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 | 610,971,484 | 9,624 | alesapin | Do not test until the repository structure will be fixed. | 2020-04-08T13:49:34 | 2020-04-08T13:49:34 | {} | MEMBER |
ClickHouse | ClickHouse | 611,158,560 | 10,121 | alexey-milovidov | > Functional stateless tests (address) — Some queries hung
- reported. | 2020-04-08T19:50:05 | 2020-04-08T19:50:05 | {} | MEMBER |
ClickHouse | ClickHouse | 611,163,633 | 10,132 | alesapin | Probably you have a very silent logging level (`trace` level is recommended). Do you have access to zookeeper? If so, I'd like to see your table schema and data from zookeeper nodes `/path_to_table/metadata` and `/path_to_table/replicas/replica_name/metadata`. | 2020-04-08T20:00:13 | 2020-04-08T20:00:13 | {} | MEMBER |
ClickHouse | ClickHouse | 611,384,724 | 10,132 | alesapin | Ok, if you will see this error again just reopen this issue and provide server logs. | 2020-04-09T07:48:38 | 2020-04-09T07:48:38 | {} | MEMBER |
ClickHouse | ClickHouse | 611,409,937 | 10,011 | filimonov | duplicate of #8779
Fixed in versions >= 20.1.3 | 2020-04-09T08:44:42 | 2020-04-09T08:44:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,489,281 | 10,149 | nvartolomei | This might be a bit controversial, open for better ideas on how to solve the problem when initial_user has readonly=1 on profile but remote user has readonly=2. | 2020-04-09T11:58:34 | 2020-04-09T11:58:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,528,335 | 9,473 | abraithwaite | Also related, it's very frustrating that one `ctrl+c` press exits the client. Most other SQL shells and REPL shells cancel the current command and put you back at the prompt again. Would it be possible to make this configurable?
I agree however that multiple `ctrl+c`s in a row should exit the client mid-query thou... | 2020-04-09T13:29:10 | 2020-04-09T13:29:10 | {} | NONE |
ClickHouse | ClickHouse | 610,994,480 | 10,116 | den-crane | What CH version do you use?
Can you show 'show create table A' and 'show create table B' | 2020-04-08T14:32:28 | 2020-04-08T14:32:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,147,175 | 10,132 | cw9 | If there is information else you need me to collect, let me know while I can still reproduce it on that node.
Also would love to know if there are any workarounds for this bug. | 2020-04-08T19:27:34 | 2020-04-08T19:27:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,179,703 | 10,132 | cw9 | Hmm, so I saw there are 2 `drop column` commands on that table stuck for a long time unfinished in `system.mutation` table, I killed those two commands and now I can drop columns now | 2020-04-08T20:34:03 | 2020-04-08T20:34:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,462,549 | 10,136 | mnkonkova | The problem may have been that the gRPC_ZLIB_PROVIDER was not set | 2020-04-09T10:47:43 | 2020-04-09T10:47:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,684,157 | 10,149 | nvartolomei | We could allow on remote nodes to downgrade from readonly=2 to readonly=1, but this would break custom settings sent by proxy. If proxy could send profile settings and then downgrade the connection to readonly=1 somehow it could work but seems over complicated.
`SELECT 1 FROM distributed WHERE 1 IN (SELECT tableFunc... | 2020-04-09T18:29:18 | 2020-04-09T18:29:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 610,847,288 | 10,098 | alesapin | ehh, stupid bug :( Will be fixed in next 20.3. | 2020-04-08T09:15:03 | 2020-04-08T09:15:03 | {} | MEMBER |
ClickHouse | ClickHouse | 610,883,242 | 2,245 | msamusenka | @blinkov thanks. I believe you could close the issue. | 2020-04-08T10:37:19 | 2020-04-08T10:37:19 | {} | NONE |
ClickHouse | ClickHouse | 610,966,025 | 10,087 | den-crane | The default value `max_part_loading_threads = auto (==logical cores)` is set assuming that a system has a small number of large parts (<1000) and slow HDD disks. You can tune it for your particular system and set as you need ~ 8 / 64 / 96 / 333. | 2020-04-08T13:39:45 | 2020-04-08T13:43:28 | {
"hooray": 3
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,022,601 | 4,432 | qoega | reproduces in 20.3+ | 2020-04-08T15:22:53 | 2020-04-08T15:22:53 | {} | MEMBER |
ClickHouse | ClickHouse | 611,311,115 | 10,085 | alexey-milovidov | @achimbab Something is wrong, look at the tests: https://clickhouse-test-reports.s3.yandex.net/10085/53b5dade5ed488f8be0f0b8e7a624be3e5bb206c/functional_stateless_tests_(ubsan)/stderr.log | 2020-04-09T03:42:07 | 2020-04-09T03:42:07 | {} | MEMBER |
ClickHouse | ClickHouse | 611,314,604 | 10,139 | weiyongyuan | sorry ,i found that my distributed table and local table 's column found_time's type is not same. | 2020-04-09T03:58:31 | 2020-04-09T03:58:31 | {} | NONE |
ClickHouse | ClickHouse | 611,543,623 | 10,127 | filimonov | Mutations are executed asynchronously, out of current query context.
```sql
create table x Engine=MergeTree ORDER BY tuple() as SELECT * FROM numbers(100);
-- that temp table exists only in current session
create temporary table y as select * from numbers(10);
-- mutations are executed out of the session con... | 2020-04-09T13:59:50 | 2020-04-09T14:02:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,607,940 | 10,146 | qoega | Can reproduce on 20.4.1 testing build.
In master #10077 could already fixed this specific case | 2020-04-09T15:59:53 | 2020-04-09T15:59:53 | {} | MEMBER |
ClickHouse | ClickHouse | 611,637,245 | 10,131 | KochetovNicolai | 01086_odbc_roundtrip is a flappy test. | 2020-04-09T16:58:51 | 2020-04-09T16:58:51 | {} | MEMBER |
ClickHouse | ClickHouse | 611,026,833 | 9,881 | filimonov | > There is a problem that `1.2` literal is treated as float, but not as decimal. For correct results for example there should be conditional parsing of literals with `.`. Otherwise there would be error if number would be converted from float to decimal.
```
select CAST(0.11111111111111111111111111111111111111 AS De... | 2020-04-08T15:30:09 | 2020-04-08T15:31:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,178,533 | 10,132 | alesapin | Have you already performed successful alters after upgrade to 20.3.5? Also, do you have records in `system.replication_queue` for your table? | 2020-04-08T20:31:48 | 2020-04-08T20:34:04 | {} | MEMBER |
ClickHouse | ClickHouse | 611,310,307 | 10,085 | alexey-milovidov | @qoega NULL is not equals to NULL.
http://sqlfiddle.com/#!9/0a28f18/437
The current variant is more efficient and better for default. | 2020-04-09T03:38:40 | 2020-04-09T03:41:22 | {} | MEMBER |
ClickHouse | ClickHouse | 611,326,099 | 10,077 | zhang2014 | > Ok. I will also add a trivial functional test to check that the query
Could this possibly broken unbundled's test?
| 2020-04-09T04:49:05 | 2020-04-09T04:49:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,327,305 | 10,141 | weiyongyuan | > What version of clickhouse-server do you use?
> Do I understand correctly that you have 50 clickhouse-servers running on localhost?
The versions are 20.3.2.1 and 20.3.4.10.I have 3 distribute nodes,50 clickhouse-servers running on localhost.I've test 20.1.6.30 ,20.1.6.30 is ok。but,20.3.2.1 and 20.3.4.10 all has th... | 2020-04-09T04:54:00 | 2020-04-09T04:54:00 | {} | NONE |
ClickHouse | ClickHouse | 611,665,547 | 10,155 | alexey-milovidov | Both of them are using the same data structure (hash table).
The difference in performance is because you have much higher number of distinct values in the second case: 108 514 884 instead of 65 536.
Another possible reason is that UInt32 value is more difficult to read from disk and it is compressed worse due to... | 2020-04-09T17:55:17 | 2020-04-09T17:55:17 | {} | MEMBER |
ClickHouse | ClickHouse | 611,026,772 | 10,128 | alex-zaitsev | > > There is no way to force TTL manually without part merge. Need to add a new SYSTEM or ALTER command to allow manual control: OPTIMIZE TABLE TTL.
>
> `ALTER MATERIALIZE TTL [IN PARTITION]`? This query also recomputes and prunes parts with existing ttl metadata.
It only re-computes the TTLs but does not trigger... | 2020-04-08T15:30:01 | 2020-04-08T15:30:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,033,447 | 9,088 | filimonov | Reason:
```
SELECT
UserID,
UserID = 0,
if(UserID = 0, 'delete', 'leave')
FROM default.test2
FORMAT TSV
2 0 leave
0 1 delete
2 0 leave
2 0 leave
2 0 leave
0 1 delete
0 1 delete
\N \N delete
``` | 2020-04-08T15:41:44 | 2020-04-08T15:50:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,160,159 | 10,110 | alexey-milovidov | As I understant, now we support tautological cases: toString of String and LowCardinality(String). | 2020-04-08T19:53:08 | 2020-04-08T19:53:08 | {} | MEMBER |
ClickHouse | ClickHouse | 611,307,686 | 10,141 | alexey-milovidov | What version of clickhouse-server do you use?
Do I understand correctly that you have 50 clickhouse-servers running on localhost? | 2020-04-09T03:28:01 | 2020-04-09T03:28:01 | {} | MEMBER |
ClickHouse | ClickHouse | 611,318,853 | 205 | blinkov | https://clickhouse.tech/docs/en/introduction/adopters/ | 2020-04-09T04:17:58 | 2020-04-09T04:17:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,359,928 | 9,968 | danlark1 | Overall looks good to me, it would be also nice to test SSE2 vs SSE4.2 on AMD processors (I have a feeling that SSE2 might be faster there) but this is just an idea for future if somebody comes. @alexey-milovidov, your turn to review or merge | 2020-04-09T06:46:00 | 2020-04-09T06:46:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,190,247 | 10,132 | cw9 | btw I have quite a few shards returning that error, and on another shard I tried killing the stuck mutations and that fixed the problem again. | 2020-04-08T20:56:36 | 2020-04-08T20:56:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,323,690 | 10,011 | excitoon | @erste @filimonov This is a known problem. It was fixed in #8800 . | 2020-04-09T04:38:24 | 2020-04-09T04:38:24 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,352,123 | 10,113 | KochetovNicolai | @Sam--Shan thank you for feedback!
Most likely it was fixed after #10099 | 2020-04-09T06:22:32 | 2020-04-09T06:22:32 | {} | MEMBER |
ClickHouse | ClickHouse | 611,401,397 | 9,681 | l1t1 | please restore tha fonts of pdf
(you don't have to strictly follow this form)
**Describe the issue**
the new font isnt well readable. espsially UKaiCN, and the font for scripts and codes
the fonts of the pdf file in Apr 9.
You can find it in server logs. | 2020-04-08T08:39:10 | 2020-04-08T08:39:10 | {} | MEMBER |
ClickHouse | ClickHouse | 610,872,953 | 2,245 | blinkov | > @blinkov am I right saying that the suggested way to select records where attribute value equals NULL is using `isNull()` function?
Yes, you can also use more common `foo IS NULL` syntax. | 2020-04-08T10:12:14 | 2020-04-08T10:12:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 610,914,315 | 5,371 | stavrolia | Full support for ranges done here #9989 | 2020-04-08T11:54:14 | 2020-04-08T11:54:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 610,931,078 | 8,109 | akuzm | Broken in #7247.
What happens is that we check each function in `alter update` to see whether it is deterministic, and we don't work correctly with special function `lambda`. So if you have any lambda expression in your `alter update`, it breaks.
It reproduces for any table with Replicated engine, but doesn't rep... | 2020-04-08T12:31:45 | 2020-04-08T12:31:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 610,930,913 | 10,123 | alesapin | If less than 8 minutes passed you can detach `ALTER TABLE DETACH PARTITION` all partitions from your table and than found source parts for your mutation in `/detached` directory. For example: `all_1_10_1_132` -- mutated part, source part can have name `all_1_10_1_100` (the fifth number is less than in mutated part) or ... | 2020-04-08T12:31:25 | 2020-04-08T12:32:14 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 610,972,777 | 10,127 | coolpsix | After one error all mutations fails with the same error even if they don't use it and worked before.
Row 5:
──────
database: default
table: tbl
mutation_id: mutation_10.txt
command: UPDATE a= 3 WHERE b= 5
create_time: 2020-04... | 2020-04-08T13:52:03 | 2020-04-08T13:55:02 | {} | NONE |
ClickHouse | ClickHouse | 611,143,816 | 10,132 | alesapin | Does error reproduce if try to execute the same query again on the same host? | 2020-04-08T19:20:41 | 2020-04-08T19:20:59 | {} | MEMBER |
ClickHouse | ClickHouse | 611,148,941 | 10,132 | alesapin | Very strange, this error has an incorrect message and type, but it should mean, that we cannot connect to zookeeper. Do tables on this node work properly (you can insert data)? | 2020-04-08T19:30:47 | 2020-04-08T19:30:47 | {} | MEMBER |
ClickHouse | ClickHouse | 611,161,090 | 10,110 | alexey-milovidov | I've added changelog entry (otherwise the "Description check" is red). | 2020-04-08T19:54:57 | 2020-04-08T19:54:57 | {} | MEMBER |
ClickHouse | ClickHouse | 611,210,952 | 10,049 | azat | >01086_odbc_roundtrip
failure looks unrelated | 2020-04-08T21:45:11 | 2020-04-08T21:45:11 | {} | MEMBER |
ClickHouse | ClickHouse | 611,211,153 | 10,049 | azat | >Yes, it's Ok. Let's update the tests.
Done | 2020-04-08T21:45:40 | 2020-04-08T21:45:40 | {} | MEMBER |
ClickHouse | ClickHouse | 611,376,005 | 5,346 | alekseipalshin | Reproduced on CH version: 19.17.6
| 2020-04-09T07:27:58 | 2020-04-09T07:27:58 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 611,537,101 | 10,128 | filimonov | > It's not correct. See test 01070_mutations_with_dependencies.
:+1: So that is done :) #8775
| 2020-04-09T13:46:54 | 2020-04-09T13:46:54 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,542,107 | 9,964 | BayoNet | I'll merge the source branch into docs/CLICKHOUSEDOCS-558-RBAC-introduction. | 2020-04-09T13:57:01 | 2020-04-09T13:57:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,673,295 | 10,149 | nvartolomei | @alexey-milovidov I'm out of ideas here. | 2020-04-09T18:08:23 | 2020-04-09T18:08:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 610,944,733 | 6,336 | leoluk | Any update on this? Is it true that CH Kafka ingestion is single-threaded on a single shard regardless of partition count? | 2020-04-08T12:59:52 | 2020-04-08T12:59:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,151,076 | 10,132 | cw9 | Yes, I can still write to and read from this table | 2020-04-08T19:34:55 | 2020-04-08T19:34:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,177,826 | 10,132 | cw9 | 5 for one replica and 3 for the other replica | 2020-04-08T20:30:28 | 2020-04-08T20:30:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,335,778 | 10,113 | Sam--Shan | @KochetovNicolai I tried the lastest testing version 20.4.1.2909. It works, that exception is not displayed again. Looking forward to its release soon. Thanks a lot . | 2020-04-09T05:27:15 | 2020-04-09T05:27:15 | {} | NONE |
ClickHouse | ClickHouse | 611,645,119 | 10,149 | nvartolomei | On proxy cluster we have 2 types of users, ones that are allowed to customize settings and ones that aren't and also have tighter limits in order to support both of these the user on remote clusters has to allow both of them hence readonly=2.
@alexey-milovidov wdyt? | 2020-04-09T17:14:12 | 2020-04-09T17:14:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 610,932,093 | 10,117 | alesapin | Will try to check tests in the release branch. | 2020-04-08T12:33:52 | 2020-04-08T12:33:52 | {} | MEMBER |
ClickHouse | ClickHouse | 611,237,486 | 10,138 | alexey-milovidov | No test required. | 2020-04-08T23:01:45 | 2020-04-08T23:01:45 | {} | MEMBER |
ClickHouse | ClickHouse | 611,435,620 | 10,145 | alesapin | Yes, just update to the latest 20.1 or 20.3. | 2020-04-09T09:42:53 | 2020-04-09T09:42:53 | {} | MEMBER |
ClickHouse | ClickHouse | 610,910,036 | 10,116 | xiashuijun | 你是不是tableA的数据没有清除干净，又往里面写。因为clickhouse要过一会才能看到数据 | 2020-04-08T11:43:11 | 2020-04-08T11:43:11 | {} | NONE |
ClickHouse | ClickHouse | 611,662,719 | 10,149 | alexey-milovidov | @qoega
> If a user has readonly=1 on proxy cluster there is no way how he can change setting that would be send further. And thus for readonly=1 scenario it looks like all settings should be sent only from different user from proxy server. Currently readonly user specific config is pushed further and it looks as a ... | 2020-04-09T17:49:53 | 2020-04-09T17:49:53 | {} | MEMBER |
ClickHouse | ClickHouse | 610,874,818 | 9,462 | qoega | Without `b` column in your example it saves the row? | 2020-04-08T10:16:35 | 2020-04-08T10:16:35 | {} | MEMBER |
ClickHouse | ClickHouse | 610,967,870 | 7,683 | alesapin | testing disabled until we fix directories structure. | 2020-04-08T13:42:49 | 2020-04-08T13:42:49 | {} | MEMBER |
ClickHouse | ClickHouse | 611,172,198 | 10,132 | cw9 | Yes ZK info:
```
get /clickhouse/databases/default/tables/foo/shards/147/replicas/1/metadata
metadata format version: 1
date column:
sampling expression:
index granularity: 8192
mode: 0
sign column:
primary key: _namespace, -_timestampMillis
data format version: 1
partition key: intDiv(_timestampMillis, 3600... | 2020-04-08T20:18:57 | 2020-04-08T20:18:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,186,042 | 10,132 | cw9 | The value is 0 in `parts_to_do` and 0 in `is_done`, I'll set logging level to `trace` and see if I can find more information | 2020-04-08T20:47:12 | 2020-04-08T20:47:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,413,661 | 10,145 | Inasayang | 20.1.6.30 (official build) | 2020-04-09T08:53:00 | 2020-04-09T08:53:00 | {} | NONE |
ClickHouse | ClickHouse | 611,496,785 | 10,149 | qoega | Actually the most controversial is the idea, that user has a right to use more privileged user than he has on a proxy, but it can be OK.
IMHO. If a user has readonly=1 on proxy cluster there is no way how he can change setting that would be send further. And thus for readonly=1 scenario it looks like all settings shou... | 2020-04-09T12:18:13 | 2020-04-09T12:18:13 | {} | MEMBER |
ClickHouse | ClickHouse | 611,685,303 | 1,604 | Avogar | @tacyuuhon. Hello, I have a question about Thrift. How do you use it in Kafka and what did you mean by "support Thrift"? Msgpack is a specific data storage format, but Thrift is not. (I could be wrong, correct me please). | 2020-04-09T18:31:31 | 2020-04-09T18:31:31 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 610,838,343 | 2,245 | msamusenka | @blinkov am I right saying that the suggested way to select records where attribute value equals NULL is using `isNull()` function? | 2020-04-08T08:55:28 | 2020-04-08T08:55:28 | {} | NONE |
ClickHouse | ClickHouse | 611,018,803 | 10,128 | CurtizJ | > There is no way to force TTL manually without part merge. Need to add a new SYSTEM or ALTER command to allow manual control: OPTIMIZE TABLE TTL.
```ALTER MATERIALIZE TTL [IN PARTITION]```? This query also recomputes and prunes parts with existing ttl metadata.
> When mutation is performed on the columns that ar... | 2020-04-08T15:16:20 | 2020-04-08T15:18:09 | {} | MEMBER |
ClickHouse | ClickHouse | 611,152,139 | 10,132 | alesapin | Ok. Strange. Can you provide logs (`/var/log/clickhouse-server/clickhouse-server.log`) from this node? Also is it possible to execute the same query on other replicas of this node? | 2020-04-08T19:37:03 | 2020-04-08T19:37:03 | {} | MEMBER |
ClickHouse | ClickHouse | 611,176,489 | 10,132 | alesapin | Do you have `path_to_table/replica/replica_name/metadata_version`? | 2020-04-08T20:27:44 | 2020-04-08T20:27:44 | {} | MEMBER |
ClickHouse | ClickHouse | 611,562,725 | 10,041 | ssky1 | hi filimonov,
just for my understanding, each insert row is one insert or each batch insert (batchSize = 1000 rows) is one insert !
I explain myself
if I have 500GB dedecated to clickhouse data ( 3 2768 000 inodes)
32 768 000 indoes / 480 (files per insert ) = 68k inserts
as I am using batchInsert wi... | 2020-04-09T14:36:19 | 2020-04-09T14:36:19 | {} | NONE |
ClickHouse | ClickHouse | 611,662,051 | 10,149 | alexey-milovidov | It looks like the query:
`SELECT 1 FROM distributed WHERE 1 IN (SELECT tableFunction(...))`
will lead to privelege escalation (readonly user will be able to execute table function on remote server).
| 2020-04-09T17:48:27 | 2020-04-09T17:50:11 | {} | MEMBER |
ClickHouse | ClickHouse | 611,738,182 | 10,049 | alexey-milovidov | I have changed changelog category to "Backward Incompatible Change" because it is Backward Incompatible Change. | 2020-04-09T20:30:51 | 2020-04-09T20:30:51 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 611,741,867 | 10,149 | alexey-milovidov | > Ie how this works when a user on the proxy has a row filter or something?
Row filters do not work on Distributed tables directly.
Instead they have to be enforced for local tables on remote nodes.
And this was implemented in a very bad way: https://github.com/ClickHouse/ClickHouse/pull/8926. The filters are sele... | 2020-04-09T20:39:37 | 2020-04-09T20:39:54 | {} | MEMBER |
ClickHouse | ClickHouse | 610,834,056 | 10,113 | Sam--Shan | [exception.txt](https://github.com/ClickHouse/ClickHouse/files/4449306/exception.txt)
Thanks for your reply,Mr. Kochetov.And this is my exception. | 2020-04-08T08:46:27 | 2020-04-08T08:46:27 | {} | NONE |
ClickHouse | ClickHouse | 611,213,188 | 10,136 | alexey-milovidov | This is not meant to be merged, please switch to "draft" pull request. | 2020-04-08T21:51:02 | 2020-04-08T21:51:02 | {} | MEMBER |
ClickHouse | ClickHouse | 611,465,548 | 10,116 | eiffel0311 | > 你是不是tableA的数据没有清除干净，又往里面写。因为clickhouse要过一会才能看到数据
i create new tableA and select * from sysytem.mutations . i am sure the query done | 2020-04-09T10:55:50 | 2020-04-09T10:55:50 | {} | NONE |
ClickHouse | ClickHouse | 611,625,398 | 5,587 | filimonov | The reason is the data in the file was corrupted (most probably some hardware issue).
If the problem is on the single replica - detach table, drop folder, attach table. It should download the part from other replica.
| 2020-04-09T16:34:46 | 2020-04-09T16:34:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,722,159 | 10,158 | filimonov | Rather it's bug #9940
It is not very clear how does it happen. Are you able to create some isolated sample? | 2020-04-09T19:53:22 | 2020-04-09T21:23:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 610,956,026 | 10,123 | mrAndersen | @alesapin more than 8 minutes passed :( | 2020-04-08T13:21:38 | 2020-04-08T13:21:38 | {} | NONE |
ClickHouse | ClickHouse | 611,095,411 | 10,120 | alesapin | Build OK, first test run finished. Merge. | 2020-04-08T17:41:26 | 2020-04-08T17:41:26 | {} | MEMBER |
ClickHouse | ClickHouse | 611,145,679 | 10,132 | cw9 | Yes it does reproduce, there is a column I want to drop from that table and I can neither drop it or add other columns to the table | 2020-04-08T19:24:46 | 2020-04-08T19:25:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,702,922 | 10,146 | alexey-milovidov | #10056 | 2020-04-09T19:08:54 | 2020-04-09T19:08:54 | {} | MEMBER |
ClickHouse | ClickHouse | 611,752,511 | 9,889 | alexey-milovidov | There is no way to disable Msgpack from build, but we will fix it later. | 2020-04-09T21:06:12 | 2020-04-09T21:06:12 | {} | MEMBER |
ClickHouse | ClickHouse | 610,845,892 | 9,337 | admsev | @filimonov It would be great to have it indeed, thanks a lot! | 2020-04-08T09:12:05 | 2020-04-08T09:12:05 | {} | NONE |
ClickHouse | ClickHouse | 597,167,201 | 4,132 | dmkoltsov | btw: in Clickhouse 19.8 ppc64le build with the workaround above - I've already met the bug (which, may be, fixed in later versions of Clickhouse) - it's the bug with clickhouse-client program.
When using first query from tutorial:
```
host :) SELECT
StartURL AS URL,
AVG(Duration) AS AvgDuration
FROM t... | 2020-03-10T16:00:46 | 2020-04-08T18:35:05 | {} | NONE |
ClickHouse | ClickHouse | 611,161,107 | 10,132 | cw9 | The log is also showing the same stack, nothing much else.
I can confirm the data written to this replica is replicated to the other replica. The `alter table add column` also fails on the other replica. | 2020-04-08T19:54:59 | 2020-04-08T19:55:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,163,099 | 10,077 | alexey-milovidov | Ok. I will also add a trivial functional test to check that the query
```
CREATE DATABASE conv_main ENGINE = MySQL('host:3306', conv_main, 'metrika', 'password')
```
does not segfault. | 2020-04-08T19:59:05 | 2020-04-08T19:59:05 | {} | MEMBER |
ClickHouse | ClickHouse | 611,180,563 | 10,132 | alesapin | Ok, it would be awesome if you will provide verbose server logs. Otherwise, I don't know how to investigate this issue :( | 2020-04-08T20:35:43 | 2020-04-08T20:35:43 | {} | MEMBER |
ClickHouse | ClickHouse | 611,402,771 | 9,968 | Enmk | > Overall looks good to me, it would be also nice to test SSE2 vs SSE4.2 on AMD processors (I have a feeling that SSE2 might be faster there) but this is just an idea for future if somebody comes. @alexey-milovidov, your turn to review or merge
Sure, I'd be happy to see more test results. I hope that it is relativel... | 2020-04-09T08:29:00 | 2020-04-09T08:29:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,663,176 | 10,149 | alexey-milovidov | @blinkov This PR is not non-significant. | 2020-04-09T17:50:51 | 2020-04-09T17:50:51 | {} | MEMBER |
ClickHouse | ClickHouse | 611,163,766 | 10,132 | alesapin | Also `SELECT * FROM system.mutations` is useful. | 2020-04-08T20:00:28 | 2020-04-08T20:00:28 | {} | MEMBER |
ClickHouse | ClickHouse | 611,180,986 | 10,132 | alesapin | > Hmm, so I saw there are 2 `drop column` commands on that table stuck for a long time unfinished in `system.mutation` table, I killed those two commands and now I can drop columns now
How did they look like? How many records they had in `parts_to_do` column? | 2020-04-08T20:36:36 | 2020-04-08T20:36:36 | {} | MEMBER |
ClickHouse | ClickHouse | 611,354,294 | 9,968 | Enmk | Benchmark code and results (please note that I had to change code a bit from excluding pieces of code with macros to excluding code with `if constexpr` in order to simplify measurements):
https://gist.github.com/Enmk/2c72399733d52f1e20ea3df2051db2d1
Conclusion:
* SSE2 && SSE4.2 versions are slower (~25%) if token... | 2020-04-09T06:29:30 | 2020-04-09T06:45:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 611,427,542 | 10,145 | Inasayang | I think i find the solution.
[#9860](https://github.com/ClickHouse/ClickHouse/pull/9860) | 2020-04-09T09:23:51 | 2020-04-09T09:23:51 | {} | NONE |
ClickHouse | ClickHouse | 611,742,483 | 10,149 | alexey-milovidov | @blinkov This is not "non-significant".
| 2020-04-09T20:41:19 | 2020-04-09T20:41:24 | {} | MEMBER |
ClickHouse | ClickHouse | 601,573,062 | 8,890 | ssserj | 20.1.6.30 release fixed that issue.
@josepowera @khamamet @bezzil85 @SergeiSadov please check on your clusters. | 2020-03-20T07:51:11 | 2020-03-20T07:51:11 | {} | NONE |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.