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 | 753,222,896 | 16,575 | den-crane | @garneja you need to `set enable_global_with_statement=1`
--enable_global_with_statement arg Propagate WITH statements to UNION queries and all subqueries
```sql
set enable_global_with_statement=1;
with 5 as q1, x as (select number as a from numbers(10)), y as (select number+100 as b, number as a f... | 2020-12-31T23:11:05 | 2020-12-31T23:11:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,330,861 | 18,655 | alexey-milovidov | But we have to fix bugs, otherwise it will make constant noise in our CI system. | 2021-01-01T15:41:11 | 2021-01-01T15:41:11 | {} | MEMBER |
ClickHouse | ClickHouse | 753,421,964 | 4,911 | kevinbin | clickhouse client --date_time_input_format=best_effort --query="INSERT INTO data.pollution_measurement FORMAT CSV" < Measurement_summary.csv
Code: 41. DB::Exception: Cannot read DateTime: unexpected word: (at row 1)
Row 1:
Column 0, name: Measurement date, type: DateTime, parsed text: "Mea"ERROR: DateTime ... | 2021-01-02T02:57:33 | 2021-01-02T02:57:33 | {} | NONE |
ClickHouse | ClickHouse | 753,561,727 | 18,264 | ucasfl | > > The difference is that cluster/remote uses sync inserts, and if you will try something like this you will see the race too:
>
> And this means that the race is in the upstream too.
>
> I guess it is totally fine to copy the query AST in the InterpreterInsertQuery, i.e.
>
> ```diff
> $ git di
> diff --git... | 2021-01-03T03:18:30 | 2021-01-03T03:18:30 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 753,650,560 | 18,694 | den-crane | ```sql
create table test(i Int64, t DateTime) Engine=Memory;
set date_time_input_format='best_effort';
insert into test values(1, '2020-11-07T20:45:00.000+3:00');
insert into test select 2, parseDateTimeBestEffort('2020-11-07T20:45:00.000+3:00');
select * from test;
┌─i─┬───────────────────t─┐
│ 2 │ 20... | 2021-01-03T17:34:08 | 2021-01-03T17:34:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,995,503 | 17,455 | den-crane | > > I edited table ZK node and it somehow got into sync. Maybe also I've reattached the table. Did not have an opportunity to find solid steps yet.
>
> The metadata in ZK seems correct to me! And unfortunately, re-attaching the table didn't help.
metadata in ZK is stored in two places. In the table node and in th... | 2020-12-31T16:06:08 | 2020-12-31T16:06:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,058,191 | 17,455 | OmarBazaraa | > Why do you have 2 different versions in the cluster?
I updated the cluster today after the issue occurred. But all nodes are having the same version now. | 2020-12-31T19:54:28 | 2020-12-31T19:54:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,221,584 | 18,676 | den-crane | @amosbird | 2020-12-31T22:58:05 | 2020-12-31T22:58:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,406,849 | 18,664 | l1t1 | systemctl start clickhouse ? | 2021-01-02T00:03:41 | 2021-01-02T00:03:41 | {} | NONE |
ClickHouse | ClickHouse | 753,583,042 | 18,264 | azat | >I have tried this, but still have the same problem.
I see, so after fixing that data race new one pops up, if you using Log engine?
With the Log engine there should be deadlock, not data race, and it should be fixed by #15260
And anyway this problem is not related to this PR. | 2021-01-03T08:10:24 | 2021-01-03T08:10:24 | {} | MEMBER |
ClickHouse | ClickHouse | 752,966,671 | 18,670 | OmarBazaraa | I tried to restart the servers, but they couldn't start as there were too many broken parts.
So, I increased `max_suspicious_broken_parts` and was able to start them.
Now, when I try to drop the index I get the following message:
```
Code: 517. DB::Exception: Received from localhost:9440. DB::Exception: Metadat... | 2020-12-31T14:04:27 | 2020-12-31T14:04:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,970,065 | 17,646 | alexey-milovidov | It is for "memory tracker blocker" and means "block everything"? | 2020-12-31T14:17:38 | 2020-12-31T14:17:43 | {} | MEMBER |
ClickHouse | ClickHouse | 752,983,653 | 8,052 | filimonov | Fixed in #18637 | 2020-12-31T15:06:35 | 2020-12-31T15:06:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,998,727 | 17,455 | OmarBazaraa | Well, let me give some context...
I have a cluster of 2 shards, 3 replicas each. In that cluster I have a table with some data skipping indexes.
When I attempted to drop those indexes using DDL queries, the indexes in the second shard got deleted, but those in the first shard couldn't be dropped.
And here is the... | 2020-12-31T16:22:22 | 2020-12-31T16:22:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,431,112 | 9,030 | amosbird | https://github.com/ClickHouse/ClickHouse/pull/17726 | 2021-01-02T04:46:23 | 2021-01-02T04:46:23 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 753,605,317 | 18,623 | zhangguiyu | Sorry, I can't seem to replicate this bug in clickhouse-server 20.11.4 anymore:
e.g., the following now works:
```
CREATE MATERIALIZED VIEW IF NOT EXISTS bug_mv
ENGINE = MergeTree
ORDER BY pmid
POPULATE AS
-- INSERT INTO ... | 2021-01-03T11:48:55 | 2021-01-03T11:48:55 | {} | NONE |
ClickHouse | ClickHouse | 753,650,640 | 18,369 | filimonov | > I've verified one more time and even if the initiator/sender does not checks the checksum the remote side should send the exception back and sender will rethrow the same exception, and the batch should be marked as broken.
Reciever never send an exception if sender has file smaller than it should be. | 2021-01-03T17:35:05 | 2021-01-03T17:35:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,951,995 | 18,668 | filimonov | You can pull my branch, or wait for CI build in #18671 | 2020-12-31T12:53:12 | 2020-12-31T12:53:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,003,821 | 17,455 | OmarBazaraa | > what about replica 2 and 3 ?
>
> get /clickhouse/tables/clarity/page_data_local_2/1-1/replicas/2/metadata
> get /clickhouse/tables/clarity/page_data_local_2/1-1/replicas/3/metadata
All replicas of shard 1 have the same metadata (i.e. still with data skipping indexes), and all replicas of shard 2 have the same ... | 2020-12-31T16:48:27 | 2020-12-31T16:48:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,370,072 | 18,681 | den-crane | https://github.com/ClickHouse/ClickHouse/issues/9684 | 2021-01-01T19:15:09 | 2021-01-01T19:15:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,509,913 | 14,893 | pingyu | I'd like to have a try.
Could you please assign to me ? @alexey-milovidov | 2021-01-02T18:05:16 | 2021-01-02T18:05:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,969,680 | 17,646 | alexey-milovidov | > Global - ignore any memory limits
BTW, why it's named "Global" (not something like "None")? It can be mistaken with server context. | 2020-12-31T14:16:11 | 2020-12-31T14:16:11 | {} | MEMBER |
ClickHouse | ClickHouse | 752,971,483 | 18,649 | den-crane | CH relays on the system file cache.
In case of allocating max_server_memory_usage (0.9*RAM) system file cache will be tiny and CH will be slow. | 2020-12-31T14:22:56 | 2020-12-31T14:23:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,054,021 | 17,455 | den-crane | Cannot alter (version 20.10.3.30
Cannot alter (version 20.12.5.14
Why do you have 2 different versions in the cluster? | 2020-12-31T19:51:31 | 2020-12-31T19:51:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,331,007 | 18,674 | alexey-milovidov | @traceon Could you please also fix the `01502_bar_overflow` test in UBSan that can be related to your changes? | 2021-01-01T15:42:29 | 2021-01-01T15:42:29 | {} | MEMBER |
ClickHouse | ClickHouse | 753,436,818 | 18,662 | awesomeleo | > @awesomeleo but if you delete some rows `select count（*）` without FINAL will be incorrect.
Yes, I know. So we need to optimzie the FINAL performace. It's 1400 times slower than the query without FINAL. It's not acceptable. | 2021-01-02T06:02:00 | 2021-01-02T06:02:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,511,761 | 8,102 | den-crane | @NikolayPraydkin clickhouse-client такого не умеет. Можно загрузить sql командой `insert into select from file()` https://clickhouse.tech/docs/en/sql-reference/table-functions/file/ но в этом случае грузить будет clickhouse-server и файл должен быть доступен clickhouse-server . | 2021-01-02T18:20:45 | 2021-01-02T18:20:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,599,902 | 18,369 | azat | >We ended with 30GB of data in Distributed Table and it was not send for over 5 days
>
> snip
>
> There are 1,1mln bin files and 300 of them are broken and freezes Distributed Sends.
@wedi-dev Also it looks like you should use `distributed_directory_monitor_batch_inserts` to avoid too much files
>Error from t... | 2021-01-03T11:01:48 | 2021-01-03T11:02:05 | {} | MEMBER |
ClickHouse | ClickHouse | 753,625,251 | 18,691 | UnamedRus | Related https://github.com/ClickHouse/ClickHouse/issues/15997 | 2021-01-03T14:24:37 | 2021-01-03T14:24:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,954,379 | 18,661 | alexey-milovidov | It contains debug info. | 2020-12-31T13:05:39 | 2020-12-31T13:05:39 | {} | MEMBER |
ClickHouse | ClickHouse | 752,964,217 | 18,618 | den-crane | > will ch support ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROWS etc clauses?
If it's god's will it will happen | 2020-12-31T13:55:33 | 2020-12-31T13:55:33 | {
"hooray": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,982,447 | 18,362 | alesapin | Ded Moroz approved this PR :) | 2020-12-31T15:01:17 | 2020-12-31T15:01:17 | {} | MEMBER |
ClickHouse | ClickHouse | 752,987,766 | 7,430 | vitlibar | The work will be continued in https://github.com/ClickHouse/ClickHouse/pull/18673 | 2020-12-31T15:26:01 | 2020-12-31T15:26:01 | {} | MEMBER |
ClickHouse | ClickHouse | 753,513,638 | 16,503 | azat | Due to some dependencies in the code it looks like it is better to simply ignore extra thread usage for the s3 writes (and fix this separately)
@Akazz what do you think? | 2021-01-02T18:38:13 | 2021-01-02T18:38:13 | {} | MEMBER |
ClickHouse | ClickHouse | 752,971,591 | 17,646 | alexey-milovidov | I have strange feeling that it's overengineered. So we are going to introduce a settings that is difficult to use. And it is only for Buffer tables that we are going to make obsolete.
But the changes to memory tracker blocker are reasonable. Maybe don't introduce any settings to Buffer engine but just stick to "igno... | 2020-12-31T14:23:21 | 2020-12-31T14:23:37 | {} | MEMBER |
ClickHouse | ClickHouse | 752,989,024 | 17,455 | simPod | I edited table ZK node and it somehow got into sync. Maybe also I've reattached the table. Did not have an opportunity to find solid steps yet. | 2020-12-31T15:33:08 | 2020-12-31T15:33:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,012,796 | 17,455 | den-crane | @OmarBazaraa check at `get /clickhouse/tables/clarity/page_data_local_2/1-1/replicas/` that you have exactly/only 3 replicas
and all
```
get /clickhouse/tables/clarity/page_data_local_2/1-1/metadata
get /clickhouse/tables/clarity/page_data_local_2/1-1/replicas/1/metadata
get /clickhouse/tables/clarity/page_data_l... | 2020-12-31T17:36:27 | 2020-12-31T17:36:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,214,665 | 16,575 | garneja | is this supposed to work?
with 5 as q1, x as (select number as a from numbers(10)), y as (select number+100 as b, number as a from numbers(10)) select a from (select * from x where a > q1) x1 left join y using a
i'm getting this:
Received exception from server (version 20.13.1): ... | 2020-12-31T21:58:04 | 2020-12-31T21:58:04 | {} | NONE |
ClickHouse | ClickHouse | 753,416,235 | 16,575 | garneja | is with substitution not supposed to be transitive? this one doesn't work for me:
with 5 as q1, x as (select number+100 as b, number as a from numbers(10) where number > q1) select * from x
Code: 47. DB::Exception: Received from 127.0.0.1:9000. DB::Exception: Missing columns: 'q1' while processing query: 'SELECT... | 2021-01-02T01:51:51 | 2021-01-02T01:52:21 | {} | NONE |
ClickHouse | ClickHouse | 752,943,029 | 18,486 | alexey-milovidov | > Stress test (undefined) — Killed by signal (in clickhouse-server.log)
#18526 | 2020-12-31T12:19:59 | 2020-12-31T12:19:59 | {} | MEMBER |
ClickHouse | ClickHouse | 752,951,346 | 18,661 | l1t1 | thanks. and one more question
why https://builds.clickhouse.tech/master/aarch64/clickhouse
is so big size? it's about 2GB. | 2020-12-31T12:50:08 | 2020-12-31T12:50:08 | {} | NONE |
ClickHouse | ClickHouse | 752,962,250 | 18,665 | den-crane | also see https://github.com/ClickHouse/ClickHouse/issues/18618 | 2020-12-31T13:45:37 | 2020-12-31T13:45:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,980,678 | 17,838 | alexey-milovidov | Ok but it means we need to rewrite the test in another way... | 2020-12-31T14:54:53 | 2020-12-31T14:54:53 | {} | MEMBER |
ClickHouse | ClickHouse | 752,982,267 | 10,177 | OmarBazaraa | I am facing a similar issue when I try to drop an existing index:
```
DB::Exception: Metadata on replica is not up to date with common metadata in Zookeeper. Cannot alter.
``` | 2020-12-31T15:00:35 | 2020-12-31T15:00:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,496,259 | 18,684 | amosbird | Test failure doesn't look like related
```
2021-01-02 12:42:05 01133_max_result_rows: [ FAIL ] 0.80 sec. - result differs with reference:
2021-01-02 12:42:05 --- /usr/share/clickhouse-test/queries/0_stateless/01133_max_result_rows.reference 2021-01-02 10:31:04.00000... | 2021-01-02T16:42:19 | 2021-01-02T16:42:19 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 753,506,236 | 17,977 | alexey-milovidov | BTW, I'm not interested in gcc build with sanitizers at all (it has worse support than clang). | 2021-01-02T17:38:26 | 2021-01-02T17:38:26 | {} | MEMBER |
ClickHouse | ClickHouse | 753,511,079 | 8,102 | NikolayPraydkin | как в интерактивном режиме clickhouse-client загрузить tsv файл? | 2021-01-02T18:14:22 | 2021-01-02T18:14:22 | {} | NONE |
ClickHouse | ClickHouse | 753,554,979 | 3,575 | gevorgyg | 2 years have passed and still no PRs. Is it really an "easy task"? :) | 2021-01-03T01:50:43 | 2021-01-03T01:50:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,582,893 | 17,300 | astudnev | Please reopen this, as i getting the same errors after updated from 19.9 to 20.11
```
2021.01.03 08:01:15.601198 [ 1968 ] {b4472a5e-e463-4852-a0ed-efd90fae6393} <Error> DynamicQueryHandler: Code: 241, e.displayText() = DB::Exception: PeekableReadBuffer: Memory limit exceed, Stack trace (when copying this message, a... | 2021-01-03T08:08:30 | 2021-01-03T08:08:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,940,116 | 18,663 | PHaroZ | You can keep your decimal type and use [`intDivOrZero`](https://clickhouse.tech/docs/en/sql-reference/functions/arithmetic-functions/#intdivorzeroa-b) to divite your decimal with a 0 denominator
```sql
select intDivOrZero(1, toDecimal64(0, 2))
```
```
┌─intDivOrZero(1, toDecimal64(0, 2))─┐
│ ... | 2020-12-31T12:08:39 | 2020-12-31T12:09:56 | {} | NONE |
ClickHouse | ClickHouse | 752,946,654 | 18,668 | filimonov | BTW - while EOS semantics will guarantee you that no duplicates will happen on the Kafka side (i.e. even if you produce the same messages few times it will be consumed once), but ClickHouse as a kafka client can guarantee only at-least-once. And in some corner cases (connection lost etc) you can get duplicates.
We n... | 2020-12-31T12:31:08 | 2020-12-31T12:31:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,954,906 | 18,661 | l1t1 | can you also supply a non debug version? | 2020-12-31T13:08:49 | 2020-12-31T13:08:49 | {} | NONE |
ClickHouse | ClickHouse | 752,988,275 | 18,264 | ucasfl | > It failed in TSan: https://clickhouse-test-reports.s3.yandex.net/18264/714a4f8f1cafa723354126470e994bff61cce214/functional_stateless_tests_(thread)/stderr.log
Data race happened when insert into a cluster created with `cluster` function which have more than one shard on localhost.
And if local table engine is `... | 2020-12-31T15:28:58 | 2020-12-31T15:31:11 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 752,993,618 | 17,455 | OmarBazaraa | > I edited table ZK node and it somehow got into sync. Maybe also I've reattached the table. Did not have an opportunity to find solid steps yet.
The metadata in ZK seems correct to me! And unfortunately, re-attaching the table didn't help. | 2020-12-31T15:56:01 | 2020-12-31T15:56:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,330,750 | 18,655 | alexey-milovidov | @chenziliang Thank you! Now the chance it will be removed is low. | 2021-01-01T15:40:11 | 2021-01-01T15:40:11 | {} | MEMBER |
ClickHouse | ClickHouse | 753,501,374 | 17,977 | azat | >Functional stateful tests (ubsan) — No output log
https://clickhouse-test-reports.s3.yandex.net/17977/ffa718d0c3a9d6166778b13e0bd3623f3e5173be/functional_stateful_tests_(ubsan).html
Tons of
```
../src/Common/MemoryTracker.h:142:32: runtime error: load of null pointer of type 'uint64_t'
../src/Common/Current... | 2021-01-02T17:10:51 | 2021-01-02T17:10:51 | {} | MEMBER |
ClickHouse | ClickHouse | 753,512,134 | 18,264 | azat | >The difference is that cluster/remote uses sync inserts, and if you will try something like this you will see the race too:
And this means that the race is in the upstream too.
I guess it is totally fine to copy the query AST in the InterpreterInsertQuery, i.e.
```patch
$ git di
diff --git a/src/Interprete... | 2021-01-02T18:24:39 | 2021-01-02T18:24:39 | {} | MEMBER |
ClickHouse | ClickHouse | 752,955,734 | 18,661 | alexey-milovidov | If you run `strip clickhouse` you will get it automatically. | 2020-12-31T13:13:59 | 2020-12-31T13:13:59 | {} | MEMBER |
ClickHouse | ClickHouse | 752,966,307 | 18,662 | den-crane | @awesomeleo but if you delete some rows `select count（*）` without FINAL will be incorrect. | 2020-12-31T14:03:20 | 2020-12-31T14:03:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,977,064 | 18,264 | alexey-milovidov | I guess @tavplubix can help. | 2020-12-31T14:42:00 | 2020-12-31T14:42:00 | {} | MEMBER |
ClickHouse | ClickHouse | 753,030,498 | 18,446 | azat | >So it looks like the CAST is required, since only it can represent the correct type for types that does not have it's own DataType, i.e. DateTime and similar.
Force pushed to the previous version, with `castAssumeNotConst` (and no CI checks, since the SHA is the same) | 2020-12-31T19:16:23 | 2020-12-31T19:16:23 | {} | MEMBER |
ClickHouse | ClickHouse | 753,328,740 | 18,661 | alexey-milovidov | No. Debug info has absolutely zero impact on performance.
It is contained in a separate section of the binary that is not even loaded.
The machine code does not depend on the presence of the debug info.
> debug version is slower than release version in windows
You are mixed up a little. You mean "debug" build.
... | 2021-01-01T15:19:49 | 2021-01-01T15:20:14 | {} | MEMBER |
ClickHouse | ClickHouse | 753,392,344 | 18,683 | genzgd | These exceptions in the err.log might be related? Is it possible that such an error means that a pool entry isn't freed?
```
2021.01.01 22:05:02.638865 [ 326969 ] {} <Error> comcast_xcr.atsec_chi (5bc32aef-bbf8-4259-a2a7-d4617955cf6e): auto DB::StorageReplicatedMergeTree::queueTask()::(anonymous class)::operator()... | 2021-01-01T22:09:01 | 2021-01-01T22:09:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,457,585 | 17,875 | UnamedRus | -Distinct combinator works correctly.
```
SELECT
groupArrayDistinct(val) AS uniq,
toTypeName(uniq),
groupArray(val) AS arr,
toTypeName(arr)
FROM
(
SELECT CAST(number % 2, 'Enum(\'hello\' = 1, \'world\' = 0)') AS val
FROM numbers(2)
)
Query id: b2301bcc-9f35-488a-836e-025236593775... | 2021-01-02T10:34:43 | 2021-01-02T10:34:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,511,306 | 18,264 | azat | >And if local table engine is Log, the sql stucked in here(write block to LogBlockOutputStream) :
I see MergeTree engine in tests, but this does not looks important for the race, since it is not in Storage anyway.
>But if we create distributed table with Distributed engine, everything is good, what is the bigest ... | 2021-01-02T18:16:30 | 2021-01-02T18:16:30 | {} | MEMBER |
ClickHouse | ClickHouse | 752,962,156 | 18,665 | den-crane | >I would expect all groupArray(number) to have the same size (3).
No. It's wrong.
postgresql:
```sql
SELECT
gr,
str,
count(*) OVER (PARTITION BY gr ORDER BY str DESC) a,
sum(1) OVER (PARTITION BY gr ORDER BY str DESC) b,
row_number() OVER (PARTITION BY gr ORDER BY str DESC) c,
strin... | 2020-12-31T13:45:04 | 2020-12-31T13:45:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,003,279 | 17,455 | den-crane | what about replica 2 and 3 ?
get /clickhouse/tables/clarity/page_data_local_2/1-1/replicas/2/metadata
get /clickhouse/tables/clarity/page_data_local_2/1-1/replicas/3/metadata | 2020-12-31T16:45:00 | 2020-12-31T16:45:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,004,069 | 17,455 | den-crane | @OmarBazaraa OK . And what exact errors / log messages do you see when you execute drop index against this `shard 1 ` ? | 2020-12-31T16:49:53 | 2020-12-31T16:49:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,015,086 | 18,655 | chenziliang | We are using rocksdb table engine as a kv store for now since we don't wanna introduce another kv store dependency. If we get this removed, may i ask if there is any other alternative in CH (ReplacingMergeTree etc may not be as convenient as kv store) ? | 2020-12-31T17:48:24 | 2020-12-31T17:48:24 | {
"heart": 5
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,411,817 | 18,662 | bohutang | MaterializeMySQL underlying engine is ReplacingMergeTree, this case looks not make sense. Can you insert data to ReplacingMergeTree directly and do the test? | 2021-01-02T00:57:54 | 2021-01-02T00:57:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,602,791 | 18,623 | zhangguiyu | Thanks amosbird, here are additional info:
dict config:
guiyu@mi:~/mia/src/docker/clickhouse-server/etc/config.d$ more dmesh_dictionary.xml
```
<yandex>
<dictionary>
<name>dmesh</name>
<source>
<clickhouse>
<host>localhost</host>
<port>9000<... | 2021-01-03T11:28:11 | 2021-01-03T11:50:15 | {} | NONE |
ClickHouse | ClickHouse | 752,940,956 | 18,661 | alexey-milovidov | We already release binaries for Linux on ARM, see https://clickhouse.tech/docs/en/getting-started/install/#from-binaries-non-linux
There are no special binaries for Windows as ClickHouse works on WSL2.
We don't release binaries for Android although it can be built easily: https://clickhouse.tech/blog/en/2020/pixe... | 2020-12-31T12:12:01 | 2020-12-31T12:12:01 | {} | MEMBER |
ClickHouse | ClickHouse | 752,967,158 | 18,667 | den-crane | Can you provide:
* `create table`
* insert query
* excerpt from the /var/log/clickhouse-server/clickhouse-server.log about the time of issue | 2020-12-31T14:06:04 | 2020-12-31T14:06:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,006,114 | 17,455 | den-crane | @OmarBazaraa No. It's not those logs.
Execute without "on cluster" at any replica at shard1 `ALTER TABLE clarity.page_data_local_2 DROP INDEX page_duration_index`
And check logs at all 3 replica /var/log/clickhouse-server/clickhouse-server.err.log | 2020-12-31T17:00:41 | 2020-12-31T17:00:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,331,261 | 18,674 | alexey-milovidov | I checked it and found that it's appeared in this PR: #11617 | 2021-01-01T15:44:50 | 2021-01-01T15:45:02 | {} | MEMBER |
ClickHouse | ClickHouse | 753,333,076 | 9,621 | alexey-milovidov | I think that @CurtizJ will do it. | 2021-01-01T16:03:13 | 2021-01-01T16:03:13 | {} | MEMBER |
ClickHouse | ClickHouse | 753,338,791 | 18,579 | pingyu | > Some changes needed.
@alexey-milovidov
All comments addressed. PTAL, thanks~ | 2021-01-01T16:29:56 | 2021-01-01T16:29:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,421,165 | 6,060 | amosbird | https://github.com/ClickHouse/ClickHouse/pull/15451 | 2021-01-02T02:49:31 | 2021-01-02T02:49:31 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 753,512,448 | 17,977 | azat | >BTW, I'm not interested in gcc build with sanitizers at all (it has worse support than clang).
@alexey-milovidov yep, I assumed that.
But the purpose is not to replace clang santitizers with gcc, but to run sanitizers against code generated by gcc, to catch possible issues like #18053 (even though gcc is not use... | 2021-01-02T18:27:44 | 2021-01-02T18:27:44 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 753,641,528 | 17,623 | alexey-milovidov | @filimonov subpartitions: #18695 | 2021-01-03T16:26:02 | 2021-01-03T16:26:02 | {} | MEMBER |
ClickHouse | ClickHouse | 752,960,246 | 18,661 | l1t1 | > If you run `strip clickhouse` you will get it automatically.
I see. but supply such an file will save network traffic, since many people just want to use it, not test it. | 2020-12-31T13:34:45 | 2020-12-31T13:34:45 | {} | NONE |
ClickHouse | ClickHouse | 752,964,969 | 18,664 | den-crane | > My OS is Linux devcloud113 3.10.107-1-tlinux2_kvm_guest-0049 #1 SMP Tue Jul 30 23:46:29 CST 2019 x86_64 x86_64 x86_64 GNU/Linux
What Linux distro do you use? `cat /etc/*rel*` | 2020-12-31T13:58:48 | 2020-12-31T13:59:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 752,985,688 | 17,455 | OmarBazaraa | > Ok seems that I have really fixed it by editing the ZK. CH asynced it probably.
>
> Anyway, is there any way to force the sync / detect inconsistencies?
How did you fix it?
I have the same issue, but as far as I can tell, both the local metadata and the metadata in ZK are the same. | 2020-12-31T15:15:44 | 2020-12-31T15:15:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,027,263 | 17,455 | OmarBazaraa | @den-crane, yes, there are only 3 replicas and all the metadata are the same:
```
$ ls /clickhouse/tables/clarity/page_data_local_2/1-1/replicas
[1, 2, 3]
$ get /clickhouse/tables/clarity/page_data_local_2/1-1/metadata
metadata format version: 1
date column:
sampling expression: intHash32(UserId)
index ... | 2020-12-31T18:58:20 | 2020-12-31T18:58:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,239,662 | 16,575 | garneja | that works! thank you so much. | 2021-01-01T01:48:56 | 2021-01-01T01:48:56 | {} | NONE |
ClickHouse | ClickHouse | 753,370,927 | 18,681 | UnamedRus | I tried to search related issues, but did not succeed in that.
Let's close my issue then? | 2021-01-01T19:22:44 | 2021-01-01T19:22:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,478,818 | 18,385 | azat | >I see that performance package should contain it (and no errors from cp), however I don't see it in the performance.tgz archive.
#18686 | 2021-01-02T14:17:21 | 2021-01-02T14:17:21 | {} | MEMBER |
ClickHouse | ClickHouse | 753,007,761 | 17,455 | OmarBazaraa | **Replica 1**
```
DB::Exception: Metadata on replica is not up to date with common metadata in Zookeeper. Cannot alter.
```
```
2020.12.31 17:02:01.128426 [ 13584 ] {b6f4cacf-d164-49e9-ae5a-6170000e2d84} <Error> executeQuery: Code: 517, e.displayText() = DB::Exception: Metadata on replica is not up to date with ... | 2020-12-31T17:09:32 | 2020-12-31T17:09:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,009,956 | 11,617 | nikitamikhaylov | Performance test is broken in master. 01502_bar_overflow (ubsan) is broken is master too. 01133_max_result_rows (memory) is a bit flaky. AST Fuzzer failed without any stacktrace. | 2020-12-31T17:22:07 | 2020-12-31T17:22:07 | {} | MEMBER |
ClickHouse | ClickHouse | 753,609,615 | 18,685 | amosbird | Nice idea. Actually we can just let `except` accept a string and treat it as regex. | 2021-01-03T12:27:32 | 2021-01-03T12:27:32 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 752,945,823 | 18,668 | filimonov | I'll help you.... | 2020-12-31T12:28:01 | 2020-12-31T12:28:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,005,039 | 17,455 | OmarBazaraa | > @OmarBazaraa OK . And what exact errors / log messages do you see when you execute drop index against this `shard 1 ` ?
```
DB::Exception: Metadata on replica is not up to date with common metadata in Zookeeper. Cannot alter.
```
```
2020.12.31 16:03:46.936106 [ 7920 ] {e07130ab-a07d-4303-80e0-76c545ac5dce} ... | 2020-12-31T16:54:47 | 2020-12-31T16:54:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,464,995 | 15,927 | UnamedRus | untuple makes it a bit clearer.
```
arrayZip(untuple(sumMap(([partition_id], [rows])))) AS rows_per_partition
```
```
SELECT
database,
table,
formatReadableSize(compressed_bytes) AS readable,
sum(data_compressed_bytes) AS compressed_bytes,
sum(data_uncompressed_bytes) AS uncompressed... | 2021-01-02T11:57:14 | 2021-01-02T11:57:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 753,521,492 | 18,369 | azat | >We have big problem in one of our production nodes
Interesting. And you reproduced the problem with 20.12? Or you just upgrade the clickhouse and it starts processing existing files?
>Why Clickhouse does not move those files to the broken Directory but gets stucked?
Can you provide full stacktrace?
Error fro... | 2021-01-02T19:49:24 | 2021-01-02T19:49:24 | {} | MEMBER |
ClickHouse | ClickHouse | 753,251,162 | 18,661 | l1t1 | c program compiled in debug version is slower than release version in windows, is that also true in linux? | 2021-01-01T04:07:41 | 2021-01-01T04:07:41 | {} | NONE |
ClickHouse | ClickHouse | 753,621,461 | 18,685 | alexey-milovidov | Yes, it will be more consistent. | 2021-01-03T13:59:27 | 2021-01-03T13:59:27 | {} | MEMBER |
ClickHouse | ClickHouse | 753,637,776 | 18,692 | den-crane | 1. Did you upgrade you DB? Or created a new one in 20.8 and re-loaded data?
2. ```sql
select part_type, formatReadableSize(sum(data_compressed_bytes)) size, sum(rows) nrows,
count() nparts, formatReadableSize(sum(marks_bytes)) marks_bytes, sum(marks) marks
from system.parts
where active = 1 and ... | 2021-01-03T15:56:55 | 2021-01-03T15:57:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 758,032,532 | 18,897 | alesapin | These changes useful and don't related to the failure. Will investigate separately. | 2021-01-11T15:42:56 | 2021-01-11T15:42:56 | {} | MEMBER |
ClickHouse | ClickHouse | 758,115,165 | 18,951 | alexey-milovidov | `clickhouse_memcpy` should not be enabled for AArch64 builds. Looks like wrong build flags.
| 2021-01-11T17:48:27 | 2021-01-11T17:48:27 | {} | MEMBER |
ClickHouse | ClickHouse | 758,767,550 | 14,648 | Enmk | Noted, will do | 2021-01-12T16:17:42 | 2021-01-12T16:17:42 | {} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.