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 | 650,575,107 | 11,940 | alexey-milovidov | @KochetovNicolai Looks like an old bug around LowCardinality. | 2020-06-27T15:28:08 | 2020-06-27T15:28:08 | {} | MEMBER |
ClickHouse | ClickHouse | 650,404,144 | 10,102 | filimonov | That was the fix probably https://github.com/ClickHouse/ClickHouse/pull/10791 | 2020-06-26T21:13:18 | 2020-06-26T21:13:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,500,150 | 11,986 | filimonov | > You could try setting your consumer group offsets such that they don't attempt to consume old messages (e.g. offset=3 for the example above)
May be we can allow using default value (by special setting ) if the requested column is absent in the Avro message instead of exception 'Field brand_id not found in Avro sch... | 2020-06-27T05:49:06 | 2020-06-27T05:49:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,507,696 | 11,498 | azat | @vitlibar
>You can store any part of the context you need to perform asynchronous INSERT beforehand. This issue doesn't seem to be difficult to solve.
Yes, but there were worries (from @abyss7 ) even about storing password in RAM, so storing password on disk looks questionable.
>provide access if both initial... | 2020-06-27T06:41:54 | 2020-06-27T06:41:54 | {} | MEMBER |
ClickHouse | ClickHouse | 650,580,686 | 11,971 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-27T16:14:25 | 2020-06-27T16:14:25 | {} | MEMBER |
ClickHouse | ClickHouse | 650,583,579 | 11,995 | azat | > test_adaptive_granularity/test.py::test_version_update_two_nodes
Is in ci fest | 2020-06-27T16:36:05 | 2020-06-27T16:36:05 | {} | MEMBER |
ClickHouse | ClickHouse | 650,749,898 | 12,004 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-28T12:50:25 | 2020-06-28T12:50:25 | {} | MEMBER |
ClickHouse | ClickHouse | 650,335,908 | 11,498 | vitlibar | > * right now protocol does not support user change w/o reconnects, so you need to issue reconnect (reconnect is not a huge problem I guess, and also protocol can be expanded to support this, or connection pool can be created for each user)
> * not always there is user/password while executing distributed query (see [... | 2020-06-26T18:42:16 | 2020-06-26T18:42:16 | {} | MEMBER |
ClickHouse | ClickHouse | 650,385,703 | 8,167 | alexey-milovidov | @CurtizJ I don't remember whether it was fixed. | 2020-06-26T20:29:41 | 2020-06-26T20:29:41 | {} | MEMBER |
ClickHouse | ClickHouse | 650,419,420 | 11,498 | vitlibar | @filimonov Ok, it sounds reasonable, thanks for the explanation. | 2020-06-26T21:51:28 | 2020-06-26T21:51:28 | {} | MEMBER |
ClickHouse | ClickHouse | 650,608,000 | 11,928 | alexey-milovidov | No need to backport. | 2020-06-27T19:22:06 | 2020-06-27T19:22:06 | {} | MEMBER |
ClickHouse | ClickHouse | 650,617,006 | 12,001 | alexey-milovidov | It helps but CI reports error nevertheless. | 2020-06-27T20:17:38 | 2020-06-27T20:17:38 | {} | MEMBER |
ClickHouse | ClickHouse | 650,673,241 | 11,629 | jackpgao | I've already remove the corrupted partition, so I cannot reproduce it. | 2020-06-28T02:19:23 | 2020-06-28T02:19:23 | {} | NONE |
ClickHouse | ClickHouse | 650,289,786 | 11,498 | vitlibar | If I understand the problem correctly you'd like to use `initial_user` (the user who logins and executes a query, not cluster's user) to check access rights on cluster's nodes. It seems to me there is an easier solution: just add an option to cluster's definition like
```
<remote_servers>
<test_cluster>
... | 2020-06-26T17:04:23 | 2020-06-26T17:06:00 | {} | MEMBER |
ClickHouse | ClickHouse | 650,394,057 | 11,401 | akuzm | > > OSWriteChars was the only suspicious metric, so I turned off the query profiler, and now there are almost no changes in performance ¯_(ツ)_/¯
> > Results here: https://clickhouse-test-reports.s3.yandex.net/11401/ae1bc3cb27225afe6d20e9beb38cad1e4643049c/performance_comparison/report.html#fail1
>
> Due to a bug in... | 2020-06-26T20:48:39 | 2020-06-26T20:48:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,404,549 | 5,131 | filimonov | Fix #10791 | 2020-06-26T21:14:24 | 2020-06-26T21:14:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,437,786 | 11,986 | oandrew | @SergeyIBM47
Schema id and schema are immutable.
Schemas are cached by their `id`.
Each kafka message contains its schema id which is used to find the schema in the registry.
Most likely you get this error because you are trying to read old messages (with old schema) which do not have new field in their schema... | 2020-06-26T22:51:14 | 2020-06-26T23:06:50 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,486,088 | 11,990 | den-crane | 1. It's not recommended to share CH with ZK. CH uses all hosts resources and it cases ZK issues.
2. Check CH logs for insertion errors (/var/log/clickhouse/clickhouse.err.log /var/log/clickhouse/clickhouse.log ) .
3. Replace a sharding key `rand()` with something more stable i.e. `id` because rand makes your insert... | 2020-06-27T04:06:48 | 2020-06-27T04:06:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,507,984 | 10,564 | azat | So, looks like this can be closed? (after https://github.com/ClickHouse/ClickHouse/issues/10564#issuecomment-646296199) | 2020-06-27T06:43:55 | 2020-06-27T06:43:55 | {} | MEMBER |
ClickHouse | ClickHouse | 650,595,672 | 11,997 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-27T18:17:13 | 2020-06-27T18:17:13 | {} | MEMBER |
ClickHouse | ClickHouse | 650,590,512 | 11,986 | SergeyIBM47 | @oandrew
Unfortunately, I read new messages from a topic which contains a new schema id. I have several times changed the consumer group id and even drop old schemas from the schema registry and underlying topic for test purpose. I get the behaviour the same as I described above - the new schema wasn't fetched.
B... | 2020-06-27T17:41:28 | 2020-06-27T19:51:11 | {} | NONE |
ClickHouse | ClickHouse | 650,619,192 | 5,843 | alexey-milovidov | @nvartolomei I think we should add virtual method in IDataType to validate a column.
For most types it will be noop, but for some types it will throw exception if some inconsistency found. | 2020-06-27T20:26:47 | 2020-06-27T20:26:47 | {} | MEMBER |
ClickHouse | ClickHouse | 650,162,534 | 11,950 | KochetovNicolai | Not waiting for tests because only comment was changed. | 2020-06-26T12:52:41 | 2020-06-26T12:52:41 | {} | MEMBER |
ClickHouse | ClickHouse | 650,385,977 | 11,498 | filimonov | > reconnect is not a huge problem I guess
- Why do we use connection pool at all in that case? :)
- There are always an extra cost for establishing the connection. Imagine thousands req/sec.
> Separate connection pools sound better than reconnections at this point.
100 users and cluster 20 shards x 2 replicas... | 2020-06-26T20:30:20 | 2020-06-26T20:30:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,388,513 | 10,102 | alexey-milovidov | Wow, it's fixed! | 2020-06-26T20:36:04 | 2020-06-26T20:36:04 | {
"rocket": 1
} | MEMBER |
ClickHouse | ClickHouse | 650,574,778 | 11,791 | alexey-milovidov | @4ertus2 It is easily reproducible in master. | 2020-06-27T15:25:16 | 2020-06-27T15:25:16 | {} | MEMBER |
ClickHouse | ClickHouse | 650,581,628 | 11,947 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-27T16:20:44 | 2020-06-27T16:20:44 | {} | MEMBER |
ClickHouse | ClickHouse | 650,594,951 | 11,995 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-27T18:14:44 | 2020-06-27T18:14:44 | {} | MEMBER |
ClickHouse | ClickHouse | 650,642,729 | 11,987 | siradjev | > There was an idea to add dedicated `comment` column to query log and to allow to specify comment in HTTP and native interfaces.
that would work too, but if there are naturally multiple comments in multiple places, how would it look in logs? is there an associated issue for that? | 2020-06-27T22:42:49 | 2020-06-27T22:42:49 | {} | NONE |
ClickHouse | ClickHouse | 650,718,943 | 10,389 | adegtyarev | ```
┌─name──────┬─type───────────────────┬─default_type─┬─default_expression─┬─comment─┬─codec_expression──┬─ttl_expression─┐
│ timestamp │ DateTime('UTC') │ │ │ │ Delta(4), ZSTD(1) │ │
│ date │ Date │ DEFAULT │ toDate(timestam... | 2020-06-28T08:42:33 | 2020-06-28T08:42:33 | {} | NONE |
ClickHouse | ClickHouse | 650,196,496 | 11,978 | filimonov | I don't see any issue here. What exactly is the problem? | 2020-06-26T14:01:48 | 2020-06-26T14:02:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,393,737 | 5,131 | alexey-milovidov | It is fixed in recent ClickHouse releases. | 2020-06-26T20:47:52 | 2020-06-26T20:47:52 | {} | MEMBER |
ClickHouse | ClickHouse | 650,397,616 | 11,986 | filimonov | /cc @oandrew | 2020-06-26T20:57:29 | 2020-06-26T20:57:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,467,374 | 11,992 | l1t1 | (version 20.4.5) | 2020-06-27T01:06:14 | 2020-06-27T01:06:14 | {} | NONE |
ClickHouse | ClickHouse | 650,555,370 | 11,903 | alexey-milovidov | @blinkov This PR failed in Description check:
```
Changelog entry is not found.
```
You have to either provide the description for changelog or choose "Not for changelog".
I will fix this mistake for you. | 2020-06-27T12:39:29 | 2020-06-27T12:39:29 | {} | MEMBER |
ClickHouse | ClickHouse | 650,571,025 | 11,996 | bohutang | Yes.
That's the universal solution for all the external engines(mongodb, cassandra etc) if they have Enum types.
Since mysqlxx(other storage driver) returns real enum string result not the enum type value, so need more works to do with it(need convert the string result to the Enum8 index) | 2020-06-27T14:53:58 | 2020-06-27T14:53:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,576,369 | 11,985 | alexey-milovidov | The crash happens only in debug build (we translate all exceptions with `LOGICAL_ERROR` to std::terminate, but only in debug build).
| 2020-06-27T15:39:17 | 2020-06-27T15:39:17 | {} | MEMBER |
ClickHouse | ClickHouse | 650,523,807 | 11,987 | siradjev | As a workaround, it would also be reasonable if we can support some custom key, so that it can be found from query text. It would be less intrusive than using query text itself.
E.g. if you define some sort of custom_identifiers allowed to be passed as settings and then recorded in settings column or additional custo... | 2020-06-27T08:36:16 | 2020-06-27T08:36:16 | {} | NONE |
ClickHouse | ClickHouse | 650,575,222 | 11,940 | alexey-milovidov | @iameugenejo
```
CREATE TABLE test.logs (
`ingest_time` DateTime,
`start` DateTime, `service`
`client_id` LowCardinality(String),
`metric_value` UInt64) ENGINE = Distributed(shards, 'test', logs_buffer, rand())
```
This query is incorrect, you forgot to provide a type for `service` column. | 2020-06-27T15:29:08 | 2020-06-27T15:29:08 | {} | MEMBER |
ClickHouse | ClickHouse | 650,595,563 | 11,998 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-27T18:16:52 | 2020-06-27T18:16:52 | {} | MEMBER |
ClickHouse | ClickHouse | 650,607,439 | 11,996 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-27T19:18:21 | 2020-06-27T19:18:21 | {} | MEMBER |
ClickHouse | ClickHouse | 650,701,172 | 8,789 | zhuyoucui0909 | @jason-godden Could you help to explain more how to insert into aggregatingMergeTree table data, without storing raw data? If possible, could you provide an example?
https://github.com/ClickHouse/ClickHouse/issues/8789#issuecomment-577472482
Is it possible to do it by go client? https://github.com/ClickHouse/clickhou... | 2020-06-28T05:49:56 | 2020-06-28T05:51:59 | {} | NONE |
ClickHouse | ClickHouse | 650,749,617 | 12,003 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-28T12:49:13 | 2020-06-28T12:49:13 | {} | MEMBER |
ClickHouse | ClickHouse | 650,749,807 | 12,003 | alexey-milovidov | @akuzm Any ideas on how to automate this?
I have tried `ispell` or `aspell` but it has too many false positives. | 2020-06-28T12:49:59 | 2020-06-28T12:49:59 | {} | MEMBER |
ClickHouse | ClickHouse | 650,212,199 | 11,932 | santoshdvn | I am trying to ingest ORC files from HDFS into Clickhouse. I tried different compression types(None,Snappy,zlib). I wast not able to ingest data in Clickhouse , where as I could able to ingest parquet files from HDFS to Clickhouse tables. This is how I am loading parquet files
hdfs dfs -cat <hdfs_parquetfiles> | ~... | 2020-06-26T14:35:06 | 2020-06-26T14:35:06 | {} | NONE |
ClickHouse | ClickHouse | 650,314,005 | 11,974 | alexey-milovidov | Flaky integration test:
test_adaptive_granularity/test.py::test_version_update_two_nodes
https://clickhouse-test-reports.s3.yandex.net/11974/2a1937625b0abf0ed8244c90284af15b57e8814e/integration_tests_(asan).html
I see it for the first time.
| 2020-06-26T17:52:29 | 2020-06-26T17:52:29 | {} | MEMBER |
ClickHouse | ClickHouse | 650,371,178 | 11,987 | alexey-milovidov | The difficulty is that query_id is generated too early - before parsing stage.
That's why we are able to report query_id in case of parse errors. | 2020-06-26T19:57:55 | 2020-06-26T19:58:17 | {} | MEMBER |
ClickHouse | ClickHouse | 650,541,898 | 11,994 | l1t1 | thanks @den-crane | 2020-06-27T10:49:38 | 2020-06-27T10:49:38 | {} | NONE |
ClickHouse | ClickHouse | 650,573,922 | 11,903 | alexey-milovidov | FYI This PR breaks integration tests in master. | 2020-06-27T15:17:47 | 2020-06-27T15:17:47 | {} | MEMBER |
ClickHouse | ClickHouse | 650,574,501 | 11,629 | alexey-milovidov | It is possible but should not lead to crash anyway. | 2020-06-27T15:22:49 | 2020-06-27T15:22:49 | {} | MEMBER |
ClickHouse | ClickHouse | 650,581,574 | 11,973 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-27T16:20:19 | 2020-06-27T16:20:19 | {} | MEMBER |
ClickHouse | ClickHouse | 650,749,500 | 12,002 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-28T12:48:47 | 2020-06-28T12:48:47 | {} | MEMBER |
ClickHouse | ClickHouse | 650,487,373 | 11,992 | den-crane | >but I run it again, it returns
do you run it the second time immediately and get another result?
It's looks like a normal behavior. CH client by default shows a query result splitted into blocks. You can change it by using another output format https://clickhouse.tech/docs/en/interfaces/formats/#prettycompactmo... | 2020-06-27T04:15:34 | 2020-06-27T04:39:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,580,649 | 11,975 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-27T16:14:08 | 2020-06-27T16:14:08 | {} | MEMBER |
ClickHouse | ClickHouse | 650,607,928 | 11,923 | alexey-milovidov | Two functional tests still has to be updated.
> Integration tests
This is Ok (Broken in #11903) | 2020-06-27T19:21:34 | 2020-06-27T19:21:34 | {} | MEMBER |
ClickHouse | ClickHouse | 650,608,171 | 11,912 | alexey-milovidov | Sorry, I don't understand the meaning of this PR. | 2020-06-27T19:23:06 | 2020-06-27T19:23:06 | {} | MEMBER |
ClickHouse | ClickHouse | 650,641,039 | 11,378 | alexey-milovidov | That's a great find! | 2020-06-27T22:29:25 | 2020-06-27T22:29:25 | {} | MEMBER |
ClickHouse | ClickHouse | 650,644,026 | 11,986 | oandrew | @SergeyIBM47
Schemas are only cached in memory and it's ok because schemas are immutable.
Keep in mind you can't change the schema of old messages because kafka messages are also immutable.
Also, what do you mean by "old schema removed from schema registry"?
If you were referring to `DELETE /subjects/(string: s... | 2020-06-27T22:52:17 | 2020-06-27T22:52:44 | {
"rocket": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,146,185 | 11,979 | hustnn | @alexey-milovidov Thanks a lot. | 2020-06-26T12:12:02 | 2020-06-26T12:12:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,195,197 | 11,978 | filimonov | It looks like a normal exception, not a segfault. | 2020-06-26T13:59:14 | 2020-06-26T13:59:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,265,920 | 11,985 | zlobober | Just to mention: I expect these comparisons to work and not any kind of "comparing nullables inside tuples is not supported" error :) | 2020-06-26T16:16:36 | 2020-06-26T16:16:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,387,784 | 5,912 | alexey-milovidov | It is also known as INSERT RETURNING but I'm not sure whether it is applicable for INSERT SELECT. | 2020-06-26T20:34:33 | 2020-06-26T20:34:33 | {} | MEMBER |
ClickHouse | ClickHouse | 650,410,544 | 8,167 | CurtizJ | It was fixed in #11042. | 2020-06-26T21:29:55 | 2020-06-26T21:29:55 | {} | MEMBER |
ClickHouse | ClickHouse | 650,489,881 | 11,498 | filimonov | > Cluster nodes check two filters for row policies - both for cluster's user and for initial user.
Interesting, i was not aware of that.
> both initial user and cluster's user on that node have corresponding grants.
That means that cluster user should have a superset of rights of all possible initial_users (i... | 2020-06-27T04:34:52 | 2020-06-27T05:25:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,563,599 | 10,679 | cw9 | Thanks for adding this!
| 2020-06-27T13:49:26 | 2020-06-27T13:49:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,575,271 | 12,000 | alesapin | The image already updated manually. | 2020-06-27T15:29:36 | 2020-06-27T15:29:36 | {} | MEMBER |
ClickHouse | ClickHouse | 650,726,628 | 11,347 | TCeason | ```
:)show create database a ;
SHOW CREATE DATABASE a
┌─statement───────────────────────────┐
│ CREATE DATABASE a ENGINE = Ordinary │
└─────────────────────────────────────┘
1 rows in set. Elapsed: 0.001 sec.
:)detach database a ;
DETACH DATABASE a
3 rows in set. Elapsed: 0.109 sec.
:... | 2020-06-28T09:49:59 | 2020-06-28T09:49:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,380,626 | 11,103 | alexey-milovidov | @KochetovNicolai Maybe the new Query Plan will help? | 2020-06-26T20:17:44 | 2020-06-26T20:17:44 | {} | MEMBER |
ClickHouse | ClickHouse | 650,497,952 | 9,568 | filimonov | Check https://github.com/ClickHouse/ClickHouse/issues/8167#event-3488197913 and #10128 p.4
| 2020-06-27T05:32:31 | 2020-06-27T05:34:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,543,575 | 8,624 | nikitamikhaylov | Merged in #11947 | 2020-06-27T11:00:49 | 2020-06-27T11:00:49 | {} | MEMBER |
ClickHouse | ClickHouse | 650,574,912 | 11,791 | alexey-milovidov | @4ertus2 Should we hide `partial_merge` option inside experimental flag if it does not have acceptable quality? | 2020-06-27T15:26:20 | 2020-06-27T15:26:20 | {} | MEMBER |
ClickHouse | ClickHouse | 650,590,179 | 11,940 | iameugenejo | @alexey-milovidov That was a copy paste issue. service shouldn't be there. | 2020-06-27T17:38:22 | 2020-06-27T17:38:22 | {} | NONE |
ClickHouse | ClickHouse | 650,617,401 | 12,001 | alexey-milovidov | I will try to delete this test completely. | 2020-06-27T20:19:01 | 2020-06-27T20:19:01 | {} | MEMBER |
ClickHouse | ClickHouse | 650,412,870 | 11,397 | filimonov | Similar for Delta: #4609 | 2020-06-26T21:32:39 | 2020-06-26T21:32:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,469,354 | 11,994 | l1t1 | and the inner query runs ok too
```
select left(toString(ad_code),2)code, case when (create_time > '2020-04-30 23:59:59') then 1 else 0 end k,admin_status,sum(1)s
from household_short
group by left(toString(ad_code),2) ,admin_status,case when (create_time > '2020-04-30 23:59:59') then 1 else 0 end
```
result
... | 2020-06-27T01:23:33 | 2020-06-27T01:23:33 | {} | NONE |
ClickHouse | ClickHouse | 650,498,783 | 11,992 | filimonov | Duplicate #118 #6860 #4948 | 2020-06-27T05:38:40 | 2020-06-27T05:38:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,612,980 | 7,529 | alexey-milovidov | Fixed in master (version 20.6) in PR #11745. | 2020-06-27T19:56:00 | 2020-06-27T19:56:20 | {} | MEMBER |
ClickHouse | ClickHouse | 650,419,356 | 11,059 | filimonov | > i've finished the optimization of the merge algorithm of replacingmerge tree and get 30% speedup on final query , and the remained work is not very difficult.
Do you plan to share your improvements / findings? | 2020-06-26T21:51:17 | 2020-06-26T21:51:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,419,469 | 11,498 | vitlibar | I have another idea. Cluster nodes check two filters for row policies - both for cluster's user and for initial user. It seems all we have to do to solve this task is to check access rights in the same way, i.e. provide access if both initial user and cluster's user on that node have corresponding grants. | 2020-06-26T21:51:37 | 2020-06-26T21:51:37 | {} | MEMBER |
ClickHouse | ClickHouse | 650,562,961 | 10,564 | cw9 | I have not tried the feature yet, still waiting for the release, but feel free to close for now. Thanks for the help! | 2020-06-27T13:43:14 | 2020-06-27T13:43:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,574,630 | 11,629 | alexey-milovidov | Does it reproduce on 20.3.12 or 20.4.6? | 2020-06-27T15:23:57 | 2020-06-27T15:23:57 | {} | MEMBER |
ClickHouse | ClickHouse | 650,580,916 | 11,971 | alexey-milovidov | > Functional stateful tests (address) — No output log
An error with Docker registry.
> Unable to find image 'yandex/clickhouse-stateful-test:latest' locally
docker: Error response from daemon: Get https://registry-1.docker.io/v2/: EOF.
See 'docker run --help'.
| 2020-06-27T16:15:59 | 2020-06-27T16:15:59 | {} | MEMBER |
ClickHouse | ClickHouse | 650,607,517 | 11,988 | alexey-milovidov | > Integration tests
Broken in #11903 | 2020-06-27T19:18:53 | 2020-06-27T19:18:53 | {} | MEMBER |
ClickHouse | ClickHouse | 650,734,432 | 11,590 | azat | >@alexey-milovidov AFAICS the problem is not in the middle-sized but in allocations >~256K (depends on the number of classes for fast path, this depends on the size of pages is used in tcmalloc, tried 8k/256k not helps a lot), and it does not looks like it can be fixed easily.
https://github.com/google/tcmalloc/issu... | 2020-06-28T11:00:21 | 2020-06-28T11:00:21 | {} | MEMBER |
ClickHouse | ClickHouse | 650,750,164 | 12,001 | alexey-milovidov | It did not help. | 2020-06-28T12:51:39 | 2020-06-28T12:51:39 | {} | MEMBER |
ClickHouse | ClickHouse | 650,157,778 | 11,903 | alesapin | Checked manually. It woks. | 2020-06-26T12:41:51 | 2020-06-26T12:41:51 | {} | MEMBER |
ClickHouse | ClickHouse | 650,314,139 | 11,815 | alexey-milovidov | test_adaptive_granularity/test.py::test_version_update_two_nodes
https://clickhouse-test-reports.s3.yandex.net/11974/2a1937625b0abf0ed8244c90284af15b57e8814e/integration_tests_(asan).html | 2020-06-26T17:52:52 | 2020-06-26T17:52:52 | {} | MEMBER |
ClickHouse | ClickHouse | 650,399,366 | 11,932 | filimonov | According to the exception problem is not with reading / parsing ORC (which is supported) but with writing ORC.
What is the definitions of table in which you trying to insert the data? | 2020-06-26T21:01:38 | 2020-06-26T21:01:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,489,896 | 11,994 | den-crane | CH does not have full support for SQL `case` syntax.
```
select (case a when 1 then a else 0 end ) from (select materialize(1) a)
DB::Exception: Argument at index 2 for function transform must be constant.
select (case when a=1 then a else 0 end ) from (select materialize(1) a)
--
1
```
you can rewrite it... | 2020-06-27T04:35:03 | 2020-06-27T04:36:22 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,548,421 | 11,987 | alexey-milovidov | There was an idea to add dedicated `comment` column to query log and to allow to specify comment in HTTP and native interfaces. | 2020-06-27T11:36:53 | 2020-06-27T11:36:53 | {} | MEMBER |
ClickHouse | ClickHouse | 650,607,653 | 11,983 | alexey-milovidov | > Integration tests
Broken in #11903
> Performance — 17 slower
Looks like a usual flap. | 2020-06-27T19:19:47 | 2020-06-27T19:19:47 | {} | MEMBER |
ClickHouse | ClickHouse | 650,618,307 | 9,523 | alexey-milovidov | Your question: is it possible to execute distributed queries out of order?
No, we don't know what are possible dependencies between queries, so it will be unsafe in general. | 2020-06-27T20:22:55 | 2020-06-27T20:22:55 | {} | MEMBER |
ClickHouse | ClickHouse | 650,331,158 | 11,986 | SergeyIBM47 |
This is parts of my error logs, I hope it helps you:
`020.06.26 15:13:30.824286 [ 269 ] {} <Error> DynamicQueryHandler: Code: 8, e.displayText() = DB::Exception: Field brand_id not found in Avro schema: While executing SourceFromInputStream, Stack trace (when copying this message, always include the lines below):... | 2020-06-26T18:31:27 | 2020-06-26T18:31:27 | {} | NONE |
ClickHouse | ClickHouse | 650,543,143 | 11,947 | nikitamikhaylov | Error in `test_adaptive_granularity/test.py::test_version_update_two_nodes` is not related to this PR. | 2020-06-27T10:57:57 | 2020-06-27T10:57:57 | {} | MEMBER |
ClickHouse | ClickHouse | 650,571,697 | 11,996 | bohutang | BTW, thanks for giving 'can be tested' free pass. | 2020-06-27T14:59:14 | 2020-06-27T14:59:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 650,581,708 | 9,711 | alexey-milovidov | Could you please also update `clickhouse-test` and integration tests to Python 3? | 2020-06-27T16:21:16 | 2020-06-27T16:21:16 | {} | MEMBER |
ClickHouse | ClickHouse | 650,613,449 | 8,550 | alexey-milovidov | The issue is no longer relevant. I will add a test. | 2020-06-27T19:58:46 | 2020-06-27T19:58:46 | {} | MEMBER |
ClickHouse | ClickHouse | 650,616,697 | 12,000 | alexey-milovidov | It did not help. | 2020-06-27T20:16:28 | 2020-06-27T20:16:28 | {} | MEMBER |
ClickHouse | ClickHouse | 637,594,116 | 11,383 | alexey-milovidov | Format `Regexp` will be mostly sufficient. | 2020-06-02T14:51:46 | 2020-06-02T14:51:46 | {} | MEMBER |
ClickHouse | ClickHouse | 637,999,158 | 11,396 | YiuRULE | If you want to set the same record on all of your server, why not just set a replicated table on each of your server ? | 2020-06-03T07:00:07 | 2020-06-03T07:00:07 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.