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 | 491,481,939 | 5,231 | daniccan | Upgraded to 19.4.3.11. Now continuously getting the following error.
> (StorageReplicatedMergeTree): DB::StorageReplicatedMergeTree::queueTask()::<lambda(DB::StorageReplicatedMergeTree::LogEntryPtr&)>: Poco::Exception. Code: 1000, e.code() = 0, e.displayText() = Null pointer: _pInstance in file "/home/centos/clickho... | 2019-05-11T05:44:30 | 2019-05-11T05:45:11 | {} | NONE |
ClickHouse | ClickHouse | 491,604,260 | 5,079 | filimonov | Also such an error can happen if clickhouse-extract-from-config fails | 2019-05-12T15:14:25 | 2019-05-12T15:14:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,763,912 | 5,250 | alexey-milovidov | Linking ClickHouse with shared libraries is discouraged and never can be considered for production usage. | 2019-05-13T10:18:01 | 2019-05-13T10:18:01 | {} | MEMBER |
ClickHouse | ClickHouse | 491,893,416 | 5,257 | filimonov | >AFAIK problems in #5005 was somehow related with database with dictionary=Engine, and some dictionaries initialization problems.
> Do you use database with Engine=Dictionary? odbc dictionaries? Can you try if ClickHouse starts without dictionaries? | 2019-05-13T16:28:20 | 2019-05-13T16:28:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,913,193 | 5,257 | agelwarg | > Are you sure that you still running 19.4.3.11 after reboot?
Yes. I'm attaching a log of startup with logger level set to `trace`.
```
2019.05.13 17:21:54.633671 [ 1 ] {} <Information> : Starting ClickHouse 19.4.3.11 with revision 54416
2019.05.13 17:21:54.633736 [ 1 ] {} <Information> Application: starting u... | 2019-05-13T17:26:12 | 2019-05-13T17:26:12 | {} | NONE |
ClickHouse | ClickHouse | 492,022,521 | 5,257 | agelwarg | @den-crane I see some of those, but not at the times I would think Kubernetes would have killed it. Maybe it was straight to `SIGKILL`? I'll look around the K8s logs to see what I can find there. | 2019-05-13T23:27:39 | 2019-05-13T23:27:39 | {} | NONE |
ClickHouse | ClickHouse | 492,091,694 | 5,254 | koxt | It seems, that complex function is not the cause. Usage field name in ORDER BY is not supported either.
Only works with alias.
```
-- fail
select
test2.id as c0
from
test1
join test2 on test2.test1_id = test1.id
join test3 on test2.test3_id = test3.id
where
test1.code in ('1', '2', '3')
... | 2019-05-14T06:07:29 | 2019-05-14T06:07:29 | {} | NONE |
ClickHouse | ClickHouse | 492,230,309 | 3,997 | mcarbonneaux | I've started test but I've not the time to finalize and test for the moment.
But they must add all source related to sasl in cmakefile of rdkafka and add os sasl library in linking phase...
The ideal must to add static sasl but is more complex change.... | 2019-05-14T13:09:15 | 2019-05-14T13:09:15 | {} | NONE |
ClickHouse | ClickHouse | 492,234,702 | 5,265 | ilyabo | Thanks for the hint! I didn't realize that it's possible to switch to CSV. This can solve our issue.
Still, is this escaping defined by this TSV standard? The [library we use for TSV parsing](https://github.com/d3/d3-dsv) does not unescape these. | 2019-05-14T13:21:36 | 2019-05-14T13:21:36 | {} | NONE |
ClickHouse | ClickHouse | 491,545,760 | 5,231 | filimonov | AFAIK problems in #5005 was somehow related with database with dictionary=Engine, and some dictionaries initialization problems.
Do you use database with Engine=Dictionary? odbc dictionaries? Can you try if ClickHouse starts without dictionaries?
It would be great it we could find the root of that problem. | 2019-05-11T21:38:05 | 2019-05-11T21:38:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,568,989 | 5,223 | andyyzh | In my MacOS, I update to the latest master branch code.
If not add ‘ - DENABLE_ICU = OFF’ link failed, If add it then build passed.
Icu4c is installed, is it version problem?
$ brew search icu
==> Formulae
icu4c ✔ | 2019-05-12T06:32:52 | 2019-05-12T06:32:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,598,549 | 5,238 | janplus | Hmm, the data analyst in my company reported this bug to me. I also double checked this logic, and found that MySQL, Postgresql, IBM DB2 all leverage the same logic, so I decide to behave the same way. | 2019-05-12T14:02:17 | 2019-05-12T14:02:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,629,261 | 5,238 | alexey-milovidov | > Hmm, the data analyst in my company reported this bug to me. I also double checked this logic, and found that MySQL, Postgresql, IBM DB2 all leverage the same logic, so I decide to behave the same way.
This is the right decision, thank you! | 2019-05-12T21:06:13 | 2019-05-12T21:06:13 | {} | MEMBER |
ClickHouse | ClickHouse | 491,791,306 | 5,250 | kindred77 | thanks for comment，but how could I close the shared libraries linking？ | 2019-05-13T11:52:24 | 2019-05-13T11:52:24 | {} | NONE |
ClickHouse | ClickHouse | 492,198,387 | 5,262 | filimonov | Hm. Looks like that fix didn't solve the problem. At least the test fails.
It relates to `bad_token` flag i'm not sure what was the goal of that. | 2019-05-14T11:27:40 | 2019-05-14T12:23:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,254,164 | 5,243 | den-crane | https://clickhouse.yandex/docs/en/query_language/syntax/#notes-on-usage
And CH computes expressions only once (it's optimized)
select rand()+1 a , rand()+2 b
```
┌──────────a─┬──────────b─┐
│ 1255903112 │ 1255903113 │
└────────────┴────────────┘
``` | 2019-05-14T14:10:58 | 2019-05-14T14:10:58 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,254,867 | 4,442 | abyss7 | Should be fixed since 19.5.2.6 | 2019-05-14T14:12:42 | 2019-05-14T14:12:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,289,099 | 5,262 | filimonov | Closing it currently. Should be implemented other way.
Function like `hasToken` should be added. | 2019-05-14T15:32:33 | 2019-05-14T15:32:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,543,005 | 5,141 | filimonov | @rpfennin does that problem reproduces? What was the select during which the select crached? | 2019-05-11T20:52:05 | 2019-05-11T20:52:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,605,860 | 5,226 | hanguyen6 | Hi,
it turns out to be CPU bound. We saw CPU limit when number of queries are running in parallel due to large number of rows the server needs to load. Using PREWHERE to filter rows we were able to achieve better parallelism.
Thanks, | 2019-05-12T15:36:01 | 2019-05-12T15:36:01 | {} | NONE |
ClickHouse | ClickHouse | 492,152,037 | 5,263 | alexey-milovidov | I see significant performance gain:
272 ms. vs 146 ms.
on my machine.
To check if the JIT is used, you can run
```
sudo perf top
```
If you will see
```
DB::BinaryOperationImplBase<unsigned long, unsigned long, DB::DivideFloatingImpl<unsigned long, unsigned long>, double>::vector_vector
```
the JIT is n... | 2019-05-14T09:01:26 | 2019-05-14T09:01:26 | {} | MEMBER |
ClickHouse | ClickHouse | 492,243,488 | 5,249 | abyss7 | Also should partially fix #4308 | 2019-05-14T13:44:50 | 2019-05-14T13:44:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,254,088 | 5,265 | ilyabo | Ok, I see. Thanks for the suggestion again. Using `format CSV` solves our issue. | 2019-05-14T14:10:47 | 2019-05-14T14:10:47 | {} | NONE |
ClickHouse | ClickHouse | 492,302,450 | 5,259 | daledude | As a reminder if it means anything I *just* upgraded from 19.4.4.33 and I've been using the external xml dictionary files since version 0.x.
Thanks @filimonov. You pointed me in the right direction. I don't have any Engine=Dictionary tables but I do have dictionary.xml files. I think I fixed the "account" error by ... | 2019-05-14T16:04:56 | 2019-05-14T16:04:56 | {} | NONE |
ClickHouse | ClickHouse | 491,544,992 | 5,236 | filimonov | /cc @nikvas0 | 2019-05-11T21:24:34 | 2019-05-11T21:24:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,550,420 | 3,156 | filimonov | > Hi, I had a few questions:
You can find some overview of ClickHouse architecture here: https://clickhouse.yandex/docs/en/development/architecture/
There is also few videos (Russian)
https://www.youtube.com/watch?v=g5CDi8Nbut4
https://www.youtube.com/watch?v=PLMSA_gDdyM
https://www.youtube.com/watch?v=vbhSrZx... | 2019-05-11T23:05:05 | 2019-05-11T23:05:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,576,789 | 4,948 | alexey-milovidov | > So I was surprised to see in this case that a group by would make it stream again.
There are two algorithms for group by: single-level and two-level. Or, in other words, single-bucket and multi-bucketed. ClickHouse automatically switch to two-level algorithm for large results. This algorithm helps with parallel me... | 2019-05-12T08:42:57 | 2019-05-12T08:43:43 | {} | MEMBER |
ClickHouse | ClickHouse | 491,598,445 | 1,464 | filimonov | See also: #4753 | 2019-05-12T14:00:42 | 2019-05-12T14:00:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,729,480 | 5,124 | abyss7 | Self-merge, again. I suggest to revert this merge. | 2019-05-13T08:38:01 | 2019-05-13T08:38:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,794,080 | 3,579 | abyss7 | Should be fixed for original problem. For similar problems in the newer versions please open new issues. | 2019-05-13T12:01:44 | 2019-05-13T12:01:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,888,401 | 5,257 | filimonov | Probably same as #5231 #5005.
We need some clue / hints helping to reproduce the bug. | 2019-05-13T16:14:03 | 2019-05-13T16:14:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,173,730 | 1,861 | filimonov | PR #3598 fixes that (should work properly in versions > 19.1)
To make it work you need to have a cluster properly configured with user credentials.
Generally, when you do some DDL on non-leader replica it forwards the request to a leader, and during that forwarding "default" passwordless user was used. | 2019-05-14T10:04:20 | 2019-05-14T10:04:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,250,975 | 4,356 | abyss7 | - What is your CH version? Some frames indicate that the version is quite old.
- Do you send multiple rows inside one message or one row per message?
- Have you set the `kafka_row_delimiter` setting for Kafka table? | 2019-05-14T14:02:59 | 2019-05-14T14:02:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,577,061 | 5,238 | alexey-milovidov | That's great! | 2019-05-12T08:47:32 | 2019-05-12T08:47:32 | {} | MEMBER |
ClickHouse | ClickHouse | 491,577,113 | 5,238 | alexey-milovidov | Although it doesn't seem logical for me. | 2019-05-12T08:48:17 | 2019-05-12T08:48:17 | {} | MEMBER |
ClickHouse | ClickHouse | 491,735,775 | 1,052 | AlexAkulov | My tool is able to store backups efficiently on the disk and upload it fully or incrementally on S3.
I will be glad to get feedback and bug reports. | 2019-05-13T08:53:33 | 2019-05-13T08:53:33 | {
"+1": 3
} | NONE |
ClickHouse | ClickHouse | 491,792,401 | 5,253 | abyss7 | Fixes #2551 | 2019-05-13T11:56:08 | 2019-05-13T11:56:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,156,803 | 5,260 | alexey-milovidov | Performance test fail:
`SELECT COUNT() FROM join_table AS left LEFT JOIN join_table AS right ON left.A = right.A`
- is it expected or just a fluctuation? | 2019-05-14T09:15:28 | 2019-05-14T09:15:28 | {} | MEMBER |
ClickHouse | ClickHouse | 491,407,461 | 5,241 | den-crane | > I’d rather move indexes out of the columns definition. Similar to how partitioning, primary key are defined. No?
>
> This is more like a bug and not a feature request.
I agree and I really don't understand why indexes were placed in columns section and it broke backward compatibility for users with columns name... | 2019-05-10T19:40:18 | 2019-05-10T19:40:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,289,261 | 4,248 | turbotankist | No, there is still with problem
```
May 14 17:40:19 clickhouse-vm-9qt0 clickhouse-server[13173]: %4|1557844819.566|MAXPOLL|ClickHouse 19.6.2.11#consumer-6| [thrd:main]: Application maximum poll interval (300000ms) exceeded by 53ms (adjust max.poll.interval.ms for long-running mes
May 14 17:40:22 clickhouse-vm-9qt0 c... | 2019-05-14T15:32:56 | 2019-05-14T15:32:56 | {} | NONE |
ClickHouse | ClickHouse | 491,730,179 | 5,124 | alexey-milovidov | It is Ok. | 2019-05-13T08:40:10 | 2019-05-13T08:40:10 | {} | MEMBER |
ClickHouse | ClickHouse | 491,859,212 | 5,254 | 4ertus2 | The failed query uses complex function (multiIf(isNull(test1.code), 1, 0)) outside of select list that is not supported yet. You may try to place this statement into selected columns like this
```
SELECT test2.id AS c0, multiIf(isNull(test1.code), 1, 0) AS ord
FROM test1
INNER JOIN test2 ON test2.test1_id = test... | 2019-05-13T15:00:55 | 2019-05-13T15:07:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,560,013 | 5,243 | knoguchi | It’s a minimal example to reproduce the problem. | 2019-05-12T02:55:37 | 2019-05-12T02:55:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,894,819 | 5,257 | agelwarg | @filimonov we don't use any dictionaries. | 2019-05-13T16:32:09 | 2019-05-13T16:32:09 | {} | NONE |
ClickHouse | ClickHouse | 492,173,993 | 5,264 | filimonov | What is the version?
Most probably duplicate of #1861
PR #3598 fixes that (should work properly in versions > 19.1)
Generally, when you do some DDL on non-leader replica it forwards the request to a leader, and during that forwarding "default" passwordless user was used. | 2019-05-14T10:05:04 | 2019-05-14T10:05:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,240,385 | 5,265 | den-crane | CH needs to escape ' because of arrays
`['arr1','arr2']<tab>someString` | 2019-05-14T13:36:28 | 2019-05-14T13:36:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,237,917 | 5,262 | filimonov | Talked with @nikvas0 it's by design behaviour - to make it possible token partial match.
We need to design some better approach for that. | 2019-05-14T13:30:00 | 2019-05-14T13:53:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,544,852 | 5,243 | filimonov | Does it work without subrequest? In your case it looks trivial to avoid subrequest. | 2019-05-11T21:23:08 | 2019-05-11T21:23:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,153,573 | 5,263 | alexey-milovidov | You can also check if the JIT working from the query profile events:
https://github.com/yandex/ClickHouse/blob/master/dbms/src/Common/ProfileEvents.cpp#L91
that are available in `system.events`, `system.processes`, `system.query_log`, `system.query_thread_log` tables. | 2019-05-14T09:06:09 | 2019-05-14T09:06:09 | {} | MEMBER |
ClickHouse | ClickHouse | 492,224,724 | 5,265 | den-crane | Default http format is TSV not CSV
```
echo "SELECT 'd\'Artagnan', 'am I' format CSV" | curl http://localhost:8123 --data-binary @-
"d'Artagnan","am I"
```
Default clickhouse-client format is pretty-compact
```
:) select 'd\'Artagnan' format TSV
d\'Artagnan
```
```
:) select 'd\'Artagnan' format CSV... | 2019-05-14T12:53:36 | 2019-05-14T12:53:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,263,729 | 5,165 | abyss7 | If it doesn't solve your problem, please, let me know about more details on your scenario. | 2019-05-14T14:33:49 | 2019-05-14T14:33:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,298,195 | 5,267 | 4ertus2 | Perf says most of time we spend in ColumnVector<*>::replicate() on page faults. | 2019-05-14T15:53:59 | 2019-05-14T17:19:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,347,499 | 5,233 | den-crane | Try to validate you build with this https://github.com/yandex/ClickHouse/issues/5197#issuecomment-489462448 . | 2019-05-10T16:21:31 | 2019-05-10T16:21:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,470,895 | 5,109 | geldot | +1. Additionally, having `SELECT ALL` (with or without the `BUT` clause) include virtual columns would be a welcome feature for columnar workflows. (Or an alternative syntax that doesn't risk conflicting with column names.) | 2019-05-11T02:07:10 | 2019-05-11T02:08:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,560,132 | 5,243 | den-crane | MV never supported select from select. AFAIK no plans to support them. | 2019-05-12T02:58:11 | 2019-05-12T02:58:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,727,017 | 5,116 | YiuRULE | Ok thanks ! The changes were applied,
Just two notes :
- Concerning how we write the progress, I reverted it on the old way. So it's apply also for how we write on JSON. The problem is currently the progress concerning the writing is not on real time (we get it on the end), so it's mean we send values who will not ... | 2019-05-13T08:30:43 | 2019-05-13T08:30:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,768,135 | 5,126 | 4ertus2 | The problem is
then we do ANY JOIN we make hash table with one value per key (for right table). That's OK for ANY LEFT JOIN. It's not OK for ANY RIGHT JOIN cause user expects symmetry of left and right joins. And also it's not clear what would be best for INNER & FULL ones. It could be interesting option to get any va... | 2019-05-13T10:32:02 | 2019-05-13T10:32:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,896,375 | 5,257 | den-crane | Are you sure that you still running 19.4.3.11 after reboot? | 2019-05-13T16:36:39 | 2019-05-13T16:36:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,545,091 | 5,027 | filimonov | @Giant1 did you find the problem? | 2019-05-11T21:26:07 | 2019-05-11T21:26:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,566,701 | 5,243 | knoguchi | If that's the case CREATE MV stmt should raise error when subqueries are given. I see #2878 and #3450 that people were trying subqueries. | 2019-05-12T05:47:29 | 2019-05-12T05:47:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,550,944 | 4,948 | filimonov | ClickHouse use streaming almost everywhere, and distibute the query between threads.
Every time when thread finished it's part of work it's either passed to the client either passed to another working thread which will countine data processing.
Some tasks (like ORDER BY) require to execute final sorting in the s... | 2019-05-11T23:15:47 | 2019-05-12T07:21:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,753,363 | 5,182 | TH-HA | Hello @filimonov
To start with , yes query subquery propagated from the "root" query within the cluster with timing information. Those information can be usefull in multiple context and are generally hard to consolidate.
Typically such information is very usefull in case of troubleshooting of course , but also du... | 2019-05-13T09:44:46 | 2019-05-13T09:44:46 | {} | NONE |
ClickHouse | ClickHouse | 491,929,365 | 5,259 | den-crane | Check the log for DDLWorker messages. Probably DDLWorker during CH start found stale uncompleted distributed alter queries and tried to apply them. | 2019-05-13T18:13:31 | 2019-05-13T18:33:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,170,689 | 4,629 | bgranvea | Thank you for your review. Regarding your points:
- I'm not very happy with the mutable member in IDataType either. But to get rid of that, I would have to change the creator of all existing datatypes to take extra parameters because they all can be wrapped in a SimpleAggregateFunction. This looks like very impactin... | 2019-05-14T09:55:22 | 2019-05-14T09:55:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,236,738 | 4,248 | abyss7 | @turbotankist can you confirm that the fix in rdkafka solves your problem? If not, I will try to make some tunings on CH side. | 2019-05-14T13:27:01 | 2019-05-14T13:27:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,263,056 | 5,165 | abyss7 | Both `kafka_broker_list` and `kafka_topic_list` may have multiple comma-separated values. | 2019-05-14T14:32:19 | 2019-05-14T14:32:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,546,664 | 5,079 | filimonov | Did you disable log in your ClickHouse configs?
https://github.com/yandex/ClickHouse/blob/f7343ecb6fbeb1f93a334f23e7c42d85fce9f0e2/docker/server/entrypoint.sh#L23-L25 | 2019-05-11T21:54:08 | 2019-05-11T21:57:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,603,661 | 5,244 | svladykin | @andyyzh Looks like I have another one for you :) | 2019-05-12T15:06:12 | 2019-05-12T15:06:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,737,091 | 1,052 | blinkov | @AlexAkulov you could have saved everyone a few clicks by providing a direct link: https://github.com/AlexAkulov/clickhouse-backup/ 😉 | 2019-05-13T08:57:24 | 2019-05-13T08:57:24 | {
"+1": 4
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,793,096 | 3,997 | abyss7 | Technically the duplicate of #2551 | 2019-05-13T11:58:37 | 2019-05-13T11:58:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,917,063 | 5,257 | agelwarg | Also of note, we are deploying into a Kubernetes cluster with a Helm chart that we have developed ourselves. You will notice some `Host not found` errors. We have gone back and forth with how to deploy the Kubernetes `Service` pieces and configure the clusters, in order to deal with scenarios where DNS caching can ca... | 2019-05-13T17:37:42 | 2019-05-13T17:37:42 | {} | NONE |
ClickHouse | ClickHouse | 492,006,887 | 5,257 | agelwarg | @filimonov @den-crane I think I may have found a big piece of the puzzle that is maybe clouding the troubleshooting. We have a `livenessProbe` configured for the Clickhouse PODs. The probe checks for socket connectivity to the HTTP port (8123). I believe with the amount of data/parts we have, it was actually taking ... | 2019-05-13T22:15:44 | 2019-05-13T22:15:44 | {} | NONE |
ClickHouse | ClickHouse | 492,155,409 | 5,262 | alexey-milovidov | Why we pass `pos` by pointer not by reference?
According to our code style, we should always prefer references. | 2019-05-14T09:11:34 | 2019-05-14T09:11:34 | {} | MEMBER |
ClickHouse | ClickHouse | 492,182,589 | 5,263 | filimonov | Do you use official build? What a CH version & OS? | 2019-05-14T10:32:03 | 2019-05-14T10:32:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,597,576 | 2,404 | filimonov | You can just login to the container with bash and change file owners.
Recent ClickHouse versions give much better control on file owners in docker, see documentation on docker hub.
Closing that issue (obsolete). Please create a new issue in case of additional questions. | 2019-05-12T13:49:29 | 2019-05-12T13:51:04 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,712,292 | 5,244 | andyyzh | Run this sql, the result is fine, and it should be a type conversion problem.
```
SELECT bitmapAndCardinality(bitmapBuild([1, 2, 5]), bitmapBuild([1, 2])) AS result
┌─result─┐
│ 2 │
└────────┘
```
| 2019-05-13T07:45:26 | 2019-05-13T07:45:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,896,844 | 5,257 | filimonov | Ok. It's also useful information. Problem appears during server startup? Can you try to figure out looking in logs the sequence of actions happened before? Does it relate to certain table loading? | 2019-05-13T16:38:03 | 2019-05-13T16:38:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,576,556 | 4,948 | alexey-milovidov | > Some tasks (like ORDER BY) require to execute final sorting in the single thread so you get the result in one block (from single thread).
That's not true.
```
SELECT DISTINCT blockNumber() FROM (SELECT rand() AS x FROM numbers(1000000) ORDER BY rand())
``` | 2019-05-12T08:38:52 | 2019-05-12T08:38:52 | {} | MEMBER |
ClickHouse | ClickHouse | 491,954,629 | 5,259 | daledude | Thanks @den-crane. I'm running a single server with no distributed or replicated tables if that means anything.
I don't see any log lines containing DDLWorker. I'm probably blind but I can't even figure out what db.table this is for. | 2019-05-13T19:29:24 | 2019-05-13T19:29:24 | {} | NONE |
ClickHouse | ClickHouse | 491,960,661 | 5,259 | filimonov | Do you have database with Engine=Dictionary? Can you share your dictionaries config? Do you have `account` field in some / few of your dictionaries? Is it listed once? | 2019-05-13T19:48:31 | 2019-05-13T19:49:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,011,355 | 5,261 | alexey-milovidov | > No description provided. | 2019-05-13T22:34:15 | 2019-05-13T22:34:15 | {} | MEMBER |
ClickHouse | ClickHouse | 492,304,662 | 2,508 | abyss7 | Fixed in the master (future v19.8+). How it works:
- Introduced new global setting `stream_poll_timeout_ms`, means, for how long to wait for incoming message before consider that read() has failed. This setting is read once on server start-up and run-time changes are ignored. Zero means instant polling - not a sync ... | 2019-05-14T16:10:59 | 2019-05-14T16:10:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,402,392 | 5,241 | nvartolomei | I’d rather move indexes out of the columns definition. Similar to how partitioning, primary key are defined. No?
This is more like a bug and not a feature request. | 2019-05-10T19:20:30 | 2019-05-10T19:31:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,494,428 | 5,231 | daniccan | Found that the above is another bug in ClickHouse v19.4.3.11 - #5005 | 2019-05-11T09:13:01 | 2019-05-11T09:13:01 | {} | NONE |
ClickHouse | ClickHouse | 491,570,728 | 5,223 | filimonov | It should be something similar:
https://stackoverflow.com/questions/37335604/icu-ustdio-h-functions-undefined-symbols-on-mac-os-when-building-with-cmake
See also
http://userguide.icu-project.org/design#TOC-ICU-Binary-Compatibility:-Using-ICU-as-an-Operating-System-Level-Library
https://stackoverflow.com/questio... | 2019-05-12T07:02:37 | 2019-05-12T07:17:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,576,913 | 4,948 | alexey-milovidov | > either passed to the client either passed to another working thread which will countine data processing
Streams of data and threads of execution are independent (orthogonal) concepts in ClickHouse. Multiple data streams can be processed by a single thread and vice-versa. | 2019-05-12T08:45:08 | 2019-05-12T08:45:08 | {} | MEMBER |
ClickHouse | ClickHouse | 491,598,371 | 2,404 | filimonov | See also: #1464 #3828 #2114 #4753 | 2019-05-12T13:59:49 | 2019-05-12T13:59:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,607,296 | 5,245 | filimonov | Workaround:
```
export HOME=/tmp
``` | 2019-05-12T15:55:19 | 2019-05-12T15:55:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,746,674 | 5,249 | abyss7 | Should fix #2508 | 2019-05-13T09:24:47 | 2019-05-13T09:24:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,790,635 | 5,245 | filimonov | Similar:
```sql
:) show databases;
Code: 106. DB::ErrnoException: Cannot append history to file /home/ubuntu/.clickhouse-client-history, errno: 13, strerror: Permission denied
# clickhouse-client terminates...
``` | 2019-05-13T11:49:59 | 2019-05-13T11:49:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,021,524 | 5,257 | den-crane | Usually with TERM signal CH logs it
```
<Information> Application: Received termination signal (Terminated)
<Debug> Application: Received termination signal.
```
| 2019-05-13T23:22:34 | 2019-05-13T23:22:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,053,204 | 5,243 | knoguchi | I found an interesting feature of ClickHouse. The values are calculated in reactive manner. Is it documented ?
```
:) select 2 as a, a *2;
SELECT
2 AS a,
a * 2
┌─a─┬─multiply(2, 2)─┐
│ 2 │ 4 │
└───┴────────────────┘
```
I was able to eliminate subqueries using this feature in some l... | 2019-05-14T02:20:55 | 2019-05-14T02:20:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,281,346 | 5,237 | Onlinehead | Thanks for a message @filimonov (and for useful key:)
So, I created all tables from scratch and insert some generated data:
```sql
CREATE TABLE events
(
`EventDate` DateTime CODEC(Delta, LZ4HC),
`UserID` LowCardinality(String) CODEC(LZ4HC),
`ServiceID` LowCardinality(String) CODEC(LZ4HC),
`Rec... | 2019-05-10T13:04:17 | 2019-05-10T16:47:32 | {} | NONE |
ClickHouse | ClickHouse | 491,543,393 | 5,115 | filimonov | >
>
> @blinkov In my case there isn't communications between client and server, it's all about the clickhouse-client tool on the same machine where clickhouse-server installed.
> Could you explain how to insert data from Native format to a CH table?
Native is on-wire format. On-disk format used to store data on... | 2019-05-11T20:58:34 | 2019-05-11T20:58:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,547,991 | 4,880 | filimonov | @pliakhavets v19.5 with the fix was released, so probably you can just pick it.
Closing the issue for now, please reopen or create new if neccessery. | 2019-05-11T22:18:20 | 2019-05-11T22:18:20 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,598,203 | 2,114 | filimonov | See also #4753 | 2019-05-12T13:57:48 | 2019-05-12T13:57:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,860,608 | 5,256 | den-crane | https://clickhouse.yandex/docs/en/operations/table_engines/replication/
Data blocks are deduplicated. For multiple writes of the same data block (data blocks of the same size containing **the same rows in the same order), the block is only written once**. The reason for this is in case of network failures when the c... | 2019-05-13T15:04:24 | 2019-05-13T15:08:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 491,959,051 | 5,257 | filimonov | @agelwarg thank you, that log may be usefull. | 2019-05-13T19:43:21 | 2019-05-13T19:43:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 492,154,027 | 5,262 | alexey-milovidov | Description for the changelog? | 2019-05-14T09:07:36 | 2019-05-14T09:07:36 | {} | MEMBER |
ClickHouse | ClickHouse | 492,154,303 | 5,262 | alexey-milovidov | A test? | 2019-05-14T09:08:18 | 2019-05-14T09:08:18 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.