019-01-28T13:15:29 | {} | MEMBER |
ClickHouse | ClickHouse | 458,134,034 | 3,983 | alexey-milovidov | So, we should just set `OPENSSLDIR` to `/etc/ssl/certs` at the build time? | 2019-01-28T13:34:55 | 2019-01-28T13:34:55 | {} | MEMBER |
ClickHouse | ClickHouse | 458,609,201 | 2,241 | megastef | Thank you for the workaround above. I developed a CH log parser with multi line parsing of CH logs, which works also for containers.
I wonder why these logger settings are not enabled by default in the docker specific configuration: https://github.com/yandex/ClickHouse/blob/master/docker/server/docker_related_confi... | 2019-01-29T16:32:40 | 2019-01-29T16:32:40 | {
"+1": 9
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,870,993 | 4,192 | alesapin | @zhang2014 Why ` SET max_threads = 1; ` helps? Maybe it hides real problem? | 2019-01-30T09:20:50 | 2019-01-30T09:20:50 | {} | MEMBER |
ClickHouse | ClickHouse | 457,873,495 | 838 | alexey-milovidov | Fixed. | 2019-01-26T22:59:11 | 2019-01-26T22:59:11 | {
"heart": 1,
"hooray": 5
} | MEMBER |
ClickHouse | ClickHouse | 457,874,840 | 1,389 | alexey-milovidov | Implemented. | 2019-01-26T23:24:06 | 2019-01-26T23:24:06 | {} | MEMBER |
ClickHouse | ClickHouse | 457,879,222 | 2,240 | alexey-milovidov | Fixed. | 2019-01-27T00:51:42 | 2019-01-27T00:51:42 | {
"hooray": 1
} | MEMBER |
ClickHouse | ClickHouse | 458,501,235 | 4,180 | MalKeshar | К сожалению, другой информации нет. | 2019-01-29T11:10:38 | 2019-01-29T11:10:38 | {} | NONE |
ClickHouse | ClickHouse | 458,501,524 | 4,180 | MalKeshar | На будущее, лучше включить debug уровень в журнале? | 2019-01-29T11:11:39 | 2019-01-29T11:11:39 | {} | NONE |
ClickHouse | ClickHouse | 457,866,513 | 124 | alexey-milovidov | Fixed long time ago.
Now we have single set of packages for all Ubuntu/Debian flavours.
| 2019-01-26T21:13:01 | 2019-01-26T21:13:01 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 457,875,023 | 1,460 | alexey-milovidov | This task is assigned to @igron99 | 2019-01-26T23:27:49 | 2019-01-26T23:27:49 | {} | MEMBER |
ClickHouse | ClickHouse | 457,875,875 | 1,815 | alexey-milovidov | Implemented: #3377 | 2019-01-26T23:42:56 | 2019-01-26T23:42:56 | {} | MEMBER |
ClickHouse | ClickHouse | 457,879,640 | 2,369 | alexey-milovidov | Protobuf output format (export from ClickHouse) has been implemented (in master).
Nested messages are not yet supported.
We are going to implement protobuf input format, and only then consider implementing nested messages. | 2019-01-27T01:00:43 | 2019-01-27T01:00:43 | {} | MEMBER |
ClickHouse | ClickHouse | 457,934,301 | 838 | nvartolomei | Closed in https://github.com/yandex/ClickHouse/commit/8cd1e8c21dcd65c4bfe8815b582bf7874842014a. | 2019-01-27T16:53:48 | 2019-01-27T16:53:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,951,274 | 157 | Abzac | Sounds amazing! | 2019-01-27T20:28:11 | 2019-01-27T20:28:11 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 458,612,731 | 4,177 | proller | https://github.com/yandex/ClickHouse/pull/4188 | 2019-01-29T16:41:19 | 2019-01-29T16:41:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,155,881 | 4,165 | alexey-milovidov | Details:
`update a = x where cond` is translated to `update a = cond ? x : a`
Function if: `date ? '2019-01-01' : date` does not support conversion from constant string to date. | 2019-01-28T14:37:10 | 2019-01-28T14:37:10 | {} | MEMBER |
ClickHouse | ClickHouse | 458,213,178 | 3,656 | alexey-milovidov | It's about aliases visibility in ClickHouse.
(Well known and documented behaviour that is not SQL compatible.)
There is a chance that we will reconsider this behaviour. | 2019-01-28T16:58:34 | 2019-01-28T16:58:34 | {} | MEMBER |
ClickHouse | ClickHouse | 457,726,294 | 4,146 | iameugenejo | I see, I'll try separating them out when we migrate to another environment. Thank you for the suggestion. | 2019-01-25T21:00:09 | 2019-01-25T21:00:09 | {} | NONE |
ClickHouse | ClickHouse | 457,336,661 | 4,137 | healiseu | That's what I have read in the [`Block Streams`][1] section of the documentation, isn't this what you are referring to more or less ? Anyway, the `TRIACLICK` release is going to be for experimental and demonstration purposes, no need to seek for the optimum solution of this problem now. I am aiming to show that `Associ... | 2019-01-24T20:02:28 | 2019-01-25T23:12:56 | {} | NONE |
ClickHouse | ClickHouse | 457,867,393 | 222 | alexey-milovidov | We have implemented `arrayDifference` function FTW. | 2019-01-26T21:27:06 | 2019-01-26T21:27:06 | {} | MEMBER |
ClickHouse | ClickHouse | 457,871,218 | 538 | alexey-milovidov | I cannot reproduce timeouts while loading huge CSV through clickhouse-client. Tested with 600 GB CSV file.
For `lineorder` table from "star schema benchmark", ClickHouse can parse about 1.5 million rows per second (170 MB/sec) on single decent CPU core - that's definitely not slow. | 2019-01-26T22:26:50 | 2019-01-26T22:26:50 | {} | MEMBER |
ClickHouse | ClickHouse | 457,873,380 | 812 | alexey-milovidov | @cir0p
According to the [operation instructions](https://clickhouse.yandex/docs/en/operations/tips/#ram):
Do not disable overcommit. The value `cat /proc/sys/vm/overcommit_memory` should be 0 or 1. Run
```
echo 0 | sudo tee /proc/sys/vm/overcommit_memory
```
| 2019-01-26T22:56:58 | 2019-01-26T22:56:58 | {} | MEMBER |
ClickHouse | ClickHouse | 457,875,687 | 1,579 | alexey-milovidov | We're aware of MaxMind database and you can use it with ClickHouse through `ip_trie` dictionary. | 2019-01-26T23:39:13 | 2019-01-26T23:39:13 | {} | MEMBER |
ClickHouse | ClickHouse | 458,205,560 | 1,766 | alexey-milovidov | `Bad address` should not be related to high IO usage.
It can be related to memory corruption.
Do you experience this issue on the latest stable release? | 2019-01-28T16:40:10 | 2019-01-28T16:40:44 | {} | MEMBER |
ClickHouse | ClickHouse | 458,469,916 | 4,175 | alesapin | @nvartolomei This issue is not known, so I need your help in investigation. Do you use custom compression codec (zstd or lzhc)? | 2019-01-29T09:31:32 | 2019-01-29T09:31:32 | {} | MEMBER |
ClickHouse | ClickHouse | 458,484,562 | 4,175 | alesapin | We have test for this https://github.com/yandex/ClickHouse/blob/master/dbms/tests/integration/test_non_default_compression/test.py#L10-L25, but seems like it does not cover all cases :(
As I can see, this segfault occurred during `SELECT` query. But can the server start? Do merges work? Do you have any LowCardinali... | 2019-01-29T10:15:48 | 2019-01-29T10:15:48 | {} | MEMBER |
ClickHouse | ClickHouse | 458,509,585 | 4,168 | zhang2014 | @alexey-milovidov Have we considered deleting them? | 2019-01-29T11:40:47 | 2019-01-29T11:40:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,570,888 | 4,175 | alesapin | Bug reproduces with `use_uncompressed_cache=1`.
Fix https://github.com/yandex/ClickHouse/pull/4186. | 2019-01-29T14:58:32 | 2019-01-29T14:58:32 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 458,612,258 | 4,156 | proller | https://github.com/yandex/ClickHouse/pull/4188 | 2019-01-29T16:40:02 | 2019-01-29T16:40:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,765,719 | 4,146 | iameugenejo | Thank you, will try out. | 2019-01-25T23:07:20 | 2019-01-25T23:07:20 | {} | NONE |
ClickHouse | ClickHouse | 457,875,034 | 1,460 | alexey-milovidov | Duplicate: #3605 | 2019-01-26T23:28:04 | 2019-01-26T23:28:04 | {} | MEMBER |
ClickHouse | ClickHouse | 458,147,211 | 2,232 | Duiesel | > Now you can build only client :
>
> ```
> cmake .. -DENABLE_CLICKHOUSE_ALL=0 -DENABLE_CLICKHOUSE_CLIENT=1 -DCMAKE_CXX_COMPILER=`which g++-7` -DCMAKE_C_COMPILER=`which gcc-7`
> make -j clickhouse-client
> ```
Doesn't work for now. | 2019-01-28T14:15:02 | 2019-01-28T14:15:02 | {} | NONE |
ClickHouse | ClickHouse | 458,200,068 | 2,718 | alexey-milovidov | These tasks are assigned to
@comunodi
@Gleb-Tretyakov
@favstovol
@FawnD2 | 2019-01-28T16:26:21 | 2019-01-28T16:26:21 | {} | MEMBER |
ClickHouse | ClickHouse | 458,215,117 | 3,656 | code-of-kpp | Can you please elaborate a little here? It seems "logical" to have `a` as String or `a` as integer in the output, but error shown is at least unhelpful. | 2019-01-28T17:03:52 | 2019-01-28T17:03:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,626,445 | 4,166 | alexey-milovidov | This PR fixes #2565 | 2019-01-29T17:16:30 | 2019-01-29T17:16:30 | {} | MEMBER |
ClickHouse | ClickHouse | 458,687,016 | 4,047 | ztlpn | @andywow reports that the `No message received` error is still present on 19.1.6. @alesapin can you take a look? | 2019-01-29T20:04:11 | 2019-01-29T20:04:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,713,343 | 863 | alexey-milovidov | The feature is implemented but not enabled by default. It can to be enabled manually by setting the `enable_optimize_predicate_expression` to 1. The implementation is considered to be in experimental state and should be enabled after thorough testing on your workload.
I would like to have the feedback about presence... | 2019-01-29T21:26:19 | 2019-01-29T21:30:34 | {} | MEMBER |
ClickHouse | ClickHouse | 457,933,346 | 3,698 | alexey-milovidov | Duplicate of https://github.com/yandex/ClickHouse/issues/4108 | 2019-01-27T16:41:22 | 2019-01-27T16:41:22 | {} | MEMBER |
ClickHouse | ClickHouse | 458,159,590 | 2,870 | alexey-milovidov | @s-mx is going to fix this issue. | 2019-01-28T14:47:10 | 2019-01-28T14:47:10 | {} | MEMBER |
ClickHouse | ClickHouse | 458,173,517 | 2,232 | proller | How exactly it doesn't work? | 2019-01-28T15:22:51 | 2019-01-28T15:22:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,248,044 | 3,611 | alexey-milovidov | CC @amosbird
I cannot reproduce the issue in a simple use case:
```
CREATE TABLE test.replicated_numbers (k UInt64, x UInt64) ENGINE = ReplicatedMergeTree('/clickhouse/tables/test/replicated_numbers', 'r1') PARTITION BY k ORDER BY x;
CREATE TABLE test.replicated_numbers_2 (k UInt64, x UInt64) ENGINE = Replicated... | 2019-01-28T18:32:33 | 2019-01-28T18:32:33 | {} | MEMBER |
ClickHouse | ClickHouse | 457,833,154 | 1,443 | qietingfengling | > @qietingfengling do you have any further questions?
This happens when I use the MergeTree engine. The solution is to find a copy of the data inconsistency and replace it. Later, I used the ReplicatedMergeTree engine and I didn’t encounter it. | 2019-01-26T13:56:06 | 2019-01-26T13:56:06 | {} | NONE |
ClickHouse | ClickHouse | 457,864,662 | 24 | alexey-milovidov | In the meantime `ssb-dbgen` tool has been fixed and now it compiles with gcc 7.3 without issues.
Please note that `ssb-dbgen` is very old fashioned third-party tool and there are numerous warnings during build, but it works fine.
Benchmark setup description can be found here:
https://www.altinity.com/blog/2017/6/1... | 2019-01-26T20:46:22 | 2019-01-26T20:50:43 | {} | MEMBER |
ClickHouse | ClickHouse | 457,870,918 | 4,160 | jvitkauskas | I know this is not the preferred way to select data from ClickHouse, but it was the first thing I have tried and even if I am no longer using that table arrangement, I think this issue (exception) should be still valid. | 2019-01-26T22:22:09 | 2019-01-26T22:22:09 | {} | NONE |
ClickHouse | ClickHouse | 457,873,878 | 904 | alexey-milovidov | When a node of ZooKeeper cluster goes down, ClickHouse have to reinitialize a session with another ZooKeeper node. INSERTs will stop to work until the session is reinitialized. In most cases it should not take longer than ZooKeeper session timeout (30 seconds by default) but usually it will do quicker.
In recent Cli... | 2019-01-26T23:05:45 | 2019-01-26T23:05:45 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 457,876,985 | 2,302 | meshpaul | Ok
On Sat, Jan 26, 2019, 19:02 alexey-milovidov <notifications@github.com
wrote:
> Closed #2302 <https://github.com/yandex/ClickHouse/issues/2302>.
>
> —
> You are receiving this because you are subscribed to this thread.
> Reply to this email directly, view it on GitHub
> <https://github.com/yandex/ClickHouse/issues/... | 2019-01-27T00:04:06 | 2019-01-27T00:04:06 | {} | NONE |
ClickHouse | ClickHouse | 458,047,628 | 4,157 | KochetovNicolai | I think now modifying weights is the best way.
You can also reshard all data using [clickhouse-copier](https://clickhouse.yandex/docs/en/single/#utils/clickhouse-copier/), but it's not so cheap. | 2019-01-28T09:06:26 | 2019-01-28T09:06:26 | {} | MEMBER |
ClickHouse | ClickHouse | 458,092,800 | 751 | PoulpiFr | > ClickHouse replaces all rows with the same primary key (or more accurately, with the same sorting key) with a single row (within a one data part) that stores a combination of states of aggregate functions.
Cf: https://clickhouse.yandex/docs/en/operations/table_engines/aggregatingmergetree/
For me, the documenta... | 2019-01-28T11:23:56 | 2019-01-28T11:23:56 | {} | NONE |
ClickHouse | ClickHouse | 458,131,871 | 4,166 | alexey-milovidov | Note: it's difficult to provide a test. | 2019-01-28T13:28:47 | 2019-01-28T13:28:47 | {} | MEMBER |
ClickHouse | ClickHouse | 458,292,963 | 4,168 | alexey-milovidov | ```
OSHierarchy
OSIn
OSToRoot
SEHierarchy
SEIn
SEToRoot
```
These functions are deprecated and should not be listed in docs. | 2019-01-28T20:41:06 | 2019-01-28T20:41:06 | {} | MEMBER |
ClickHouse | ClickHouse | 458,481,580 | 4,175 | nvartolomei | @alesapin zstd compression is used. | 2019-01-29T10:06:12 | 2019-01-29T10:06:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,674,794 | 4,146 | iameugenejo | 1. We don't have clickhouse running on the boxes zookeeper is running. Though we have ingestor (inserter) running on the zookeeper boxes. The ingestors do not write on disks, perform less memory and high cpu usage tasks, whereas zookeeper uses decent memory and less cpu with high disk usages. So we decided to have them... | 2019-01-25T18:35:05 | 2019-01-25T20:44:36 | {} | NONE |
ClickHouse | ClickHouse | 457,867,097 | 157 | alexey-milovidov | Our management has approved usage of Yandex internal fast library for user-agent parsing in ClickHouse. But the task is still out of scope of our plans. | 2019-01-26T21:22:38 | 2019-01-26T21:22:38 | {} | MEMBER |
ClickHouse | ClickHouse | 457,877,150 | 2,150 | alexey-milovidov | Действительно, там дурацкий код. | 2019-01-27T00:06:52 | 2019-01-27T00:06:52 | {} | MEMBER |
ClickHouse | ClickHouse | 458,201,882 | 2,565 | alexey-milovidov | Duplicate of #2870 | 2019-01-28T16:30:57 | 2019-01-28T16:30:57 | {} | MEMBER |
ClickHouse | ClickHouse | 458,204,107 | 2,304 | alexey-milovidov | Embedded dictionaries implementation is totally determined by initial Yandex.Metrica use case. We are not going to extend embedded dictionaries. Better to improve usability of external dictionaries instead. Even Yandex.Metrica is using external dictionaries for all new use cases. Some embedded dictionaries (`OS` and `S... | 2019-01-28T16:36:33 | 2019-01-28T16:36:33 | {} | MEMBER |
ClickHouse | ClickHouse | 458,524,186 | 3,891 | andywow | I got same problem on 19.1.6. Downgrade to 18.14.11 helped. | 2019-01-29T12:36:09 | 2019-01-29T12:36:09 | {} | NONE |
ClickHouse | ClickHouse | 458,657,263 | 3,891 | andywow | @ztlpn, `No message received` message | 2019-01-29T18:41:36 | 2019-01-29T18:41:36 | {} | NONE |
ClickHouse | ClickHouse | 458,776,798 | 1,510 | den-crane | > i encounter this error on large request (900M lines on table2) like that :
>
> ```
> INSERT INTO <table1> SELECT * FROM <table2>
> ```
How is your table partitioned ? By month? | 2019-01-30T01:35:15 | 2019-01-30T01:35:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,724,212 | 4,146 | filimonov | > Though we have ingestor (inserter) running on the zookeeper boxes. The ingestors do not write on disks, perform less memory and high cpu usage tasks, whereas zookeeper uses decent memory and less cpu with high disk usages. So we decided to have them in the same box to utilize the resources efficiently.
That kind o... | 2019-01-25T20:53:41 | 2019-01-25T20:57:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 457,876,511 | 2,180 | alexey-milovidov | PS. `-D USE_STATIC_LIBRARIES:BOOL=False`
This option is discouraged. Static linking is always preferred for ClickHouse. | 2019-01-26T23:55:47 | 2019-01-26T23:55:47 | {} | MEMBER |
ClickHouse | ClickHouse | 457,922,070 | 4,159 | healiseu | #### Answer:
These are my two-penny thoughts on the problem. For a relatively small result set one can group by all columns. A more efficient approach requires to **check grouping columns for cardinality and NULL values**. The **number of columns to include** in the grouping is also another parameter.
Deduplicatio... | 2019-01-27T14:24:06 | 2019-01-27T20:16:00 | {} | NONE |
ClickHouse | ClickHouse | 458,196,559 | 2,232 | alexey-milovidov | You have to replace `make` with `ninja`.
| 2019-01-28T16:18:11 | 2019-01-28T16:18:11 | {} | MEMBER |
ClickHouse | ClickHouse | 457,872,534 | 658 | alexey-milovidov | But the task remains. | 2019-01-26T22:47:49 | 2019-01-26T22:47:49 | {} | MEMBER |
ClickHouse | ClickHouse | 458,091,154 | 4,156 | proller | Can you show ```show create table db1.pixel_data_local```
And some sample data from insert ?
Better something like test.sh with server crash after run | 2019-01-28T11:18:32 | 2019-01-28T11:18:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,125,683 | 4,164 | blinkov | No | 2019-01-28T13:11:57 | 2019-01-28T13:11:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 458,494,023 | 4,166 | alexey-milovidov | Failed fuzz test with UBSan: #4179 | 2019-01-29T10:45:42 | 2019-01-29T10:45:42 | {} | MEMBER |
ClickHouse | ClickHouse | 458,501,791 | 4,180 | MalKeshar | С другой стороны, включены query_log, query_thread_log. Посмотреть там?
UP: нет, к сожалению, там данных за январь нет. | 2019-01-29T11:12:44 | 2019-01-29T11:30:40 | {} | NONE |
ClickHouse | ClickHouse | 458,616,346 | 863 | panandr | Is there any update on this ? I see this is merged : https://github.com/yandex/ClickHouse/pull/2881/commits
This means that this is fixed now and optimized selects work for normal views ? | 2019-01-29T16:50:10 | 2019-01-29T16:50:10 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.