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 | 583,452,368 | 9,016 | tavplubix | [Reverted](https://github.com/ClickHouse/ClickHouse/pull/9033#issuecomment-582869393), backport not required | 2020-02-07T15:27:38 | 2020-02-07T15:27:38 | {} | MEMBER |
ClickHouse | ClickHouse | 583,453,549 | 9,033 | tavplubix | Backport to older branches not required | 2020-02-07T15:30:00 | 2020-02-07T15:30:00 | {} | MEMBER |
ClickHouse | ClickHouse | 584,093,440 | 9,045 | filimonov | @den-crane drop partition on Distributed?
Distributed doesn't have any partitions, why should you drop partitions from it?
@karlchan-cn, try
```sql
ALTER TABLE hits_local on cluster cluster_test DROP PARTITION 20200114;
``` | 2020-02-10T12:11:50 | 2020-02-10T12:12:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,136,037 | 8,573 | dstruck | Will do. Can't build the docker test environment at the moment, get a **BADSIG** (also on the servers). Reported the issue in a separate ticket: #9063 | 2020-02-10T14:00:52 | 2020-02-10T14:00:52 | {} | NONE |
ClickHouse | ClickHouse | 583,610,101 | 9,049 | alesapin | Will perform experiments after release. | 2020-02-07T21:02:15 | 2020-02-07T21:02:15 | {} | MEMBER |
ClickHouse | ClickHouse | 583,806,126 | 9,027 | onine007 | @den-crane it may be delayed inserts from Distributed tables.
I droped Distributed tables and recreated it , then solved ! | 2020-02-09T05:20:39 | 2020-02-09T05:20:39 | {} | NONE |
ClickHouse | ClickHouse | 584,432,719 | 9,058 | chu1070y | Thanks for your kindly answer | 2020-02-11T00:59:54 | 2020-02-11T00:59:54 | {} | NONE |
ClickHouse | ClickHouse | 583,448,814 | 8,932 | tavplubix | Backport conflict in 20.1 | 2020-02-07T15:20:15 | 2020-02-07T15:20:15 | {} | MEMBER |
ClickHouse | ClickHouse | 583,468,730 | 3,571 | sfod | I've resolved this issue on September, 2019.
Current master returns exception on the example from the initial comment.
Corresponding pull request is https://github.com/ClickHouse/ClickHouse/pull/6976
I thought that PR name is enough to link it with the issue but apparently I was wrong. | 2020-02-07T15:58:48 | 2020-02-07T15:58:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,610,425 | 9,017 | filimonov | `system.query_log` should have that info AFAIR (should be enabled). | 2020-02-07T21:02:41 | 2020-02-07T21:03:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,727,444 | 9,032 | sakthi7 | Hi @den-crane ,
Thanks for the revert .
[root@ip-172-31-8-156 ~]# clickhouse-server --version
ClickHouse server version 19.17.4.11.
[root@ip-172-31-8-156 ~]# clickhouse-client --version
ClickHouse client version 19.17.4.11.
| 2020-02-08T11:19:07 | 2020-02-08T11:19:07 | {} | NONE |
ClickHouse | ClickHouse | 583,869,585 | 9,048 | tavplubix | Backport conflict in 19.17 | 2020-02-09T17:14:11 | 2020-02-09T17:14:11 | {} | MEMBER |
ClickHouse | ClickHouse | 584,007,871 | 9,050 | alesapin | > Actually the mistake was that [here](https://github.com/ClickHouse/ClickHouse/blob/016afd6c0edc4201ef2d1c752a5ad1e10ffb3cef/dbms/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp#L793-L798) ranges were passed to MergeTreeSelectProcessor in reversed order, while it expected them in direct order. Previosly ranges ... | 2020-02-10T08:25:39 | 2020-02-10T08:25:39 | {} | MEMBER |
ClickHouse | ClickHouse | 584,101,611 | 8,995 | filimonov | BTW if we will just create a test showing the memory leak - won't address sanitizer just show the place?
https://github.com/google/sanitizers/wiki/AddressSanitizerLeakSanitizer | 2020-02-10T12:35:09 | 2020-02-10T12:35:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,220,917 | 9,059 | den-crane | >I tested but sometimes #1 is better and sometimes #2 is better.
Should be the same speed. Use what is more natural for your code.
>BETWEEN toDate('2019-12-21') AND toDate('2019-12-27')
>BETWEEN '2019-12-21' AND '2019-12-27'
No difference from performance perspective.
BUT toDate('2019-12-21') is more reliab... | 2020-02-10T16:53:47 | 2020-02-10T16:53:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,484,547 | 8,567 | KochetovNicolai | That was the reason: https://github.com/ClickHouse/ClickHouse/commit/42692ad07b6e37fe3697c6611fcc672f358048bb
Unfortunately, I've missed the slides source, so can't fix mistake there. | 2020-02-11T05:35:44 | 2020-02-11T05:35:44 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 584,511,342 | 9,063 | dstruck | Weird. Works today (2020-02-11) fine, closing. | 2020-02-11T07:42:38 | 2020-02-11T07:42:38 | {} | NONE |
ClickHouse | ClickHouse | 583,492,593 | 8,964 | CurtizJ | I reproduced it by this example:
```
drop table if exists t;
create table t (a Int, b Int) engine = MergeTree order by (a, b) settings index_granularity = 400;
insert into t select 0, 0 from numbers(50);
insert into t select 0, 1 from numbers(350);
insert into t select 1, 2 from numbers(400);
insert into t... | 2020-02-07T16:44:02 | 2020-02-07T16:44:02 | {} | MEMBER |
ClickHouse | ClickHouse | 584,073,879 | 8,989 | alesapin | > key condition negated incorrectly for monotonic functions #9034 -- a 2-year old bug, can release without fix
This seems like a blocker for 20.2, but not for patch versions 20.1 and 19.17. | 2020-02-10T11:15:39 | 2020-02-10T11:15:39 | {} | MEMBER |
ClickHouse | ClickHouse | 584,110,692 | 9,060 | 4ertus2 | Most probably it means there's a mutation on some of your parts and this function returned false
```
auto can_merge = [this, &lock] (const DataPartPtr & left, const DataPartPtr & right, String *)
{
return !currently_merging_mutating_parts.count(left) && !currently_merging_mutating_part... | 2020-02-10T13:00:19 | 2020-02-10T13:00:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,126,617 | 9,040 | filimonov | Very general info about use case:
1) dictionary source is ClickHouse
2) dictionary is configured via xml
3) underling table has id field of type UInt32 (not UInt64)
4) crash happen during (probably scheduled by lifetime) reload during the following query:
`INSERT INTO table SELECT <cols>, dictGetString( ... ) FRO... | 2020-02-10T13:39:03 | 2020-02-10T13:40:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,130,210 | 8,943 | CurtizJ | Looks ok, but what about performance tests slow down. Is it unrelated? | 2020-02-10T13:47:35 | 2020-02-10T13:47:35 | {} | MEMBER |
ClickHouse | ClickHouse | 584,288,748 | 9,045 | den-crane | My understanding that `test.hits` is MergeTree not Distributed | 2020-02-10T18:43:46 | 2020-02-10T18:43:46 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,432,510 | 9,058 | den-crane | you can.
DOUBLE QUOTE in tsv is a usual character, not special.
```
SELECT
'"aaaa',
111
FORMAT TSV
"aaaa 111
``` | 2020-02-11T00:59:06 | 2020-02-11T00:59:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,442,075 | 9,043 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/8165 | 2020-02-07T15:08:15 | 2020-02-07T15:08:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,681,608 | 9,027 | den-crane | >No , I don't insert any data . how long time can be delayed ? it takes 2 days ?
Maybe.
>Sorry , they are folders not dictionaries .
Can you output `ls -l`. I don't quite understand what folders. | 2020-02-08T01:16:15 | 2020-02-08T01:16:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,091,954 | 7,415 | filimonov | Fixed in #8602 | 2020-02-10T12:07:24 | 2020-02-10T12:07:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,315,616 | 8,999 | hatrena | a simple query for checking stats like
`SELECT SUM(clicks), date FROM default.vm_click_line_chart WHERE date > '2020-02-01' AND organization_id = 'XXX' group by date`
The result of that query differs in every pod. | 2020-02-10T19:44:19 | 2020-02-10T19:44:19 | {} | NONE |
ClickHouse | ClickHouse | 584,757,913 | 9,075 | nikitamikhaylov | @nvartolomei Do you use copier? Do you disable merges before copying? | 2020-02-11T17:34:30 | 2020-02-11T17:34:30 | {} | MEMBER |
ClickHouse | ClickHouse | 584,348,976 | 9,009 | akuzm | Alexey asked me not to do this, because these function are actually deterministic. Will probably have to introduce `canBeCalculatedAtQueryOptimizationStage`, or continue the wonderful `name == "ignore"` practice. | 2020-02-10T20:54:18 | 2020-02-10T20:54:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,412,252 | 9,058 | chu1070y | Can`t I use DOUBLE QUOTE in tsv also? | 2020-02-10T23:38:27 | 2020-02-10T23:38:52 | {} | NONE |
ClickHouse | ClickHouse | 584,643,153 | 9,078 | filimonov | @excitoon | 2020-02-11T13:50:32 | 2020-02-11T13:50:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,787,757 | 9,081 | den-crane | `POPULATE AS`
you have 2 replicas
clicks_sharded R1 (1 row XXXX)
clicks_sharded R2 (1 row XXXX) -- the same one row
Now on `create MV POPULATE` the first replica will select `1 row XXXX` and will insert it locally to it's storage. The second replica also will select 1 row XXXX and will insert it locally to it... | 2020-02-11T18:39:59 | 2020-02-11T18:40:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,153,648 | 9,046 | VaBezruchko | Artem,
does "select distinct " guarantee unique results? Is it matter data replication and sharding for unique result with "select distinct" query?
Data was inserted by followed query:
`INSERT INTO data.table1_d (dated, id, column1, is_surrogate) values(?,?,?,?)`
The problem is in parameter 'optimize_skip_u... | 2020-02-10T14:39:22 | 2020-02-10T14:39:22 | {} | NONE |
ClickHouse | ClickHouse | 584,248,847 | 8,999 | hatrena | I tested the MV in our staging where we have 1 shard and 2 replicas. everything is fine.
Then I have tested it in our production where we have 2 shards and 2 replicas in each. we also have 4 Kubernetes pods.
I have checked in all 4 pods and the MV does exist in all of them, but the result of the same query is d... | 2020-02-10T17:53:12 | 2020-02-10T17:53:12 | {} | NONE |
ClickHouse | ClickHouse | 584,499,971 | 2,601 | filimonov | https://github.com/ClickHouse/ClickHouse/pull/7769 | 2020-02-11T06:55:29 | 2020-02-11T06:55:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,714,502 | 9,078 | aadant | Please also make sure the error message is explicit rather than "DB::Exception: Trying to ALTER key column" | 2020-02-11T16:13:54 | 2020-02-11T16:13:54 | {} | NONE |
ClickHouse | ClickHouse | 583,431,854 | 9,044 | den-crane | By default CH expects in TSV Nulls encoded as \N
You can use `--input_format_tsv_empty_as_default arg Treat empty fields in TSV input as default values.`
```
cat test.tsv | clickhouse-client --input_format_tsv_empty_as_default=1 --query="INSERT INTO default.test3 FORMAT TSV"
SELECT *
FROM default.test3
... | 2020-02-07T14:59:43 | 2020-02-07T14:59:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,443,889 | 9,032 | den-crane | CH version?
cc @yurriy | 2020-02-07T15:11:34 | 2020-02-07T15:11:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,733,605 | 9,048 | alesapin | > Did you find out why the order of rows changes? Theoretically, when reading sequentially in a single thread, it should be the same.
Still not, but I'll try to find the next week. Anyway, when we use `InterpreterSelectQuery` and want to preserve order we have to add `ORDER BY`. We don't provide any guarantees even ... | 2020-02-08T12:47:13 | 2020-02-08T12:47:13 | {} | MEMBER |
ClickHouse | ClickHouse | 583,755,606 | 9,051 | filimonov | 1) conf.d content added both to users.xml and config.xml (yes, it's a bit weird)
2) i think warning may be good, but exception can make pain for updates (clickhouse just don't start with same config is bad UX)
| 2020-02-08T17:01:44 | 2020-02-08T17:02:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,092,217 | 8,584 | filimonov | Fixed in #8602 | 2020-02-10T12:08:08 | 2020-02-10T12:08:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,104,786 | 9,050 | CurtizJ | [This](https://github.com/ClickHouse/ClickHouse/blob/2734347cc0c679fb979f5754cf8fe11b67790ad8/dbms/src/Storages/MergeTree/MergeTreeThreadSelectBlockInputProcessor.cpp#L69) and [this](https://github.com/ClickHouse/ClickHouse/blob/9251fb803afeb8109aaa612f4d5252f1d787fd50/dbms/src/Storages/MergeTree/MergeTreeThreadSelectB... | 2020-02-10T12:44:27 | 2020-02-10T12:45:45 | {} | MEMBER |
ClickHouse | ClickHouse | 584,105,426 | 9,004 | filimonov | Related https://github.com/ClickHouse/ClickHouse/pull/8464/ (we prefer to avoid PIC) | 2020-02-10T12:46:19 | 2020-02-10T12:46:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,119,876 | 9,021 | filimonov | Maybe it would be ok to send an exception like 'Column port is ambiguous` in both cases, instead of trying to resolve the conflict. | 2020-02-10T13:23:17 | 2020-02-10T13:23:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,197,938 | 9,054 | 4ertus2 | There're workarounds to build with system library. Use these cmake definitions
```
option (USE_LIBCXX "Use libc++ and libc++abi instead of libstdc++" ${NOT_UNBUNDLED})
option (USE_INTERNAL_LIBCXX_LIBRARY "Set to FALSE to use system libcxx and libcxxabi libraries instead of bundled" ${USE_INTERNAL_LIBCXX_LIBRARY_DE... | 2020-02-10T16:06:27 | 2020-02-10T16:06:36 | {
"confused": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,608,433 | 9,071 | 4ertus2 | You've got some error for query that run through ConnectionPoolWithFailover. Probably it's some distributed query that aborted at destination point for some reason. Or it could be just an anavaliable node. You may get the times of errors and find other errors on destinations. There could be real errors there. | 2020-02-11T12:18:07 | 2020-02-11T12:18:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,776,502 | 9,000 | oandrew | @alexey-milovidov | 2020-02-11T18:15:41 | 2020-02-11T18:15:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,808,212 | 9,053 | den-crane | Check this https://github.com/ClickHouse/ClickHouse/issues/6720#issuecomment-526045768 | 2020-02-09T06:00:22 | 2020-02-09T06:00:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,124,548 | 9,046 | 4ertus2 | It's not clear what nodes are used for replication and what are distributed. It's also not clear where you insert same keys: in any replica or "one key - one replica" mode.
I can imagine a configuration then you will have stale ids cause of insertion is in all replicas, Distributed is on replicated nodes and there's... | 2020-02-10T13:34:14 | 2020-02-10T13:34:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,369,073 | 8,964 | doom369 | @CurtizJ created VM, email sent. Thanks. | 2020-02-07T12:26:12 | 2020-02-07T12:26:12 | {} | NONE |
ClickHouse | ClickHouse | 583,419,816 | 4,242 | nvartolomei | It looks like there is no code at all for doing these sort of optimisations. @4ertus2 any plans on roadmap for them? | 2020-02-07T14:45:18 | 2020-02-07T14:45:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,440,696 | 9,044 | chu1070y | Thanks a lot. It worked | 2020-02-07T15:05:29 | 2020-02-07T15:05:29 | {} | NONE |
ClickHouse | ClickHouse | 583,536,763 | 8,964 | doom369 | @CurtizJ I see, thanks. From what I understand I'm the only one who see this issue is because our test env. using low-end VM with only 1 core (and thus `max_threads=1`). Interesting. However, those still look like regression introduced in the latest version. Before the update the were no errors. (and we have multiple s... | 2020-02-07T18:15:50 | 2020-02-07T18:19:01 | {} | NONE |
ClickHouse | ClickHouse | 583,541,524 | 8,988 | nikitamikhaylov | It will be necessary to determine conditions of applicability of this feature. I mean structure of old and new MV. Only add colums as in tests? Or remove some etc. | 2020-02-07T18:26:38 | 2020-02-07T18:26:38 | {} | MEMBER |
ClickHouse | ClickHouse | 583,609,944 | 9,049 | alesapin | gcc looks for linker:
```
[pid 13608] stat("/usr/lib/gcc/x86_64-linux-gnu/9/real-ld", 0x7ffef6e21300) = -1 ENOENT (No such file or directory)
[pid 13608] stat("/usr/lib/gcc/x86_64-linux-gnu/9/... | 2020-02-07T21:01:49 | 2020-02-07T21:01:49 | {} | MEMBER |
ClickHouse | ClickHouse | 583,623,138 | 8,964 | CurtizJ | > reading in 1 thread
Only this case was affected.
| 2020-02-07T21:32:07 | 2020-02-07T21:32:15 | {} | MEMBER |
ClickHouse | ClickHouse | 583,634,444 | 9,048 | akuzm | Did you find out why the order of rows changes? Theoretically, when reading sequentially in a single thread, it should be the same. | 2020-02-07T22:00:44 | 2020-02-07T22:00:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,201,938 | 9,058 | den-crane | \ -- (char(92)) is a special character for TSV and must be escaped with `slash (\)`
```
SELECT char(92)
FORMAT TSV
\\
``` | 2020-02-10T16:14:43 | 2020-02-10T16:31:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,529,264 | 9,032 | yurriy | > Now, why it is not working with MySQL 8 ? It is because of the character set ? Because the default charset is UTF8MB4 on MySQL 8 .
I will look into that. | 2020-02-11T08:46:41 | 2020-02-11T08:46:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,351,459 | 8,964 | doom369 | @CurtizJ I can prepare VM with the issue and give you access to it. Would that help? | 2020-02-07T11:28:46 | 2020-02-07T11:29:26 | {} | NONE |
ClickHouse | ClickHouse | 583,400,331 | 8,957 | alesapin | Performance not related. | 2020-02-07T13:59:32 | 2020-02-07T13:59:32 | {} | MEMBER |
ClickHouse | ClickHouse | 583,448,354 | 8,908 | tavplubix | Backport conflict in 20.1 | 2020-02-07T15:19:29 | 2020-02-07T15:19:29 | {} | MEMBER |
ClickHouse | ClickHouse | 583,538,054 | 9,027 | den-crane | > result is increase
>this result is fast changing .
So you are still inserting data. It could be delayed inserts from Distributed tables.
>I can see a lot of dictionary like 20200205-9*** in table data dictionary
What stands for dictionary? table data dictionary? What do you mean? | 2020-02-07T18:18:47 | 2020-02-07T18:18:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,712,480 | 9,051 | filimonov | Not sure if it's a good idea | 2020-02-08T07:37:20 | 2020-02-08T07:37:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,739,222 | 9,051 | akuzm | > Not sure if it's a good idea
Why not, is it going to break something? | 2020-02-08T14:01:55 | 2020-02-08T14:01:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,133,307 | 8,943 | akuzm | > Looks ok, but what about performance tests slow down. Is it unrelated?
Probably unrelated. It doesn't reproduce, and we only have query_log and query_thread_log enabled in performance tests, so there shouldn't be any influence. | 2020-02-10T13:54:50 | 2020-02-10T13:55:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,827,133 | 9,081 | hatrena | In the meantime, I also made this change and it solved the problem. how does it explain?
```
POPULATE AS
SELECT
SUM(rejected) AS clr,
(count() - clr) AS cla,
sumIf(revenue, rejected = 1) AS rr,
sumIf(revenue, rejected = 0) AS ra,
sumIf(payout, rejected = 1) AS pr,
sumIf(payout, reject... | 2020-02-11T20:08:31 | 2020-02-11T20:08:31 | {} | NONE |
ClickHouse | ClickHouse | 583,619,221 | 9,050 | CurtizJ | Actually the mistake was that [here](https://github.com/ClickHouse/ClickHouse/blob/016afd6c0edc4201ef2d1c752a5ad1e10ffb3cef/dbms/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp#L793-L798) ranges were passed to MergeTreeSelectProcessor in reversed order, while it expected them in direct order. Previosly ranges we... | 2020-02-07T21:23:23 | 2020-02-07T21:24:13 | {} | MEMBER |
ClickHouse | ClickHouse | 583,728,699 | 9,032 | sakthi7 | Hi @den-crane ,
Thanks for the hint . I have tried with the MySQL 5.7 client . It is working fine for me .
**Logs :**
[root@ip-172-31-8-156 bin]# ./mysql --version
./mysql Ver 14.14 Distrib 5.7.28, for linux-glibc2.12 (x86_64) using EditLine wrapper
[root@ip-172-31-8-156 bin]#
[root@ip-172-31-8-156 bin... | 2020-02-08T11:37:55 | 2020-02-08T11:38:23 | {} | NONE |
ClickHouse | ClickHouse | 584,157,569 | 9,064 | grootman2007 | crashes even with this query:
SELECT
date,
client,
SUM(shows) AS sshows,
SUM(clicks) AS clicks,
SUM(sum) AS sum,
SUM(trans) AS trans,
SUM(val) AS val,
SUM(marg) AS marg,
SUM(trans_mcf) AS trans_mcf,
SUM(val_mcf) AS val_mcf
FROM yad_auto.`yad_stats_b`
WHERE ( (client=... | 2020-02-10T14:47:30 | 2020-02-10T15:22:25 | {} | NONE |
ClickHouse | ClickHouse | 584,220,547 | 9,064 | CurtizJ | I'm not sure, that it's the same issue. @grootman2007, please, can you try to reproduce on 20.1.4? | 2020-02-10T16:53:04 | 2020-02-10T16:53:47 | {} | MEMBER |
ClickHouse | ClickHouse | 584,290,628 | 8,999 | den-crane | > but the result of the same query is different on each pod
What query? | 2020-02-10T18:48:12 | 2020-02-10T18:48:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,631,618 | 9,077 | 4ertus2 | It looks like a very special case to get `a chunk of numbers`. Could you describe your task where you need it? | 2020-02-11T13:22:49 | 2020-02-11T13:22:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,716,534 | 9,074 | den-crane | value UInt64 CODEC(DefaultCompression) -- solves https://github.com/ClickHouse/ClickHouse/issues/6227 | 2020-02-11T16:17:53 | 2020-02-11T16:17:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,347,058 | 8,964 | CurtizJ | This bug is very hard to reproduce. It would be great if you could send data parts from table privately to our mailing list clickhouse-feedback@yandex-team.ru or make a reproducible example on synthetic data. | 2020-02-07T11:15:45 | 2020-02-07T11:15:45 | {} | MEMBER |
ClickHouse | ClickHouse | 584,090,510 | 8,573 | filimonov | @dstruck Can you please recheck on 19.17.9 ? Fix for timeouts was released there. | 2020-02-10T12:03:01 | 2020-02-10T12:03:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,095,720 | 7,798 | filimonov | Still there at 19.17.9 | 2020-02-10T12:17:59 | 2020-02-10T12:17:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,246,801 | 9,063 | filimonov | That's strange. It works for me. Can you retry with official instruction:
```
sudo apt-get install dirmngr
sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv E0C56BD4
## ^^ ensure that command succeeded (sometimes keyservers are slow / unresponding, in that case you can use another keyserver).
ech... | 2020-02-10T17:48:29 | 2020-02-10T17:48:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,578,438 | 8,976 | sundarv85 | @den-crane The logs when the issue happened in Clickhouse are not available any longer. However, now the clickhouse is working correctly and there are no other logs that we see in Clickhouse.
The only error seems to be in zookeeper itself and its log shows only this information, as posted in the issue. Posting it ag... | 2020-02-11T10:53:59 | 2020-02-11T10:53:59 | {} | NONE |
ClickHouse | ClickHouse | 584,593,847 | 4,469 | filimonov | From discussions:
> Возможные проблемы:
> 1) Мы можем сделать materialize index на колонку, которая сама еще не материализована
> И если поменять default, то индекс и колонка разбегутся
>
> 2) вот это: https://github.com/ClickHouse/ClickHouse/issues/6913
>
> Наверное правильно - запретить делать MATERIALIZE... | 2020-02-11T11:35:08 | 2020-02-11T11:37:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,753,629 | 9,075 | nikitamikhaylov | Earlier when Copier faced some troubles, it had to drop the whole partition. Also each partition may be "smeared" to many shards (month as a partition key for example) because of using Distributed table. Operation of copying each partition was successful only if all the shards that contain particular partition done co... | 2020-02-11T17:25:43 | 2020-02-11T17:25:43 | {} | MEMBER |
ClickHouse | ClickHouse | 584,753,789 | 9,075 | nikitamikhaylov | How we split out partition:
``` SELECT * FROM ... WHERE partition_key = key AND cityHash64(*) % N= n```
And if N is big enough, distribution will be almost uniform (I hope it will ) :neutral_face: | 2020-02-11T17:26:00 | 2020-02-11T17:30:23 | {} | MEMBER |
ClickHouse | ClickHouse | 584,772,677 | 9,072 | den-crane | jemalloc (jemalloc.net) caches memory to save allocations/deallocations.
```
create table ttt(c1 Int64, c2 Int64) Engine=MergeTree order by c1;
insert into ttt select number,number%103 from numbers(1000000)
```
1 hour / 2 benchmarks in parallel
```
echo 'SELECT c2, topK(10)(c1) FROM ttt GROUP BY c2'|clickhouse... | 2020-02-11T18:07:13 | 2020-02-11T18:07:13 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,449,317 | 8,949 | tavplubix | Backport to 20.1 not required | 2020-02-07T15:21:14 | 2020-02-07T15:21:14 | {} | MEMBER |
ClickHouse | ClickHouse | 584,102,202 | 9,001 | filimonov | Not sure if it's always easy (i.e. StripeLog) | 2020-02-10T12:36:52 | 2020-02-10T12:36:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,460,253 | 9,045 | filimonov | > My understanding that `test.hits` is MergeTree not Distributed
You probably right: 'test.hist to store data created by below script: CREATE TABLE test.hits_local'. Misleading. :confused:
| 2020-02-11T03:02:34 | 2020-02-11T03:02:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,440,843 | 9,045 | den-crane | Not sure about 19.11. In latest (supported) 19.16+ versions you can.
>in just one execution.
Anyway, even `on cluster` is not atomic. | 2020-02-07T15:05:43 | 2020-02-07T15:05:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,551,502 | 9,048 | alesapin | Unit tests — Segmentation fault -- already fixed on master, not related to changes. | 2020-02-07T18:49:30 | 2020-02-07T18:49:30 | {} | MEMBER |
ClickHouse | ClickHouse | 583,612,597 | 9,025 | filimonov | What happened? | 2020-02-07T21:07:37 | 2020-02-07T21:07:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,647,309 | 9,025 | akuzm | > What happened?
We had to revert it while investigating a test failure. It's back already:
https://github.com/ClickHouse/ClickHouse/commit/d9073f27b77f14c13636c0fe4ac1dc4260e433a6 | 2020-02-07T22:40:32 | 2020-02-07T22:40:32 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,137,814 | 9,059 | filimonov | > I tested but sometimes 1 is better and sometimes 2 is better.
> I can't know which one is better for performance.
I think they should work ~ the same speed. Maybe you have difference due to cache hit/miss? | 2020-02-10T14:05:05 | 2020-02-10T14:05:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,688,698 | 9,075 | nvartolomei | 🥳
> Now clickhouse-copier splits each partition in number of pieces and copies them independently.
Why is this "Performance Improvement"? Because we can use multiple replicas as partition source or for other reason?
How atomicity is handled in this case? On error abort all jobs, drop partition and try again a... | 2020-02-11T15:23:09 | 2020-02-11T15:23:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 583,678,376 | 9,027 | onine007 | @den-crane
> > result is increase
> > this result is fast changing .
>
> So you are still inserting data. It could be delayed inserts from Distributed tables.
>
No , I don't insert any data . how long time can be delayed ? it takes 2 days ?
> > I can see a lot of dictionary like 20200205-9*** in table ... | 2020-02-08T00:55:12 | 2020-02-08T00:55:12 | {} | NONE |
ClickHouse | ClickHouse | 584,088,720 | 8,553 | filimonov | https://github.com/ClickHouse/ClickHouse/pull/8892 | 2020-02-10T11:57:49 | 2020-02-10T11:57:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,117,938 | 9,021 | filimonov | It's because `port` exists in `system.query_log`. So the workaround is trivial, just pick another alias to avoid ambiguity.
Nevertheless, it's a bug. Current behavior introduced somewhere around 19.4, but before that, it was also wrong (it was silently masking other value)
Minimal test case:
```
drop table if e... | 2020-02-10T13:18:20 | 2020-02-10T13:20:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,148,395 | 9,064 | akuzm | Looks like a duplicate of #8997 | 2020-02-10T14:28:21 | 2020-02-10T14:28:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,181,775 | 9,035 | 4ertus2 | Local tests with tsan are OK. | 2020-02-10T15:36:09 | 2020-02-10T15:36:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 584,657,028 | 9,077 | ramazanpolat | > It looks like a very special case to get `a chunk of numbers`. Could you describe your task where you need it?
This may seem like it is specific to my problem but considering that other databases also implemented something like this, it shows that in some cases it helps a lot.
### Use case 1
Some systems pla... | 2020-02-11T14:19:20 | 2020-02-11T14:19:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 591,567,134 | 8,886 | alexey-milovidov | Found the cause, will fix. | 2020-02-26T18:15:50 | 2020-02-26T18:16:01 | {
"+1": 1
} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.