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 | 572,049,933 | 7,646 | ghuname | @filimonov Thanks for an answer. Can you please answer to https://github.com/ClickHouse/ClickHouse/issues/7646#issuecomment-571939568 as well? | 2020-01-08T13:27:13 | 2020-01-08T13:27:13 | {} | NONE |
ClickHouse | ClickHouse | 572,125,549 | 8,568 | filimonov | When you try to upgrade clickhouse-common-static, it tries to uninstall the previous version first. But it is required by another package (clickhouse-server) and can't be uninstalled because of that.
The solution is to update both packages in one transaction, for example
```rpm -Fvh *.rpm```
| 2020-01-08T15:38:38 | 2020-01-08T15:38:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,624,408 | 8,583 | den-crane | @kuncao you mean it's a secret or what? Try to rename columns / tables and obfuscate your data sample.
Email has no sense I am not a CH developer I am a member of CH community the same as you are. | 2020-01-09T15:55:49 | 2020-01-09T15:56:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,644,349 | 8,578 | traceon | I'll try that.
| 2020-01-09T16:38:41 | 2020-01-09T16:38:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,726,276 | 8,588 | alesapin | Broken test fixed. | 2020-01-09T19:51:00 | 2020-01-09T19:51:00 | {} | MEMBER |
ClickHouse | ClickHouse | 572,894,231 | 1,761 | SulaymonHursanov | drop dictionary dict_name;
-- works on version 19.17
| 2020-01-10T06:32:29 | 2020-01-10T06:32:29 | {
"+1": 1
} | NONE |
ClickHouse | ClickHouse | 571,881,359 | 7,507 | yuzhichang | I added some debug code at dbms/src/Common/isLocalAddress.cpp to dump all interfaces' address.
The following is log at when the dddl hang problem occur,
```
Jan 8 11:42:35 shoyclp clickhouse-server: interface address of lo: 127.0.0.1
Jan 8 11:42:35 shoyclp clickhouse-server: interface address of lo: ::1%lo
Jan ... | 2020-01-08T03:59:30 | 2020-01-08T03:59:30 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,264,997 | 8,549 | excitoon | Yep, I shall make it soon. | 2020-01-08T21:28:31 | 2020-01-08T21:28:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,593,799 | 8,578 | alexey-milovidov | > No symbol table info available.
If you run stripped binary, you have to install `clickhouse-common-static-dbg` package. | 2020-01-09T14:47:50 | 2020-01-09T14:51:45 | {} | MEMBER |
ClickHouse | ClickHouse | 572,598,008 | 8,578 | alexey-milovidov | Please check what versions of ClickHouse are affected. | 2020-01-09T14:56:51 | 2020-01-09T14:56:51 | {} | MEMBER |
ClickHouse | ClickHouse | 572,598,714 | 8,578 | alexey-milovidov | Maybe your VM uses files from the host system. You suspend the VM, then the `clickhouse` binary is changed, and when you continue VM, the code in this file becomes completely unrelated. | 2020-01-09T14:58:28 | 2020-01-09T14:58:28 | {} | MEMBER |
ClickHouse | ClickHouse | 572,942,494 | 6,462 | ghuname | I can confirm that partitions are not dropped in:
ClickHouse client version 19.16.3.6.
Connected to ClickHouse server version 19.16.3 revision 54427.
When you look at system.parts table, you can see a lot of old partitions (before 2020-01-09) that doesn't have any rows, but they remain in system.parts table as wel... | 2020-01-10T09:12:56 | 2020-01-10T09:12:56 | {} | NONE |
ClickHouse | ClickHouse | 573,003,946 | 8,329 | alesapin | From a user's point of view, it would be much easier to use `[db.]dictionary` name without quotation. But as far as I know, we have a very complicated aliases resolution rules and probably it will become even more complex... | 2020-01-10T11:44:19 | 2020-01-10T11:44:19 | {} | MEMBER |
ClickHouse | ClickHouse | 573,014,893 | 8,591 | alesapin | Works for `gentoo` and `archlinux`!
Already unpacked `tgz` example:
```
$ docker run --network=host --volume=`pwd`/tgzs:/package_dir -it gentoo/stage3-amd64 bash
click / # cd /package_dir/clickhouse-common-static-20.1.1.2088/install
click /package_dir/clickhouse-common-static-20.1.1.2088/install # ./doinst.sh
cl... | 2020-01-10T12:22:24 | 2020-01-10T12:22:59 | {} | MEMBER |
ClickHouse | ClickHouse | 573,022,173 | 5,690 | 4ertus2 | > One approach would to introduce an AST-only type of `LiteralDecimal(precision,scale)` and convert on demand to Decimal or Float64 when an operator or assignment is applied
Do you mean we should implement comparison with literals but not implement comparison with Float and String columns? It's something strange. Co... | 2020-01-10T12:47:07 | 2020-01-10T12:47:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,120,999 | 8,571 | filimonov | test? | 2020-01-08T15:28:37 | 2020-01-08T15:28:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,126,210 | 6,368 | HungUnicorn | @StephenSorriaux hi, I use the same Settings but saw the following error.
`DB::Exception: Cannot parse input: expected \t at end of stream.: (at row 1)`
have you seen it? | 2020-01-08T15:40:09 | 2020-01-08T15:40:09 | {} | NONE |
ClickHouse | ClickHouse | 572,405,027 | 8,568 | huang--yijian | > When you try to upgrade clickhouse-common-static, it tries to uninstall the previous version first. But it is required by another package (clickhouse-server) and can't be uninstalled because of that.
>
> The solution is to update both packages in one transaction, for example
> `rpm -Fvh *.rpm`
We put 19.6.2 RP... | 2020-01-09T06:07:28 | 2020-01-09T06:07:28 | {} | NONE |
ClickHouse | ClickHouse | 572,597,136 | 8,578 | alexey-milovidov | Do you compile debug or release version?
Debug version use more (anonymous) memory mappings and mprotect some of them.
VMWare may have bugs associated with this behaviour. | 2020-01-09T14:54:53 | 2020-01-09T14:55:33 | {} | MEMBER |
ClickHouse | ClickHouse | 573,011,790 | 5,690 | barrkel | One approach would to introduce an AST-only type of `LiteralDecimal(precision,scale)` and convert on demand to Decimal or Float64 when an operator or assignment is applied, and the type of the other side known. If both sides are LiteralDecimal, they can both be converted to Float64 - relative comparison and ordering sh... | 2020-01-10T12:12:07 | 2020-01-10T12:12:07 | {} | NONE |
ClickHouse | ClickHouse | 573,020,444 | 5,690 | 4ertus2 | > or at the very least a feature deficiency compared with standard SQL
Yes. ClickHouse has strict data typing and it's incompatible with SQL in some cases. It's not a bug. It's a choise convert data types in runtime as seldom as possible. ClickHouse use fast but not safe data types. That means if you call some inter... | 2020-01-10T12:41:42 | 2020-01-10T12:49:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,130,788 | 8,505 | filimonov | It is not ignored, but it kills queries 'gracefully'.
I.e. it just set a flag "it's enough" and that flag is checked at some moments of query execution by working threads. | 2020-01-08T15:50:06 | 2020-01-08T15:50:06 | {
"-1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,439,658 | 8,511 | YiuRULE | Related to https://github.com/ClickHouse/ClickHouse/issues/8469 btw | 2020-01-09T08:01:48 | 2020-01-09T08:01:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,633,578 | 8,579 | traceon | Even a special flag that needs to be turned on manually to enable filtering/transformation would be useful. | 2020-01-09T16:15:39 | 2020-01-09T16:15:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,635,062 | 8,579 | alexey-milovidov | It is possible to configure with `query_masking_rules`.
```
<rule>
<name>ascii_esc</name>
<regexp>\x1b</regexp>
<replace>\\e</replace>
</rule>
```
and your terminal will be safe.
About bad UTF-8. Terminal will replace them to � automatically. | 2020-01-09T16:19:01 | 2020-01-09T20:11:26 | {} | MEMBER |
ClickHouse | ClickHouse | 572,933,600 | 7,799 | zydm521 | I also have this problem，But I did。
Add one level of sub query to solve the problem
select *
from A
ANY LEFT JOIN B
ON A.a=B.a
WHERE A.a IN (XXX)
AND A.b IN (XXX);
ok，change
select *
from （select * from A ）AS A
ANY LEFT JOIN B
ON A.a=B.a
WHERE A.a IN (XXX)
AND A.b IN (XXX);
Then right！
... | 2020-01-10T08:47:04 | 2020-01-10T08:47:04 | {} | NONE |
ClickHouse | ClickHouse | 573,044,011 | 8,612 | alesapin | Need to test own instruction. I will wait for CI build. | 2020-01-10T13:53:31 | 2020-01-10T13:53:31 | {} | MEMBER |
ClickHouse | ClickHouse | 572,263,855 | 2,269 | den-crane | > I'm facing the same problem and overcommit settings seem fine (it is zero).
>
> `max_memory_usage` is 80G and `max_bytes_before_external_group_by` is 40G.
>
> Any idea?
@ramazanpolat
What EXACT error message did you get. What CH version do you use?
How many memory your sever has? | 2020-01-08T21:25:28 | 2020-01-08T21:26:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,552,023 | 8,591 | alesapin | Unfortunately, now it works out of the box only for centos and ubuntu :(
```centos:
[root@click package_dir]# tar -xzvf clickhouse-common-static-20.1.1.2074.tgz
[root@click package_dir]# tar -xzvf clickhouse-server-20.1.1.2074.tgz
[root@click package_dir]# tar -xzvf clickhouse-client-20.1.1.2074.tgz
[root@click pa... | 2020-01-09T13:04:10 | 2020-01-09T13:04:10 | {} | MEMBER |
ClickHouse | ClickHouse | 572,590,888 | 8,579 | alexey-milovidov | I think that it's better to write bytes to log without any modification (including zero bytes, ASCII esc, illegal UTF-8 sequences). This is useful for post-mortem analysis. Sometimes you will need the exact contents of these bytes. | 2020-01-09T14:41:22 | 2020-01-09T14:43:33 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 572,595,582 | 8,578 | alexey-milovidov | The error `Non-existant physical address` is possible if you have changed the files that were associated with memory mapping. We usually don't mmap files, so the only suspected file is the `clickhouse` binary (that is mmapped when executable is run by the OS).
https://stackoverflow.com/questions/58118497/a-possible-... | 2020-01-09T14:51:35 | 2020-01-09T14:51:35 | {} | MEMBER |
ClickHouse | ClickHouse | 572,632,768 | 8,578 | traceon | ClickHouse is built manually and the executable from the build folder is executed. Build type is `RelWithDebInfo`.
The build (and source) dirs are on the mounted fs (a mounted "shared folders" of VMware), binaries are not changed outside the vm though.
Happened with all CH buld that I had, but can tell for sure o... | 2020-01-09T16:13:51 | 2020-01-09T16:13:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,757,190 | 8,466 | pachico | No, not in chproxy but, depending on what you mean by restricting or filtering data, you can do it in CH.
So yes, it seems like a combination of the two might be what you need. | 2020-01-09T21:09:23 | 2020-01-09T21:09:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 573,004,246 | 8,608 | 865357384 | this question2:
select
0 as PGID,
sg.PGID AS PGID2
from tmp_data.sy_game sg
this result:
ClickHouse exception, code: 10, host: X.X.X.126, port: 8123; Code: 10, e.displayText() = DB::Exception: Not found column PGID in block. There are only columns: ID, 0 (version 19.15.4.10)
this question3:... | 2020-01-10T11:45:23 | 2020-01-10T11:57:17 | {} | NONE |
ClickHouse | ClickHouse | 573,010,207 | 5,690 | barrkel | I think this is a bug, or at the very least a feature deficiency compared with standard SQL. Such comparisons work in almost every other database product, because they don't perform early conversion of strings of decimal digits to floating point.
A literal `1000.0000` can be safely interpreted as a Decimal(18,4); it... | 2020-01-10T12:06:11 | 2020-01-10T12:06:11 | {} | NONE |
ClickHouse | ClickHouse | 571,726,844 | 8,564 | luqasz | Thx for info. | 2020-01-07T19:08:27 | 2020-01-07T19:08:27 | {} | NONE |
ClickHouse | ClickHouse | 572,955,813 | 8,466 | pachico | This is exactly what we do.
We use chproxy for:
- roundrobin between all nodes (it performs healthcheck to point only to healthy ones, which was our first and original reason to adopt it)
- avoid exposing internal credentials
- apply rate limit | 2020-01-10T09:49:17 | 2020-01-10T09:49:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 571,935,981 | 7,646 | ghuname | Can you make messages like this one more descriptive?
`2019.11.29 05:05:47.154529 [ 37 ] {} <Error> void DB::StorageKafka::threadFunc(): std::exception. Code: 1001, type: cppkafka::HandleException, e.what() = Local: Timed out, version = 19.16.3.6`
What this message tells me? What should I do if I see it?
Anoth... | 2020-01-08T08:06:32 | 2020-01-08T08:06:32 | {} | NONE |
ClickHouse | ClickHouse | 571,939,568 | 7,646 | ivanprostran | "max.poll.interval.ms" is most common error when dealing with kafka streams using processor api.
In other words long running message "handlers/punctuators" could trigger this. I am trying to say this is perfectly OK and logical for re-balancing purposes.
What I don't understand is kafka threading model in CH.
- ... | 2020-01-08T08:18:55 | 2020-01-08T08:18:55 | {} | NONE |
ClickHouse | ClickHouse | 572,132,269 | 8,504 | filimonov | Stupid question: are clocks in sync between replicas? Do you have ntpd running? | 2020-01-08T15:53:12 | 2020-01-08T15:53:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,353,325 | 5,601 | oandrew | Hi everyone. I ended up creating a PR to support Avro + AvroConfluent.
https://github.com/ClickHouse/ClickHouse/pull/8571
Anyone interested - give it a try. I'd appreciate any feedback. | 2020-01-09T02:30:06 | 2020-01-09T02:37:05 | {
"hooray": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,389,739 | 8,583 | den-crane | >following the page 12 and page 14
Are you kidding? Nobody never will do it. | 2020-01-09T05:08:19 | 2020-01-09T05:09:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,507,811 | 8,568 | filimonov | You had only rpms from 19.17 in that folder before :) Now you are trying to install both 19.16 and 19.17 at the same time. But you need to install v19.17, right? :)
Please list all rpms which need to be installed explicitly, or use appropriate wildcard.
```
rpm -Fvh *19.17.6*.rpm
```
or
```
rpm -Fvh clickhous... | 2020-01-09T10:54:09 | 2020-01-09T10:54:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,976,478 | 8,591 | alesapin | Not it should work out of the box for `gentoo` and `arch`. I'll test the package after CI build. | 2020-01-10T10:21:42 | 2020-01-10T10:24:00 | {} | MEMBER |
ClickHouse | ClickHouse | 573,016,147 | 8,591 | alexey-milovidov | Ok. Now we need a snippet for the website
`Install packages for Ubuntu/Debian or CentOS/RedHat or Other Linux:` | 2020-01-10T12:26:50 | 2020-01-10T12:26:50 | {} | MEMBER |
ClickHouse | ClickHouse | 573,048,164 | 8,593 | 4ertus2 | There's also silent error with Delta codec:
```
create table test.t64_bug (
epoch UInt64 CODEC(Delta,LZ4),
_time_dec Float64
) Engine = MergeTree ORDER BY (epoch);
desc table test.t64_bug;
alter table test.t64_bug modify column epoch DEFAULT toUInt64(_time_dec) CODEC(Delta,LZ4);
desc table test.t64_... | 2020-01-10T14:05:19 | 2020-01-10T14:05:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 571,974,935 | 7,800 | filimonov | Duplicate of #8152 and #8003
Fixed in #8016
Released in 19.16.10 and newer | 2020-01-08T09:56:04 | 2020-01-08T09:56:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,272,361 | 2,269 | ramazanpolat | @den-crane
This **was** the message and the CH version **was** `19.17.2`.:
```
Received exception from server (version 19.17.2):
Code: 240. DB::Exception: Received from localhost:9000. DB::ErrnoException.
DB::ErrnoException:
Allocator: Cannot mremap memory chunk from 5.00 GiB to 10.00 GiB.,
errno: 14, strerr... | 2020-01-08T21:47:14 | 2020-01-08T21:47:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,348,410 | 8,550 | yuzhichang | No. The cluster has been up for ~ 40 days.
I managed to delete those rows by `ALTER TABLE cdp_orders ON CLUSTER ccc DROP PARTITION 'YYYYMMDD'` for each day in last 30 days.
Then `ALTER TABLE default.cdp_orders DELETE WHERE order_time >= toDateTime('2019-12-03 00:00:00')` can work.
I suspect the exception was caused ... | 2020-01-09T02:10:59 | 2020-01-09T02:17:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,942,901 | 8,541 | tisonkun | Ref #8600 | 2020-01-10T09:13:56 | 2020-01-10T09:13:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,208,096 | 8,504 | genzgd | Yes, we have ntpd running, and other tables on these servers replicate fine. We also are running 7 other shards (three replicas each) with exactly the same tables and configuration without issue. I only see the checksum problems when I alter a table to include multiple TTL columns. | 2020-01-08T18:57:52 | 2020-01-08T18:57:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,758,201 | 8,466 | volfco | Cool. I'll poke around and see how it works. If I could expose Clickhouse semi-directly to the user then that drastically reduces complexity of my system. | 2020-01-09T21:11:53 | 2020-01-09T21:11:53 | {} | NONE |
ClickHouse | ClickHouse | 572,893,891 | 8,590 | filimonov | Ok. The original case is cleared.
@ifoo can you say smth more about your case? Did you update from one of the prev versions? Or it's new setup? Any non standard settings? Table schema and format?
In my tests i have a speed about 150000 rows/sec. How can i repeat your result?
| 2020-01-10T06:30:58 | 2020-01-10T06:32:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,127,356 | 8,521 | filimonov | That is the data you have inserted to the Distributed, which was not (yet?) sent to the shards. Check the logs to understand why it was not send. | 2020-01-08T15:42:50 | 2020-01-08T15:42:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,277,821 | 2,269 | den-crane | @ramazanpolat It probably because of JOIN (LEFT JOIN AB_CS1 ON ... LEFT JOIN PB_CS1)
CH JOIN is a HashJoin it places `right` tables AB_CS1 and PB_CS1 into memory into hash-tables at whole. There is no settings to change this behavior.
How many rows in AB_CS1 and PB_CS1 ?
PS: actually CH has experimental featur... | 2020-01-08T22:00:55 | 2020-01-08T22:06:43 | {
"heart": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,464,308 | 5,601 | Sugaroverdose | @oandrew any chance that it will support basic authentication on schema registry, as well as multiple urls? | 2020-01-09T09:11:49 | 2020-01-09T09:11:49 | {} | NONE |
ClickHouse | ClickHouse | 572,530,363 | 4,295 | tavplubix | Implemented in #5209 | 2020-01-09T11:58:28 | 2020-01-09T11:58:28 | {} | MEMBER |
ClickHouse | ClickHouse | 572,555,266 | 6,128 | alesapin | https://github.com/ClickHouse/ClickHouse/pull/8591
| 2020-01-09T13:13:55 | 2020-01-09T13:13:55 | {} | MEMBER |
ClickHouse | ClickHouse | 572,633,667 | 8,578 | alexey-milovidov | Does the issue reproduce if you install ClickHouse from packages? | 2020-01-09T16:15:51 | 2020-01-09T16:15:51 | {} | MEMBER |
ClickHouse | ClickHouse | 572,690,435 | 8,592 | den-crane | 723.47 MiB ridiculously unrealistic target. CH was designed to use 10GB as a base point.
try `set max_threads=1, max_read_buffer_size=100000, max_compress_block_size=100000, min_compress_block_size=100000` | 2020-01-09T18:27:08 | 2020-01-09T18:27:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,715,063 | 8,466 | volfco | @pachico It looks like chproxy can be used in conjunction with clickhouse user filtering. It doesn't look like chproxy can restrict or filter data? | 2020-01-09T19:24:49 | 2020-01-09T19:24:49 | {} | NONE |
ClickHouse | ClickHouse | 572,953,710 | 8,571 | oandrew | > Nice 'Date Types matching:' table. Enum in input listed twice.
Yep, that's intentional - Avro enum can be inserted into either String or Enum in Clickhouse. | 2020-01-10T09:43:35 | 2020-01-10T09:43:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,071,334 | 8,570 | azat | The same change to the `debian/clickhouse-server.init` would be nice, I guess smth like below should work [^1]:
```shell
# Required-Start: $network
# Required-Stop: $network
```
[^1]: https://wiki.debian.org/LSBInitScripts | 2020-01-08T14:10:19 | 2020-01-08T14:10:19 | {} | MEMBER |
ClickHouse | ClickHouse | 572,124,716 | 8,550 | den-crane | >What's the root cause?
@yuzhichang No idea, seems a bug.
>DDLWorker
Did you run it **ON CLUSTER** `ALTER TABLE default.cdp_orders ON CLUSTER ..... ` ? | 2020-01-08T15:36:48 | 2020-01-08T15:36:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,277,242 | 8,577 | filimonov | Duplicate https://github.com/ClickHouse/ClickHouse/issues/8160
| 2020-01-08T21:59:29 | 2020-01-08T21:59:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,323,839 | 8,550 | yuzhichang | @den-crane Yes. I used to run `ALTER TABLE default.cdp_orders ON CLUSTER DELETE WHERE order_time >= 'YYYY-MM-DD 00:00:00' ` to delete orders of last 30 days(where YYYY-MM-DD depends on when it runs). The `DELETE` has worked for months and failed yesterday. | 2020-01-09T00:30:27 | 2020-01-09T00:35:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,364,986 | 8,571 | oandrew | Updated PR with more information and added some usage examples.
I'll be adding tests soon. | 2020-01-09T03:20:16 | 2020-01-09T03:25:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,409,846 | 2,269 | ramazanpolat | @den-crane The size of tables are 210M and 170M.
After setting `partial_merge_join = 1`, it WORKED! 👍
Do you think that I should open an issue for the crash?
Because being able to bring down whole CH server with an SQL looks like a bug that needs to be addressed. | 2020-01-09T06:23:59 | 2020-01-09T06:23:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,416,853 | 8,583 | kuncao | @den-crane
Thanks for your reply.
I can not upload the detail information.
Can you send me the email : caokun.nj@huawei.com and I will reply with the detail information.
Thanks a lot. | 2020-01-09T06:48:30 | 2020-01-09T06:48:30 | {} | NONE |
ClickHouse | ClickHouse | 572,561,163 | 8,591 | alexey-milovidov | I remember that @proller has replaced `useradd` to `adduser` for unknown reason, but it introduced extra dependency. | 2020-01-09T13:28:59 | 2020-01-09T13:28:59 | {} | MEMBER |
ClickHouse | ClickHouse | 572,591,575 | 8,579 | alexey-milovidov | Some tweaks are possible. For example, if log is written to terminal and nowhere else, we can apply some filters. | 2020-01-09T14:42:57 | 2020-01-09T14:42:57 | {} | MEMBER |
ClickHouse | ClickHouse | 572,644,180 | 8,579 | traceon | Thanks, will try. | 2020-01-09T16:38:19 | 2020-01-09T16:38:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,825,148 | 8,590 | den-crane | Today we cleared up with @KushnerykPavel that actually he had no problem, because a consuming rate is 50k/s, so a lag in that system is only 1second. | 2020-01-10T00:56:48 | 2020-01-10T00:56:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,137,451 | 8,425 | filimonov | /cc @4ertus2 | 2020-01-08T16:04:27 | 2020-01-08T16:04:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,264,785 | 8,549 | filimonov | > Nope so far.
Doesn't seems complicated to create, but quite important. We should have a test for that. | 2020-01-08T21:27:59 | 2020-01-08T21:27:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,329,444 | 8,550 | den-crane | real puzzle, maybe you upgraded your CH recently? | 2020-01-09T00:53:09 | 2020-01-09T00:53:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,694,754 | 8,590 | ifoo | I can confirm this. I'm running ClickHouse server version 20.1.1 revision 54431 and I see max. ingestion speeds of around 1-2k/s and accumulating a lot of lag. `kafka_num_consumers` is set to 1. I'm consuming from a single topic with one partition. | 2020-01-09T18:38:31 | 2020-01-09T18:38:31 | {} | NONE |
ClickHouse | ClickHouse | 571,973,879 | 8,003 | filimonov | Fixed in #8016
Released in 19.16.10 and newer | 2020-01-08T09:53:36 | 2020-01-08T09:53:36 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,025,083 | 7,646 | filimonov | > Can you make messages like this one more descriptive?
> ` Local: Timed out`
I have problem noted already. I will see what we can do / how we can improve that.
> If I set parameter, how to check its current value?
AFAIK there is no way to do that currently.
I will check if it's possible to show effective va... | 2020-01-08T12:22:39 | 2020-01-08T12:22:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,156,046 | 8,505 | den-crane | @filimonov No, it's a real issue
```
create table X (A Int64) Engine=MergeTree order by tuple();
insert into X select cityHash64(number) from numbers(10000000000);
```
```
19.4.5.35.
time cl --max_execution_time 1 -q 'select count() from X where A=1'
Received exception from server (version 19.4.5):
Code: 1... | 2020-01-08T16:46:28 | 2020-01-08T17:52:12 | {
"+1": 1,
"confused": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,249,681 | 2,269 | ramazanpolat | I'm facing the same problem and overcommit settings seem fine (it is zero).
`max_memory_usage` is 80G and `max_bytes_before_external_group_by` is 40G.
Any idea? | 2020-01-08T20:46:46 | 2020-01-08T20:46:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,507,088 | 2,269 | artemkamolodets | btw in my case the problem was:
i'm using clickhouse in docker container and there are special mount point /shm which is 64mb in size by default. This value is too small. Increasing it resolved the issue.
p.s. my english is best, i know :) | 2020-01-09T10:52:15 | 2020-01-09T10:52:15 | {} | NONE |
ClickHouse | ClickHouse | 572,682,527 | 2,269 | den-crane | > Do you think that I should open an issue for the crash?
I would not. I believe OOM crashes are normal behavior. | 2020-01-09T18:06:37 | 2020-01-09T18:06:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,878,146 | 8,566 | zhang2014 | The `zkutil.multi_async` failure seems to have nothing to do with the this patch. But it shouldn't have happened. Do we need to look into why ?
BTW: Maybe should provide application log and commit log of zookeeper | 2020-01-10T05:19:48 | 2020-01-10T05:19:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,942,879 | 8,601 | alexey-milovidov | Ok. BTW, 1.1 releases were all different, you should not group them inside one category. | 2020-01-10T09:13:53 | 2020-01-10T09:13:53 | {} | MEMBER |
ClickHouse | ClickHouse | 572,131,201 | 8,160 | HungUnicorn | @filimonov I receive `Cannot parse JSON string: expected opening quote: while reading the key of properties`. any suggestion? | 2020-01-08T15:50:56 | 2020-01-08T15:50:56 | {} | NONE |
ClickHouse | ClickHouse | 572,151,130 | 8,521 | volfco | Is there a way to resend this data or have it purged every couple of hours? | 2020-01-08T16:34:50 | 2020-01-08T16:34:50 | {} | NONE |
ClickHouse | ClickHouse | 572,347,857 | 8,570 | yuzhichang | @azat Updated per your suggestion. However I'm not able to test this `LSBInitScripts` change since it's obsoleted init system. | 2020-01-09T02:08:51 | 2020-01-09T02:08:51 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,458,783 | 8,577 | HungUnicorn | I saw #8056 and #8160. it's fine for my case to import the json as String type and not json type, or perhaps I misunderstand? | 2020-01-09T08:57:44 | 2020-01-09T08:59:58 | {} | NONE |
ClickHouse | ClickHouse | 572,576,742 | 7,572 | abyss7 | I will try to rebase and resolve conflicts, and take a look at the code. | 2020-01-09T14:07:59 | 2020-01-09T14:07:59 | {
"eyes": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,618,540 | 8,584 | SaltTan | This is because `drop partition` only drops active parts (and inactive parts get attached after a restart or detach/attach).
Related https://github.com/ClickHouse/ClickHouse/issues/7415 | 2020-01-09T15:42:57 | 2020-01-09T15:42:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,646,444 | 8,575 | alexey-milovidov | Ok.
I have changed to the new syntax. | 2020-01-09T16:43:09 | 2020-01-09T16:43:09 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 572,814,314 | 8,541 | meob | I solved the problem rising the minimum macos version in cmake/darwin/default_libs.cmake
I've already submitted a pull request on it | 2020-01-10T00:09:28 | 2020-01-10T00:09:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 571,915,793 | 8,550 | yuzhichang | @den-crane What's the root cause? | 2020-01-08T06:52:07 | 2020-01-08T06:52:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,060,251 | 7,646 | filimonov | > Does CH have dedicated thread for the kafka processing or Thread Poll is used (common static thread pool)?
'Schedule' thread pool is used for that, the size of that is configured by `background_schedule_pool_size` setting.
You can check the actual usage of the pool by looking at system.metrics metric name `Bac... | 2020-01-08T13:47:30 | 2020-01-08T14:05:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,263,806 | 8,549 | excitoon | > Ok. The question about passing TTL info for downloaded part is still actual, can be addressed in other PR.
>
> BTW: Do we have a test for current PR?
Yes.
Nope so far. | 2020-01-08T21:25:22 | 2020-01-08T21:25:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,455,450 | 8,571 | filimonov | Nice 'Date Types matching:' table. Enum in input listed twice. | 2020-01-09T08:49:06 | 2020-01-09T08:49:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 572,543,856 | 7,646 | ivanprostran |
> > Does CH have dedicated thread for the kafka processing or Thread Poll is used (common static thread pool)?
>
> 'Schedule' thread pool is used for that, the size of that is configured by `background_schedule_pool_size` setting.
>
> You can check the actual usage of the pool by looking at system.metrics metri... | 2020-01-09T12:39:31 | 2020-01-09T12:39:31 | {} | NONE |
ClickHouse | ClickHouse | 572,788,110 | 8,581 | excitoon | Possibly related: #7887 | 2020-01-09T22:34:07 | 2020-01-09T22:34:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 559,874,263 | 7,972 | den-crane | Yeah, you can rename them later.
And you can backup existing data using freeze https://clickhouse.yandex/docs/en/query_language/alter/#alter_freeze-partition
And use parts from shadow folder.
Steps:
1. create a new replicated table TEMP
2. ALTER TABLE table_name FREEZE
3. mv /var/lib/clickhouse/shadow/...tabl... | 2019-11-29T20:10:13 | 2019-11-29T20:14:05 | {
"+1": 6
} | CONTRIBUTOR |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.