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 | 680,797,798 | 11,684 | alexey-milovidov | Integration test fail sometimes, need to investigate. | 2020-08-26T10:30:29 | 2020-08-26T10:30:29 | {} | MEMBER |
ClickHouse | ClickHouse | 680,875,640 | 12,155 | den-crane | > executeQuery: Read 117847707 rows, 16.89 GiB in 3.824 sec., 30815014 rows/sec., 4.42 GiB/sec.
> executeQuery: Read 117847707 rows, 16.89 GiB in 13.651 sec., 8633189 rows/sec., 1.24 GiB/sec.
It is expected behavior in some cases. When CH needs to transfer huge volume of intermediate results from shards to an initi... | 2020-08-26T13:21:27 | 2020-08-26T13:21:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,987,636 | 13,941 | alexey-milovidov | > Integration tests (release) — fail: 1, passed: 778, error: 0
Not enough timeout in operations because our CI environment is too chaotic. #14117 | 2020-08-26T16:29:26 | 2020-08-26T16:29:26 | {} | MEMBER |
ClickHouse | ClickHouse | 681,023,082 | 14,060 | akuzm | `01396_inactive_replica_cleanup_nodes` is already reported https://github.com/ClickHouse/ClickHouse/issues/12508 | 2020-08-26T17:36:20 | 2020-08-26T17:36:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,058,067 | 14,124 | 4ertus2 | We probably need some uneversal solution to check if Decimal has number of digits less or equal to its precision. Something like `countDigits(x)`. Then we could both check decimal overflow and use honest precision for extracted values. I.e. when we need it for extraction in external DBMS. | 2020-08-26T18:47:28 | 2020-08-26T18:48:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,100,919 | 14,045 | alexey-milovidov | > version 20.6.4.44
> DB::FunctionIf::executeGeneric
The same stack trace:
> version 20.5.4.40
> DB::FunctionIf::executeGeneric
It means that the issue does not relate to:
> after upgrade from v20.5.5.74 to v20.6.4.44
| 2020-08-26T20:16:51 | 2020-08-26T20:17:05 | {} | MEMBER |
ClickHouse | ClickHouse | 680,795,660 | 13,691 | alexey-milovidov | 01396_inactive_replica_cleanup_nodes - fixed
01393_benchmark_secure_port - fixed | 2020-08-26T10:25:21 | 2020-08-26T10:25:21 | {} | MEMBER |
ClickHouse | ClickHouse | 680,902,030 | 13,027 | alexey-milovidov | Named tuples are not obsolete and should be supported and covered by tests. But they are rarely used. | 2020-08-26T14:07:06 | 2020-08-26T14:07:06 | {} | MEMBER |
ClickHouse | ClickHouse | 680,962,754 | 14,014 | alesapin | > @alesapin I wanted to test the checks that I had implemented and for that I'm using the following query:
>
> ```
> CREATE TABLE indexed_table ( `tm` DateTime, `log_message` String, INDEX log_message log_message TYPE tokenbf_v1(4096, 2, 0) GRANULARITY 1 ) ENGINE = MergeTree ORDER BY (tm) SETTINGS min_i... | 2020-08-26T15:46:34 | 2020-08-26T15:46:34 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 680,999,961 | 14,122 | alexey-milovidov | ExpressionJIT passes empty number of arguments (because of constants?).
This is still to investigate in #8482 | 2020-08-26T16:52:24 | 2020-08-26T16:52:24 | {} | MEMBER |
ClickHouse | ClickHouse | 681,023,460 | 14,060 | akuzm | > > Backwards-incompatible change
>
> @akuzm worth changing the labels?
They are mostly for convenience of filtering and don't influence the changelog, but yes, I'll change. | 2020-08-26T17:37:02 | 2020-08-26T17:37:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,031,168 | 12,413 | alexey-milovidov | This modification is practially untestable. The only way is to test manually... I have tested and it does not work. | 2020-08-26T17:53:00 | 2020-08-26T17:53:00 | {} | MEMBER |
ClickHouse | ClickHouse | 680,718,668 | 13,991 | liangli1213 | @BohuTANG got it, thanks very much, and by the way, when do you think the GTID replication feature be merged into master? | 2020-08-26T07:48:11 | 2020-08-26T07:48:11 | {} | NONE |
ClickHouse | ClickHouse | 680,852,351 | 14,096 | den-crane | it was intentional change https://github.com/ClickHouse/ClickHouse/pull/10049
and \n is showed because default format is TSV
I don't see UB here.
https://github.com/ClickHouse/ClickHouse/blob/master/CHANGELOG.md#clickhouse-release-v20429-2020-05-12
ClickHouse release v20.4.2.9, 2020-05-12
Backward Incompati... | 2020-08-26T12:36:18 | 2020-08-26T12:36:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,974,224 | 14,111 | alexey-milovidov | Returned bug again :) | 2020-08-26T16:06:07 | 2020-08-26T16:06:07 | {} | MEMBER |
ClickHouse | ClickHouse | 681,062,014 | 14,129 | alexey-milovidov | I've read the fix but I don't understand how it's related to LowCardinality. | 2020-08-26T18:55:45 | 2020-08-26T18:55:45 | {} | MEMBER |
ClickHouse | ClickHouse | 681,098,846 | 14,045 | alexey-milovidov | Please provide full logs for `from thread 5606`. You can `grep -F '[ 5606 ]'`
Also look at `dmesg`.
| 2020-08-26T20:12:26 | 2020-08-26T20:12:26 | {} | MEMBER |
ClickHouse | ClickHouse | 680,896,053 | 14,109 | johnskopis | > "johnskopis wants to merge 1 commit into ClickHouse:19.18"
>
> Can you do the same to master branch please? 19.18 was never released and is not supported.
Sure. FWIW I developed against 19.13 and picked what I thought was highest version in v19.x series. | 2020-08-26T13:56:59 | 2020-08-26T13:56:59 | {} | NONE |
ClickHouse | ClickHouse | 681,038,571 | 13,597 | epochslee | > @Renampme
> **you can run the command below**
>
> ```
> $ brew install binutils
> $ mdfind -name objcopy
> /Users/mac/Documents/GitHub/ClickHouse/contrib/llvm/llvm/utils/gn/secondary/llvm/tools/llvm-objcopy
> /usr/local/Cellar/binutils/2.35/share/man/man1/objcopy.1
> /usr/local/Cellar/binutils/2.35/x86_64-ap... | 2020-08-26T18:07:40 | 2020-08-26T18:07:40 | {} | NONE |
ClickHouse | ClickHouse | 681,072,543 | 14,130 | alexey-milovidov | I see dumb code in this line: https://clickhouse.tech/codebrowser/html_report/ClickHouse/src/Storages/Distributed/DirectoryMonitor.cpp.html#217 | 2020-08-26T19:15:47 | 2020-08-26T19:15:47 | {} | MEMBER |
ClickHouse | ClickHouse | 680,910,026 | 14,045 | radiatorius | `/etc/cron.d/clickhouse-server ` was uncommented, after commenting out, the error (_Another server instance in same directory is already running._) does not seem to recur, thanks.
But the segmentation fault issue is still recurring:
```
2020.08.26 14:16:50.381725 [ 5593 ] {} <Warning> onea_lake.orders_lv: Removi... | 2020-08-26T14:20:01 | 2020-08-26T14:20:01 | {} | NONE |
ClickHouse | ClickHouse | 680,938,779 | 14,113 | johnskopis | fixes #14106 | 2020-08-26T15:06:33 | 2020-08-26T15:06:33 | {} | NONE |
ClickHouse | ClickHouse | 681,000,234 | 14,122 | alexey-milovidov | Right now `compile_expressions` is rather dangerous experimental feature that is disabled by default. | 2020-08-26T16:52:58 | 2020-08-26T16:52:58 | {} | MEMBER |
ClickHouse | ClickHouse | 681,064,560 | 14,124 | 4ertus2 | We have a discussion and decide to make two functions: `countDigits(x)` and `isDecimalOverflow(x, P)`. First one is helpful for other types too. Second one should have one or two arguments. If second arg is ommited it should extract P from the first arg's data type. | 2020-08-26T19:00:46 | 2020-08-26T19:06:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,107,796 | 14,123 | alexey-milovidov | To compensate, we can raise the number of threads to 64. It will definitely saturate CPU. | 2020-08-26T20:31:01 | 2020-08-26T20:31:01 | {} | MEMBER |
ClickHouse | ClickHouse | 680,780,000 | 14,045 | radiatorius | when the server is operational we still get errors:
```(version 20.5.4.40 (official build))
2020.08.26 09:43:59.285056 [ 1596 ] {} <Error> Application: DB::Exception: Cannot lock file /var/lib/clickhouse/status. Another server instance in same directory is already running.
2020.08.26 09:44:29.534904 [ 1777 ] {} <Er... | 2020-08-26T09:51:39 | 2020-08-26T10:22:04 | {} | NONE |
ClickHouse | ClickHouse | 680,795,081 | 14,096 | filimonov | clickhouse-client -q "show create table default.Test" --format=TSVRaw | 2020-08-26T10:24:00 | 2020-08-26T10:24:00 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,796,884 | 14,043 | alexey-milovidov | Is there a chance that the bug you have encountered is the same as described above?
(It was fixed in master.) | 2020-08-26T10:28:17 | 2020-08-26T10:28:27 | {} | MEMBER |
ClickHouse | ClickHouse | 680,823,345 | 12,155 | Rafael-Luo | @filimonov | 2020-08-26T11:31:35 | 2020-08-26T11:31:35 | {} | NONE |
ClickHouse | ClickHouse | 680,837,670 | 12,508 | alexey-milovidov | > seems like table is not dropped from zookeeper
AFAIK there is DROP NO DELAY, @tavplubix ? | 2020-08-26T12:05:16 | 2020-08-26T12:05:16 | {} | MEMBER |
ClickHouse | ClickHouse | 680,902,075 | 14,109 | johnskopis | v20 pr: https://github.com/ClickHouse/ClickHouse/pull/14113 | 2020-08-26T14:07:12 | 2020-08-26T14:07:12 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 681,000,734 | 14,120 | alexey-milovidov | TODO: Add performance test at least for Null.
Something like `INSERT INTO null() SELECT * FROM numbers_mt(1000000000)`.
#14121 | 2020-08-26T16:53:52 | 2020-08-26T16:53:52 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 681,004,210 | 14,121 | alexey-milovidov | Actually `file` table function with `Null` format will do the trick.
But without parallel inserts. | 2020-08-26T16:59:34 | 2020-08-26T16:59:51 | {} | MEMBER |
ClickHouse | ClickHouse | 681,025,811 | 13,845 | den-crane | >two last query return same date in result,in diffrence records.
It's normal.
Because they are two different hours in UTC rendered into the same String in Asia/Tehran because of timeshift.
```
SELECT
toDateTime('2020-03-21 23:00:01', 'UTC') AS xUTC,
toDateTime('2020-03-21 00:00:01') AS yUTC,
to... | 2020-08-26T17:42:01 | 2020-08-26T17:46:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,033,041 | 12,413 | alexey-milovidov | Test 1:
Install debian package and check that it's asking for password.
It should not break with this PR.
Test 2:
Install ClickHouse without debian package (by any other means), set up default password either with plaintext or with sha256 (both options should be tested). Then install ClickHouse with debian pa... | 2020-08-26T17:56:46 | 2020-08-26T17:56:46 | {} | MEMBER |
ClickHouse | ClickHouse | 681,060,162 | 13,861 | siradjev | From ODBC trace:
```
[ODBC][24028][1598463846.531187][../contrib/unixodbc/DriverManager/SQLGetInfo.c][236]
Entry:
Connection = 0x7fd1e825e000
Info Type = SQL_SCHEMA_USAGE (91)
Info Value = (nil)
Buffer... | 2020-08-26T18:51:48 | 2020-08-26T18:51:48 | {} | NONE |
ClickHouse | ClickHouse | 681,107,408 | 14,125 | alexey-milovidov | Yes, it works. | 2020-08-26T20:30:14 | 2020-08-26T20:30:14 | {} | MEMBER |
ClickHouse | ClickHouse | 680,706,875 | 12,508 | alesapin | > @alesapin The file `skip_list.json` is in JSON format - we cannot have comments there. It's very important for development.
> I can add some very long tests to skip under TSan but refuse to do so if I cannot write comments.
We can use `YAML`, but we will get one more dependency for `clickhouse-test` `PyYAML` (htt... | 2020-08-26T07:21:13 | 2020-08-26T07:21:13 | {} | MEMBER |
ClickHouse | ClickHouse | 680,817,017 | 14,093 | UnamedRus | A little correction: looks like a certain combination of parts being merged raises that issue, not a certain part itself. | 2020-08-26T11:16:38 | 2020-08-26T11:16:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,892,244 | 13,991 | tavplubix | It's merged into master, you can download the package here: https://clickhouse-builds.s3.yandex.net/0/cff92c8ad3dea2d96b1dfaadc76c3c10b3d701b2/clickhouse_build_check/report.html | 2020-08-26T13:50:29 | 2020-08-26T13:51:16 | {} | MEMBER |
ClickHouse | ClickHouse | 681,120,919 | 13,687 | bharatnc | > We are discussing about CCache - and it will probably be disabled by default on systems with no ccache installed. Currently you can overcome this with `-D"ENABLE_CCACHE"=OFF` (or by installing ccache - which will save you some time probably).
>
> With `ENABLE_THINLTO` error is completely legit. One way or another ... | 2020-08-26T20:59:53 | 2020-08-26T20:59:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,705,913 | 12,508 | alesapin | `00633_materialized_view_and_too_many_parts_zookeeper` with database atomic https://clickhouse-test-reports.s3.yandex.net/14049/e5bc5ea419d49c807dada50b1e33966537fd0ada/functional_stateless_tests_(release,_databaseatomic).html#fail1:
seems like table is not dropped from zookeeper:
```
2020-08-26 00:19:57 [0c996d3... | 2020-08-26T07:18:52 | 2020-08-26T07:19:06 | {} | MEMBER |
ClickHouse | ClickHouse | 680,276,170 | 14,045 | radiatorius | > /var/lib/clickhouse/data/db/table/delete_tmp_all_130_2078_5_2215/
after deleting all "bad" tables completely (/var/lib/clickhouse/data/db/tableX/delete_tmp_all_1_13383_3898_13462/checksums.txt) the service no longer stops. Can you explain in detail the reason for this issue and how to avoid it in the future? Thank... | 2020-08-25T21:21:07 | 2020-08-26T10:32:21 | {} | NONE |
ClickHouse | ClickHouse | 681,002,730 | 14,121 | filimonov | Vote up. Should accept any inserts with any columns and drop them away. | 2020-08-26T16:56:50 | 2020-08-26T16:56:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,130,674 | 14,045 | filimonov | > > Please provide full logs for `from thread 5606`. You can `grep -F '[ 5606 ]'`
> > Also look at `dmesg`.
>
> ```
> 2020.07.15 14:53:54.258387 [ 5606 ] {} <Error> Application: DB::Exception: Cannot lock file /var/lib/clickhouse/status. Another server instance in same directory is already running.
> 2020.08.23 1... | 2020-08-26T21:21:14 | 2020-08-26T21:21:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,836,781 | 12,508 | alexey-milovidov | I prefer JSON with comments - either "JSON5" or just simple function that strips comments before passing to JSON parser.
| 2020-08-26T12:03:27 | 2020-08-26T12:03:27 | {} | MEMBER |
ClickHouse | ClickHouse | 680,840,787 | 12,508 | tavplubix | > > seems like table is not dropped from zookeeper
>
> AFAIK there is DROP NO DELAY, @tavplubix ?
Race condition is still possible, because `NO DELAY` doesn't make `DROP` blocking. It can be "fixed" by increasing sleep time or by making all ZooKeeper paths in functional tests unique.
| 2020-08-26T12:11:48 | 2020-08-26T12:11:48 | {} | MEMBER |
ClickHouse | ClickHouse | 680,876,328 | 14,109 | filimonov | "johnskopis wants to merge 1 commit into ClickHouse:19.18"
Can you do the same to master branch please? 19.18 was never released and is not supported.
| 2020-08-26T13:22:38 | 2020-08-26T13:22:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,059,977 | 14,129 | CurtizJ | Currently it's impossible to add performance test for such short queries. | 2020-08-26T18:51:26 | 2020-08-26T18:51:26 | {} | MEMBER |
ClickHouse | ClickHouse | 681,065,966 | 14,129 | jameshartig | @CurtizJ looking forward to this! After your change, is there still a difference with `experimental_use_processors`? | 2020-08-26T19:03:29 | 2020-08-26T19:03:29 | {} | NONE |
ClickHouse | ClickHouse | 681,102,053 | 14,045 | alexey-milovidov | Please provide the info about ALTER UPDATE mutations you have executed.
You can list the contents of all `mutation_N.txt` files.
To quickly fix the issue, remove all the `mutation_N.txt` files. | 2020-08-26T20:19:09 | 2020-08-26T20:19:32 | {} | MEMBER |
ClickHouse | ClickHouse | 681,104,816 | 14,045 | radiatorius | > Please provide full logs for `from thread 5606`. You can `grep -F '[ 5606 ]'`
> Also look at `dmesg`.
```
2020.07.15 14:53:54.258387 [ 5606 ] {} <Error> Application: DB::Exception: Cannot lock file /var/lib/clickhouse/status. Another server instance in same directory is already running.
2020.08.23 10:50:00.688578... | 2020-08-26T20:24:59 | 2020-08-26T20:40:38 | {} | NONE |
ClickHouse | ClickHouse | 680,722,614 | 13,991 | bohutang | > @BohuTANG got it, thanks very much, and by the way, when do you think the GTID replication feature be merged into master?
If all goes well, maybe merged today :D | 2020-08-26T07:56:41 | 2020-08-26T07:56:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,723,000 | 12,563 | little-circle-circle | I have the same problem, I read the code about mark cache, Found this
- mark_cache_size: this config item only control mark file data size, but mark cache has some meta data for lru
- lru meta data: lru meta data size depend on number of mark files, if your mark file is small, lru meta data size will be greater t... | 2020-08-26T07:57:25 | 2020-08-26T07:57:25 | {} | NONE |
ClickHouse | ClickHouse | 680,855,497 | 14,045 | radiatorius | Situation immediately after `kill -9 $(pidof -s clickhouse-server)` and **service clickhouse-server start**
**> service clickhouse-server status**
clickhouse-server service is running
Situation after about 10 min. later (no changes were made):
**> service clickhouse-server status**
clickhouse-server servic... | 2020-08-26T12:43:18 | 2020-08-26T12:47:52 | {} | NONE |
ClickHouse | ClickHouse | 680,914,645 | 13,880 | alexey-milovidov | Merged in #14030. | 2020-08-26T14:27:42 | 2020-08-26T14:27:42 | {} | MEMBER |
ClickHouse | ClickHouse | 680,707,715 | 14,014 | bharatnc | @alesapin I wanted to test the checks that I had implemented and for that I'm using the following query:
```
CREATE TABLE indexed_table ( `tm` DateTime, `log_message` String, INDEX log_message log_message TYPE tokenbf_v1(4096, 2, 0) GRANULARITY 1 ) ENGINE = MergeTree ORDER BY (tm) SETTINGS min_index_gra... | 2020-08-26T07:23:18 | 2020-08-26T20:22:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,726,737 | 13,991 | liangli1213 | >
>
> > @BohuTANG got it, thanks very much, and by the way, when do you think the GTID replication feature be merged into master?
>
> If all goes well, maybe merged today :D
Thanks very much, and when do you think the 20.8.2 testing build be available? | 2020-08-26T08:04:53 | 2020-08-26T08:04:53 | {} | NONE |
ClickHouse | ClickHouse | 680,787,051 | 14,042 | lecranek | Thanks for reply. I've solved it. | 2020-08-26T10:06:17 | 2020-08-26T10:06:17 | {} | NONE |
ClickHouse | ClickHouse | 680,075,086 | 14,045 | radiatorius | we reverted back to version 20.5.4.40,
now in the log we find such errors:
```
_2020.08.25 14:43:43.185849 [ 2165 ] {} <Error> db.table: Cannot quickly remove directory /var/lib/clickhouse/data/db/table/delete_tmp_all_130_2078_5_2215 by removing files; fallback to recursive removal. Reason: Poco::Exception. Code: ... | 2020-08-25T14:53:56 | 2020-08-26T10:28:10 | {} | NONE |
ClickHouse | ClickHouse | 680,861,635 | 13,861 | siradjev | Same issue in 20.6.4.44 | 2020-08-26T12:55:55 | 2020-08-26T12:55:55 | {} | NONE |
ClickHouse | ClickHouse | 680,938,252 | 14,109 | johnskopis | Closing in favor of PR against master | 2020-08-26T15:05:42 | 2020-08-26T15:05:42 | {} | NONE |
ClickHouse | ClickHouse | 680,954,774 | 14,116 | alesapin | Based on #14049. Merge after it. | 2020-08-26T15:32:54 | 2020-08-26T15:32:54 | {} | MEMBER |
ClickHouse | ClickHouse | 681,036,267 | 14,079 | alexey-milovidov | @akuzm did not answer what's wrong with AST fuzzer. | 2020-08-26T18:02:54 | 2020-08-26T18:02:54 | {} | MEMBER |
ClickHouse | ClickHouse | 681,133,188 | 12,552 | akuzm | > When user creates a data type `AggregateFunction(...)`, e.g. `AggregateFunction(avg, UInt64)`, we transform it adding parameter with version at front with the most recent version, e.g. `AggregateFunction(v1, avg, UInt64)`.
Can we do without this part? If we speak about the version of serialized data, it exists onl... | 2020-08-26T21:27:22 | 2020-08-26T21:27:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,149,211 | 12,552 | alexey-milovidov | > Maybe we can use the protocol version number as a serialization version number? It's kind of big now, the current version is 54226.
It changes more frequently than binary formats of aggregate functions. So, we can use it to negotiate supported version. But not to persist as a version number. | 2020-08-26T22:08:40 | 2020-08-26T22:08:40 | {} | MEMBER |
ClickHouse | ClickHouse | 680,794,390 | 14,097 | filimonov | https://github.com/ClickHouse/ClickHouse/pull/10744 | 2020-08-26T10:22:17 | 2020-08-26T10:22:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,817,763 | 14,105 | nikitamikhaylov | Also I don't know is it a bug or not. | 2020-08-26T11:18:19 | 2020-08-26T11:18:19 | {} | MEMBER |
ClickHouse | ClickHouse | 680,838,492 | 14,109 | johnskopis | fixes #14106 | 2020-08-26T12:07:07 | 2020-08-26T12:07:07 | {} | NONE |
ClickHouse | ClickHouse | 680,862,274 | 12,508 | alexey-milovidov | @tavplubix I'm Ok changing ZooKeeper paths for specific tests...
Or maybe we can do it by enabling Atomic database by default and utilizing the new UUID feature?
Or just rename them on the fly with silly regexp in `clickhouse-test`? | 2020-08-26T12:57:17 | 2020-08-26T12:57:52 | {} | MEMBER |
ClickHouse | ClickHouse | 680,869,041 | 14,045 | filimonov | systemd vs sysV issue?
It looks your clickhouse is fired with init.d script (contains `--daeamon` parameter)
https://github.com/ClickHouse/ClickHouse/blob/master/debian/clickhouse-server.init#L180
May be some uncleaned files from older versions like `/etc/cron.d/clickhouse-server` (check content, should be commen... | 2020-08-26T13:10:11 | 2020-08-26T13:18:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,966,910 | 14,111 | akuzm | ```
*2* :) select toDecimal64(-9223372036854775808, 0);
┌─toDecimal64(-9223372036854775808, 0)─┐
│ --9223372036854775808 │
└──────────────────────────────────────┘
1 rows in set. Elapsed: 0.008 sec.
*2* :) select toDecimal64(toString(toDecimal64(-9223372036854775808, 0)), 0)
Received exce... | 2020-08-26T15:53:43 | 2020-08-26T15:53:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,975,025 | 14,111 | akuzm | The fix is simple, I'll make a PR when it compiles. | 2020-08-26T16:07:34 | 2020-08-26T16:07:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,059,622 | 14,129 | CurtizJ | Look at QPS.
Before:
```
echo "SELECT col1, col2 FROM test PREWHERE col1 = '042fecec-ea68-1b26-00a4-a4f3b3987574'" | clickhouse-benchmark -c 1 -i 2000 -d 0
Loaded 1 queries.
Queries executed: 2000.
localhost:9000, queries 2000, QPS: 735.422, RPS: 94134.037, MiB/s: 1.436, result RPS: 0.000, result MiB/s: 0.00... | 2020-08-26T18:50:42 | 2020-08-26T18:50:42 | {} | MEMBER |
ClickHouse | ClickHouse | 681,067,736 | 14,129 | CurtizJ | > After your change, is there still a difference with experimental_use_processors?
I can't check it now, because in new versions old code is already removed and `experimental_use_processors` is exists just for backward compatibility and does nothing. | 2020-08-26T19:07:11 | 2020-08-26T19:07:11 | {} | MEMBER |
ClickHouse | ClickHouse | 681,119,334 | 14,079 | akuzm | > @akuzm did not answer what's wrong with AST fuzzer.
I did https://t.me/c/1078904674/61755 (very old branch, must be meged to master) | 2020-08-26T20:56:28 | 2020-08-26T20:56:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,119,656 | 13,687 | code-of-kpp | We are discussing about CCache - and it will probably be disabled by default on systems with no ccache installed. Currently you can overcome this with `-D"ENABLE_CCACHE"=OFF` (or by installing ccache - which will save you some time probably).
With `ENABLE_THINLTO` error is completely legit. One way or another you ar... | 2020-08-26T20:57:04 | 2020-08-26T20:57:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,112,140 | 13,687 | bharatnc | I see a couple of errors when I build project using cmake:
1. ccache which I was able to remediate by installing it `apt install ccache`
2. `ThinLTO`. This check fails:
```
set (CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFO "${CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFO} -flto=thin")
elseif (ENABLE_THINLTO)
... | 2020-08-26T20:40:26 | 2020-08-26T21:05:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,719,965 | 14,043 | markpapadakis | @alexey-milovidov apologies for the confusion. It happened for me once -- I tried to track it down and it was a matter of ignoring a block, which led to this partial rewrite.
If you are going to use this implementation/alternative, you should update `MemorySource` as well like so:
- is_finished is no longer necessar... | 2020-08-26T07:50:59 | 2020-08-26T07:50:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,702,067 | 8,467 | alesapin | The idea is clear to me, but I think we have to move this logic into `StorageMergeTree` level because in the current implementation it will work both for replicated and non replicated engines. I'll try to make some improvements. | 2020-08-26T07:09:22 | 2020-08-26T15:58:52 | {} | MEMBER |
ClickHouse | ClickHouse | 680,990,500 | 13,941 | alexey-milovidov | AST fuzzer has found a new issue: https://github.com/ClickHouse/ClickHouse/issues/14118 | 2020-08-26T16:34:37 | 2020-08-26T16:34:37 | {} | MEMBER |
ClickHouse | ClickHouse | 681,003,425 | 14,121 | alexey-milovidov | But I see no way to accept arbitrary data. We just don't have a concept of table without structure. | 2020-08-26T16:58:07 | 2020-08-26T16:58:24 | {} | MEMBER |
ClickHouse | ClickHouse | 681,064,325 | 13,845 | mbtolou | > > `SELECT toDateTime('2020-03-21 00:00:00'),toDateTime('2020-03-21 01:00:00'),toDateTime('2020-03-21 23:00:00')`
> > 
>
> Do you mean it's a wrong day for the time shift ?
This is normal alsoo ? | 2020-08-26T19:00:18 | 2020-08-26T19:00:18 | {} | NONE |
ClickHouse | ClickHouse | 681,144,167 | 12,552 | akuzm | > We'd want to use a tightly packed small number to minimize overhead, so probably we'll have to use some kind of lookup tables...
Thinking more about it, we don't have to put the version number into the function state itself -- we can't have mixed versions of states within block. If we put the version into column m... | 2020-08-26T21:55:39 | 2020-08-26T21:55:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,801,232 | 14,043 | markpapadakis | @alexey-milovidov I think it's very likely that was the culprit.
Thank you, Alexey | 2020-08-26T10:38:50 | 2020-08-26T10:38:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,832,372 | 13,964 | zhang2014 | > Changelog entry does not state what bug was present and what was changed.
>
>I will edit by myself...
Done. | 2020-08-26T11:53:23 | 2020-08-26T11:53:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,899,305 | 14,111 | alexey-milovidov | I remove "bug" label because no bug exists. | 2020-08-26T14:02:40 | 2020-08-26T14:02:40 | {} | MEMBER |
ClickHouse | ClickHouse | 680,980,644 | 14,002 | akuzm | Some workarounds that might help:
1. Don't build it at all, download a pre-built binary from the CI as described here: https://clickhouse.tech/docs/en/development/build/#you-dont-have-to-build-clickhouse
The configuration you need is `clang-10-aarch64`.
2. Use `clang-10`. This is the way ClickHouse is compiled f... | 2020-08-26T16:17:37 | 2020-08-26T16:17:37 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,028,536 | 14,121 | alexey-milovidov | null('table structure'), example: `null('x UInt64')`
\- creates StorageNull with that structure. | 2020-08-26T17:47:34 | 2020-08-26T17:47:52 | {} | MEMBER |
ClickHouse | ClickHouse | 681,120,051 | 12,413 | akuzm | > I've changed "bugfix" to "improvement" so no one will dare to backport this insignificant mostly unneeded useless fix.
The bots fight you again... better use `pr-no-backport`. | 2020-08-26T20:57:58 | 2020-08-26T20:57:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,831,749 | 14,045 | filimonov | please stop the server with kill command
```
kill -9 $(pidof -s clickhouse-server)
```
or just restart the system.
Check if your system working properly (maybe some os/memory/disk or similar issues - check logs, `dmesg` etc). | 2020-08-26T11:51:54 | 2020-08-26T11:52:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 680,853,042 | 14,002 | rubentorresbonet | Any update on this? Clickhouse doesn't run on arm at all because of this. | 2020-08-26T12:37:53 | 2020-08-26T12:37:53 | {} | NONE |
ClickHouse | ClickHouse | 681,033,927 | 12,413 | alexey-milovidov | I've changed "bugfix" to "improvement" so no one will dare to backport this insignificant mostly unneeded useless fix. | 2020-08-26T17:58:26 | 2020-08-26T17:58:26 | {} | MEMBER |
ClickHouse | ClickHouse | 681,050,388 | 13,956 | excitoon | > We shall not merge that for a while, I have few failing tests on my local machine.
My failing tests were caused by OOMs, everything is fine now. | 2020-08-26T18:31:40 | 2020-08-26T18:31:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 681,066,955 | 14,129 | CurtizJ | > I've read the fix but I don't understand how it's related to LowCardinality
The problem appeared, when we read dictionary of low cardinality. Since it has low size, a lot of marks point to one place, and previously we a lot of time was spent in loop, while trying to find mark to next compressed block. | 2020-08-26T19:05:36 | 2020-08-26T19:05:36 | {} | MEMBER |
ClickHouse | ClickHouse | 680,711,661 | 14,045 | radiatorius | unfortunately, after 3-4 hours of work the **Segmentation fault** was repeated again:
```(version 20.5.4.40 (official build))
2020.08.25 22:11:21.144249 [ 10397 ] {} <Error> void DB::BackgroundProcessingPool::workLoopFunc(): Code: 43, e.displayText() = DB::Exception: Illegal type String of argument of function toYYYY... | 2020-08-26T07:32:14 | 2020-08-26T08:19:18 | {} | NONE |
ClickHouse | ClickHouse | 680,796,554 | 14,043 | alexey-milovidov | It's not possible to make last not inclusive, because it will lead to race condition on parallel insertions:
- it may point to the `end` of list;
- `end` is not invalidated on insertions;
- but the saved `end` iterator always points to the list end regardless of new inserted elements;
- so the last range will "exte... | 2020-08-26T10:27:29 | 2020-08-26T10:29:09 | {} | MEMBER |
ClickHouse | ClickHouse | 681,099,790 | 14,045 | alexey-milovidov | Also please provide some info about your setup:
- cluster configuration and size;
- structures of the relevant tables;
- some example of queries;
- try to describe your usage scenario, especially: foreign tables, distributed queries, mutation queries, dictionaries, join tables... | 2020-08-26T20:14:24 | 2020-08-26T20:14:33 | {} | MEMBER |
ClickHouse | ClickHouse | 676,218,544 | 13,097 | 4ertus2 | There were several attempts to fix unbundled build in this PR. The reason of fail is hack for `numeric_limits<__int128>` is not working anymore.
There was a pef degradation in UUID hashes. Improved by perf optimization. There're some other strange perf degradations - not related to this patch. | 2020-08-19T11:51:09 | 2020-08-19T11:51:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 676,435,412 | 13,190 | blinkov | @Slach there're some failed tests:
```
WARNING:mkdocs.structure.pages:Documentation file 'operations/server-configuration-parameters/settings.md' contains a link to 'operations/system-tables.md' which is not found in the documentation files.
WARNING:mkdocs.structure.pages:Documentation file 'operations/server-config... | 2020-08-19T14:19:36 | 2020-08-19T14:19:36 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.