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 | 638,467,465 | 11,412 | den-crane | >ReplicatedMergeTree
Check `select * from system.replication_queue ` at both servers. Seems one server could not connect to another. They (ReplicatedMergeTree) use `hostname -f` hostnames and announce itself using zookeeper.
ReplicatedMergeTree does't use `remote_servers`.
BTW it must be true `<internal_repli... | 2020-06-03T21:18:36 | 2020-06-03T21:19:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,872,944 | 9,113 | CurtizJ | Waiting for some additional tests, because now `optimize_aggregation_in_order` is disabled by default. | 2020-06-04T14:08:31 | 2020-06-04T14:35:44 | {} | MEMBER |
ClickHouse | ClickHouse | 638,938,777 | 11,228 | crakjie | I have the same problem on the stable version `20.3.8.53`.
```
SELECT
dictGet('countryId', 'country', toUInt64(number)) AS country
FROM numbers(2)
GROUP BY country
```
` DB::Exception: Function dictGet... takes 3 or 4 arguments.`
This query was working with `19.15.3.6`
| 2020-06-04T15:45:58 | 2020-06-04T15:48:19 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 630,286,453 | 10,353 | 4ertus2 | In this PR we are interested in codecs implemented here:
- https://github.com/lemire
- https://github.com/powturbo
- https://github.com/schizofreny/middle-out (need check how much joke is in this joke)
They are expected to be performance optimized. | 2020-05-18T16:11:40 | 2020-06-02T14:36:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 637,589,989 | 11,335 | alexey-milovidov | And I don't understand why one of the builds failed. | 2020-06-02T14:45:19 | 2020-06-02T14:45:19 | {} | MEMBER |
ClickHouse | ClickHouse | 637,804,744 | 10,242 | vitlibar | Please fix compilation errors with `ENABLE_SSL=0`.
That must fix Yandex synchronization check. | 2020-06-02T21:02:43 | 2020-06-02T21:02:43 | {} | MEMBER |
ClickHouse | ClickHouse | 637,877,191 | 8,683 | azat | @alexey-milovidov taken | 2020-06-03T00:03:08 | 2020-06-03T00:03:08 | {} | MEMBER |
ClickHouse | ClickHouse | 638,205,651 | 11,405 | den-crane | I think it should be `WHERE stamp_updated > toDateTime(0)` and it is the expected behavior. | 2020-06-03T13:42:45 | 2020-06-03T13:42:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,758,008 | 11,405 | hyberdk | Hi Denis (@den-crane),
thanks for the quick reply, that did indeed fix the issue.. Do you have any idea why it would work in one case and not the other? | 2020-06-04T10:15:28 | 2020-06-04T10:15:28 | {} | NONE |
ClickHouse | ClickHouse | 638,792,240 | 11,422 | qoega | Converted to draft for now. | 2020-06-04T11:32:45 | 2020-06-04T11:32:45 | {} | MEMBER |
ClickHouse | ClickHouse | 638,947,986 | 11,439 | alesapin | tidy build passed. | 2020-06-04T15:57:34 | 2020-06-04T15:57:34 | {} | MEMBER |
ClickHouse | ClickHouse | 637,620,169 | 11,238 | qoega | Related to #10411 | 2020-06-02T15:28:29 | 2020-06-02T15:28:29 | {} | MEMBER |
ClickHouse | ClickHouse | 637,680,558 | 11,319 | Turum | When I changed max_execution_time parameter, exception was still thrown as "Read timed out", but all records ware inserted into click house table. Is it possible to fix it somehow? | 2020-06-02T16:57:26 | 2020-06-02T16:57:26 | {} | NONE |
ClickHouse | ClickHouse | 637,809,599 | 11,350 | alexey-milovidov | `test_ttl_move/test.py::test_concurrent_alter_with_ttl_move[concurrently_altering_ttl_mt-MergeTree()]`
Flaky test, fixed in #11355 | 2020-06-02T21:12:48 | 2020-06-02T21:12:48 | {} | MEMBER |
ClickHouse | ClickHouse | 638,075,987 | 11,259 | akuzm | msan unit test failure doesn't look related to this PR. | 2020-06-03T09:23:01 | 2020-06-03T09:23:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,224,260 | 11,406 | thyn | Updates. I've found a long queue in distributed table path awaitng to upload. But I can't find any logs or reason why they are not sent.
But at different time I have Connection reset by peer, Stack trace (when copying this message, always include the lines below):
```
2020.06.03 09:38:38.508736 [ 1460 ] {} <... | 2020-06-03T14:11:35 | 2020-06-03T14:36:16 | {} | NONE |
ClickHouse | ClickHouse | 603,771,818 | 4,567 | 4ertus2 | related #5046 #10276 | 2020-03-25T10:48:47 | 2020-06-03T16:41:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,447,807 | 11,417 | amartin916 | Ok my example didn't actually reproduce what I was seeing, I made it too simple :)
What I found that caused the issue was adding the AS to the column in the materialized view.
So
```
CREATE MATERIALIZED VIEW summary_test_source_table
(`replay_date` Date, `test_column_max` AggregateFunction(max,... | 2020-06-03T20:36:27 | 2020-06-03T20:36:27 | {} | NONE |
ClickHouse | ClickHouse | 638,715,173 | 11,135 | zhuyoucui0909 | v20.4.4.18 is a good version for us. We have adopted to v20.4.4.18 version.
You can add the reproduce steps on this issue to your regression.
And you can also use distributed table to replace the remote table call. | 2020-06-04T08:56:56 | 2020-06-04T08:56:56 | {} | NONE |
ClickHouse | ClickHouse | 638,940,624 | 11,288 | filimonov | Kafka changes can be reviewed/merged: #11388 | 2020-06-04T15:47:28 | 2020-06-04T15:47:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 637,573,527 | 11,348 | nikvas0 | `SELECT database, name, uuid, metadata_modification_time FROM system.tables` should not reload tables.
I think a wrapper can be a good solution. If the current table does not exist, it can simply return "dropped" as name.
Another approach is to implement some counters (like shared_ptr) for tables in CachedTable struc... | 2020-06-02T14:19:06 | 2020-06-02T14:19:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,045,045 | 11,392 | qoega | @alesapin AFAIK it was intended incompatible change | 2020-06-03T08:28:08 | 2020-06-03T08:28:08 | {} | MEMBER |
ClickHouse | ClickHouse | 638,046,415 | 11,392 | alesapin | Yes, now you have to specify a datatype for columns when you add them with `ALTER` query. | 2020-06-03T08:30:53 | 2020-06-03T08:30:53 | {} | MEMBER |
ClickHouse | ClickHouse | 638,142,938 | 11,386 | KochetovNicolai | Integration tests are not connected with changes form pr. | 2020-06-03T11:42:47 | 2020-06-03T11:42:47 | {} | MEMBER |
ClickHouse | ClickHouse | 638,199,540 | 11,234 | traceon | Pushed a fix. | 2020-06-03T13:32:49 | 2020-06-03T13:32:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,875,632 | 11,223 | 4ertus2 | Cannot reproduce on master without data. | 2020-06-04T14:13:01 | 2020-06-04T14:13:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,098,751 | 11,372 | akuzm | > > Is there some place where signed int is used for row number?
>
> Yes. It is not so easy for fix.
> https://github.com/ClickHouse/ClickHouse/blob/06446b4f08a142d6f1bc30664c47ded88ab51782/src/Processors/LimitTransform.cpp#L283-L286
It should be possible to calculate absolute row numbers for chunk start and chu... | 2020-06-03T10:10:03 | 2020-06-03T10:10:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,202,603 | 11,409 | den-crane | Do you mean you have constant number of rows and xxx does not grow with time? | 2020-06-03T13:37:56 | 2020-06-03T13:37:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,315,513 | 10,276 | 4ertus2 | Similar problem with simple join. Redefinition of `x`:
```
CREATE TABLE t (`x` UInt32, `lc` LowCardinality(String)) ENGINE = Memory;
CREATE TABLE nr (`x` Nullable(UInt32), `lc` Nullable(String)) ENGINE = Memory;
INSERT INTO t VALUES (0, 'str');
INSERT INTO nr VALUES (0, 'str');
SELECT r.lc, length(r.lc) x F... | 2020-06-03T16:40:28 | 2020-06-03T16:40:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,413,373 | 11,396 | den-crane | @AlexanderChiuluvB Do you use Replicated Engine or plain MergeTree ?
If you don't use Replicated you need to describe one more cluster with with one shard many replicas.
If you use Replicated table then just omit {shard} macros in ZK table path and don't use Distributed table. | 2020-06-03T19:28:59 | 2020-06-03T19:32:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,433,807 | 7,513 | SergeyLossev | @alexey-milovidov, So what about this bug/feature? | 2020-06-03T20:07:31 | 2020-06-03T20:07:31 | {} | NONE |
ClickHouse | ClickHouse | 638,641,570 | 11,356 | YiuRULE | @azat Alright, fixed normally every points who has been mentioned, we have now a special character supports, depending of the context (if it's an username, password, etc...) | 2020-06-04T06:52:05 | 2020-06-04T06:52:05 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 637,565,103 | 11,354 | alexey-milovidov | `test_ttl_move/test.py::test_concurrent_alter_with_ttl_move[concurrently_altering_ttl_mt-MergeTree()]`
- flaky test. | 2020-06-02T14:04:59 | 2020-06-02T14:04:59 | {} | MEMBER |
ClickHouse | ClickHouse | 637,898,321 | 157 | taythebot | > The library for User-Agent parsing was published in #10822
> But the task is in progress (now you can see just the exported sources of the library and it cannot be used).
Will this be available in the next release? | 2020-06-03T01:21:53 | 2020-06-03T01:21:53 | {} | NONE |
ClickHouse | ClickHouse | 637,952,359 | 11,396 | AlexanderChiuluvB | > Do you need data distributed among all nodes or duplicated to all nodes?
> For example you have 2 servers and inserted 10 rows. Do you need 5 rows at the first node and another 5 at the second node? Or do you need data duplicated? Both nodes have all 10 rows?
Both nodes have all 10 rows | 2020-06-03T04:56:49 | 2020-06-03T04:56:49 | {} | NONE |
ClickHouse | ClickHouse | 638,235,623 | 5,500 | kevinKaiF | > > Your advises really helped us.
>
> What did you change? And to what?
>
> min_merge_bytes_to_use_direct_io -- this parameter controls a size of merge which should use direct disk read/write (without using Linux cache). The idea is to save more cache for selects. But it can cause more I/O because during merges ... | 2020-06-03T14:28:55 | 2020-06-03T14:28:55 | {} | NONE |
ClickHouse | ClickHouse | 638,312,479 | 11,406 | thyn | Problem is resolved - bad table schema. Too many parts. Only one strange thing that it was written rarely. I didn't find it from first time | 2020-06-03T16:34:39 | 2020-06-03T16:34:39 | {} | NONE |
ClickHouse | ClickHouse | 638,451,224 | 11,417 | den-crane | In this case BTW you can use SimpleAggregateFunction
```
create table test_source_table
(
replay_date Date,
test_column Nullable(UInt32)
)
engine = MergeTree PARTITION BY toYYYYMM(replay_date)
ORDER BY (replay_date) SETTINGS index_granularity = 8192;
CREATE MATERIALIZED VIEW summary_test_source_tabl... | 2020-06-03T20:43:49 | 2020-06-03T20:43:49 | {
"+1": 2
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,812,325 | 11,385 | nikitamikhaylov | Performance tests show slightly speed-up :) | 2020-06-04T12:19:19 | 2020-06-04T12:19:19 | {} | MEMBER |
ClickHouse | ClickHouse | 637,604,058 | 11,384 | Slach | I want to explain why this feature is required and which corner cases it covers
- `system.metrics`, `system.events` and `system.asyncronous_metrics` can be monitored via prometheus
- prometheus has a `scraping interval`
- when "system.events" doesn't contain data with some critical events, like `DelayedInserts` or `... | 2020-06-02T15:07:57 | 2020-06-02T15:07:57 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,053,926 | 11,326 | weiyongyuan | Yestoday i met this problem again,the below log shows that the inserts and merges all don't work now!
-rw-r----- 1 root root 1 Jun 2 14:12 format_version.txt
drwxr-x--- 2 root root 4096 Jun 2 16:13 tmp_insert_441968_79020_79020_0
drwxr-x--- 2 root root 4096 Jun 2 16:13 tmp_insert_441968_79021_79021_0
drwxr-x-... | 2020-06-03T08:45:04 | 2020-06-03T08:45:04 | {} | NONE |
ClickHouse | ClickHouse | 638,212,620 | 11,398 | den-crane | What API do you use cpp / go ? | 2020-06-03T13:52:45 | 2020-06-03T13:52:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,637,938 | 11,396 | AlexanderChiuluvB | > @AlexanderChiuluvB Do you use Replicated Engine or plain MergeTree ?
> If you don't use Replicated you need to describe one more cluster with with one shard many replicas.
> If you use Replicated table then just omit {shard} macros in ZK table path and don't use Distributed table.
alright thanks~ | 2020-06-04T06:43:56 | 2020-06-04T06:43:56 | {} | NONE |
ClickHouse | ClickHouse | 638,731,979 | 11,424 | alesapin | Reproduced on master: https://gist.github.com/alesapin/7a816d3ea871476cbcf7ad19b98f6813 | 2020-06-04T09:25:17 | 2020-06-04T09:25:17 | {} | MEMBER |
ClickHouse | ClickHouse | 638,823,520 | 10,397 | muzammilSG | facing the same issue ... !! any updates...!! | 2020-06-04T12:43:33 | 2020-06-04T12:44:01 | {} | NONE |
ClickHouse | ClickHouse | 637,662,010 | 11,372 | KochetovNicolai | This case is integer overflow, because `SELECT number FROM numbers(100) ORDER BY number ASC LIMIT 2, 100000000 WITH TIES` works fine.
Don't know how to fix it properly. | 2020-06-02T16:22:29 | 2020-06-02T16:22:29 | {} | MEMBER |
ClickHouse | ClickHouse | 637,742,443 | 10,716 | kssenii | Continued in https://github.com/ClickHouse/ClickHouse/pull/11069. | 2020-06-02T18:54:38 | 2020-06-02T18:54:38 | {} | MEMBER |
ClickHouse | ClickHouse | 638,748,903 | 11,356 | azat | >Also maybe it will be better to add function to extract userinfo instead and netloc() will be concat(userinfo(url), domain(url)) (yep, will do extra job) ?
I guess that right now you need exactly that function and using `concat` will be slower? | 2020-06-04T09:57:21 | 2020-06-04T09:57:21 | {} | MEMBER |
ClickHouse | ClickHouse | 638,787,896 | 11,410 | akuzm | ```
[----------] 5 tests from TransformQueryForExternalDatabase
[ RUN ] TransformQueryForExternalDatabase.InWithSingleElement
unknown file: Failure
C++ exception with description "Database catalog is initialized twice. This is a bug.
" thrown in the test body.
```
https://clickhouse-test-reports.s3.yandex... | 2020-06-04T11:22:51 | 2020-06-04T11:22:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,723,758 | 11,388 | filimonov | It seems like `queued.min.messages=kafka_max_block_size` allows to archive very good performance.
Also it seem that the issue #11216 affects performance (when we drain librdkafka internal queue it start reporting 'stalled' status). | 2020-06-04T09:11:04 | 2020-06-04T12:59:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,866,115 | 11,216 | filimonov | The same happens if we empty librdkafka internal queue by fast polls. We need to distinguish those 2 situations. | 2020-06-04T13:57:13 | 2020-06-04T13:57:23 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,903,187 | 10,177 | shuttie | Experienced the same issue while upgrading from 20.1.4.14. We also did ALTER TABLE MODIFY TTL, but never dropped and recreated tables afterwards. | 2020-06-04T14:54:58 | 2020-06-04T14:54:58 | {} | NONE |
ClickHouse | ClickHouse | 638,935,286 | 6,336 | filimonov | Rechecked. That is not true currently.
The situation was improved in #11149 and #11388, but even before that with some tuning of the settings (mostly max_block_size) num_consumers were working properly.
**Results of 2 benchmarks:**
**from current master (before #11388, kafka_max_block_size=524288, topic with ... | 2020-06-04T15:43:04 | 2020-06-04T15:43:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 637,578,396 | 11,355 | alesapin | > And we have two more:
>
> ```
> test_insert_distributed_load_balancing/test.py::test_prefer_localhost_replica_0_load_balancing_nearest_hostname
> test_ttl_move/test.py::test_concurrent_alter_with_ttl_move[concurrently_altering_ttl_mt-MergeTree()]
> ```
lol :) | 2020-06-02T14:27:09 | 2020-06-02T14:27:09 | {} | MEMBER |
ClickHouse | ClickHouse | 637,905,034 | 157 | alexey-milovidov | Most likely not in the next release but in one of the releases this summer.
The code is available, it is working fine, but the pull request is not ready to be merged - we have to perform multiple cleanups to make it ready. | 2020-06-03T01:48:26 | 2020-06-03T01:48:26 | {} | MEMBER |
ClickHouse | ClickHouse | 638,019,022 | 11,395 | qoega | @alesapin probably you can answer | 2020-06-03T07:38:13 | 2020-06-03T07:38:13 | {} | MEMBER |
ClickHouse | ClickHouse | 638,034,405 | 11,396 | AlexanderChiuluvB | > If you want to set the same record on all of your server, why not just set a replicated table on each of your server ?
Because replica table can only be replicated between replicas of the same shard, but can not be replicated between diffetent shards.
I want that the small dimension table can be replicated bet... | 2020-06-03T08:06:51 | 2020-06-03T08:06:51 | {} | NONE |
ClickHouse | ClickHouse | 638,131,913 | 11,395 | alesapin | Yes, for normal non-replicated merge tree alters are always synchronous. Why this behavior is not expected? And what would you like to expect? | 2020-06-03T11:20:58 | 2020-06-03T11:21:47 | {} | MEMBER |
ClickHouse | ClickHouse | 638,169,520 | 11,234 | vitlibar | Integration test `test_grant_and_revoke` needs to be fixed because you've changed the data type in table `system.users`. | 2020-06-03T12:37:36 | 2020-06-03T12:37:36 | {} | MEMBER |
ClickHouse | ClickHouse | 638,207,716 | 10,212 | den-crane | @kterrita it's a new behavior for good. | 2020-06-03T13:45:38 | 2020-06-03T13:45:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,234,920 | 11,411 | alexey-milovidov | It can be simply translated to SELECT * FROM system.clusters WHERE ... | 2020-06-03T14:27:49 | 2020-06-03T14:27:49 | {} | MEMBER |
ClickHouse | ClickHouse | 638,404,477 | 9,361 | jangorecki | @filimonov would you mind claryfing what is the behavior of that? | 2020-06-03T19:12:16 | 2020-06-03T19:12:16 | {} | NONE |
ClickHouse | ClickHouse | 638,510,806 | 8,296 | Har01d | Indeed, that setting worked, I was trying to use it in the wrong config section (server.xml instead of users.xml). It’s not really clear why optimization settings are on the user level.
Thanks for your help! | 2020-06-03T23:21:02 | 2020-06-03T23:21:02 | {} | NONE |
ClickHouse | ClickHouse | 637,674,369 | 11,372 | akuzm | > This case is integer overflow, because `SELECT number FROM numbers(100) ORDER BY number ASC LIMIT 2, 100000000 WITH TIES` works fine.
> Don't know how to fix it properly.
Is there some place where signed int is used for row number? Maybe if we convert it all to UInt64, it will work? And if we have some other kind... | 2020-06-02T16:45:24 | 2020-06-02T16:45:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,879,872 | 8,836 | alesapin | Closed in favor of https://github.com/ClickHouse/ClickHouse/pull/11442 | 2020-06-04T14:20:05 | 2020-06-04T14:20:05 | {} | MEMBER |
ClickHouse | ClickHouse | 637,690,360 | 11,372 | KochetovNicolai | > Is there some place where signed int is used for row number?
Yes. It is not so easy for fix.
https://github.com/ClickHouse/ClickHouse/blob/06446b4f08a142d6f1bc30664c47ded88ab51782/src/Processors/LimitTransform.cpp#L283-L286 | 2020-06-02T17:16:45 | 2020-06-02T17:16:45 | {} | MEMBER |
ClickHouse | ClickHouse | 638,493,543 | 11,394 | alexey-milovidov | > Yandex third-party checks (only for Yandex employees)
It's internal issue in CI, actually everything is perfect... | 2020-06-03T22:26:40 | 2020-06-03T22:26:40 | {} | MEMBER |
ClickHouse | ClickHouse | 638,500,472 | 11,416 | alexey-milovidov | > 00754_alter_modify_order_by_replicated_zookeeper
Reported. | 2020-06-03T22:47:18 | 2020-06-03T22:47:18 | {} | MEMBER |
ClickHouse | ClickHouse | 638,641,581 | 11,412 | asurovenko | One of my node has empty replication_queue, second one has a lot of follow errors:
```
Code: 86, e.displayText() = DB::Exception: Received error from remote server ?endpoint=DataPartsExchange%3A%2Fclickhouse%2Ftables%2Fprod%2Fevents_dev%2Fconsent_aggregate%2Freplicas%2Fclickhouse-200&part=1591092000_1_1_0&client_prot... | 2020-06-04T06:52:06 | 2020-06-04T07:01:44 | {} | NONE |
ClickHouse | ClickHouse | 638,771,280 | 11,356 | YiuRULE | > > Also maybe it will be better to add function to extract userinfo instead and netloc() will be concat(userinfo(url), domain(url)) (yep, will do extra job) ?
>
> I guess that right now you need exactly that function and using `concat` will be slower?
On my side, it was was mostly for helping a bit the ClickHous... | 2020-06-04T10:43:53 | 2020-06-04T10:43:53 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 637,581,125 | 11,383 | qoega | I don't believe it will be easy :)
is it related #11317? | 2020-06-02T14:31:18 | 2020-06-02T14:31:18 | {} | MEMBER |
ClickHouse | ClickHouse | 637,956,022 | 11,394 | filimonov | Maybe also last exception (as in system dictionaries and replication queue)? | 2020-06-03T05:08:40 | 2020-06-03T05:08:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,077,380 | 9,623 | akuzm | Fixed here https://github.com/ClickHouse/ClickHouse/pull/10431 | 2020-06-03T09:25:41 | 2020-06-03T09:25:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,430,448 | 10,621 | KochetovNicolai | This pr also fixes #11413 | 2020-06-03T20:00:23 | 2020-06-03T20:00:23 | {} | MEMBER |
ClickHouse | ClickHouse | 637,566,467 | 11,354 | alexey-milovidov | Ok. I understand it in the following way:
- parts are 100% immutable;
- we don't need the lock if we read them, unless table is dropped;
- table is not dropped during fetches because fetches are cancelled on drop and we wait for them to finish. | 2020-06-02T14:07:30 | 2020-06-02T14:07:30 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 637,954,274 | 11,396 | AlexanderChiuluvB | > Do you need data distributed among all nodes or duplicated to all nodes?
> For example you have 2 servers and inserted 10 rows. Do you need 5 rows at the first node and another 5 at the second node? Or do you need data duplicated? Both nodes have all 10 rows?
>
> Both nodes have all 10 rows
Actually I want to ... | 2020-06-03T05:03:03 | 2020-06-03T05:09:29 | {} | NONE |
ClickHouse | ClickHouse | 638,173,249 | 10,212 | ibeleychev | 20.4.2.9
Error is still reproduced | 2020-06-03T12:45:16 | 2020-06-03T12:45:16 | {} | NONE |
ClickHouse | ClickHouse | 638,309,624 | 9,008 | NiyazNz | The fix works on the Memory engine and does not work on the MergeTree engine. Not tested other engines.
Clickhouse-server version: 20.4.3.16
```sql
test_default> select version()
[2020-06-03 19:25:47] 1 row retrieved starting from 1 in 44 ms (execution: 3 ms, fetching: 41 ms)
20.4.3.16
test_default> CREATE ... | 2020-06-03T16:29:47 | 2020-06-03T16:29:47 | {} | NONE |
ClickHouse | ClickHouse | 638,455,478 | 11,412 | asurovenko | Yes, configuration the same in both nodes. Both table schemas the same. And it's engine = ReplicatedMergeTree('/clickhouse/tables/{shard}/db/table_repl', '{replica}') PARTITION BY date ORDER BY date SETTINGS index_granularity = 8192;
What's right configuration? Is there any solution to continue populated tables in 01 ... | 2020-06-03T20:53:09 | 2020-06-03T20:53:09 | {} | NONE |
ClickHouse | ClickHouse | 638,490,959 | 11,420 | alexey-milovidov | I checked all other implementations of `IStorage::startup` methods and the only suspicious was for StorageKafka. | 2020-06-03T22:19:33 | 2020-06-03T22:19:33 | {} | MEMBER |
ClickHouse | ClickHouse | 638,887,963 | 11,443 | 4ertus2 | Related #11288 | 2020-06-04T14:33:31 | 2020-06-04T14:33:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 637,580,704 | 7,096 | akuzm | Part 2 is done in https://github.com/ClickHouse/ClickHouse/pull/11137 and further. We now use fmt for logging. | 2020-06-02T14:30:39 | 2020-06-02T14:30:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 637,593,539 | 11,382 | m0003r | Temporarily fixed my case via adding `UNION ALL 0, 0` to query to ensure at least 1 row in result. | 2020-06-02T14:50:52 | 2020-06-02T14:50:52 | {} | NONE |
ClickHouse | ClickHouse | 637,811,590 | 11,324 | alexey-milovidov | `test_ttl_move/test.py::test_concurrent_alter_with_ttl_move[concurrently_altering_ttl_mt-MergeTree()]`
- flaky test, fixed in #11355 | 2020-06-02T21:17:22 | 2020-06-02T21:17:22 | {} | MEMBER |
ClickHouse | ClickHouse | 638,149,809 | 11,296 | KochetovNicolai | @alexey-milovidov you should not add tags before bugfix is backported. Otherwise release script will think that you have already done it manually. | 2020-06-03T11:57:54 | 2020-06-03T11:57:54 | {} | MEMBER |
ClickHouse | ClickHouse | 638,273,889 | 7,795 | Akazz | I have reworked it somewhat and I am to push new changes within this week. | 2020-06-03T15:31:22 | 2020-06-03T15:31:22 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,530,542 | 11,409 | gubinjie | @den-crane The total amount of data in the table has basically not changed much | 2020-06-04T00:32:10 | 2020-06-04T00:32:10 | {} | NONE |
ClickHouse | ClickHouse | 638,751,489 | 11,412 | asurovenko | So, I updated my clickhouse-server version and replication started. Thanks. | 2020-06-04T10:02:07 | 2020-06-04T10:02:07 | {} | NONE |
ClickHouse | ClickHouse | 638,893,203 | 11,388 | filimonov | Benchmark results: consume speed (measured by https://github.com/filimonov/ch-kafka-consume-perftest):
That PR (v20.5.1.3632. 26d93fdfba21f504b9338184ebe1fd44920dae1c)
- all settings default: ~466000 rows/sec
- kafka_num_consumers=4: ~1243000 rows/sec
Base master commit:
- all settings default: ~180000 rows/se... | 2020-06-04T14:42:12 | 2020-06-04T14:48:39 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,409,228 | 11,412 | den-crane | ```
<internal_replication>false</internal_replication>
<replica>
<host>clickhouse-00</host>
<port>9000</port>
</replica>
<replica>
<host>clickhouse-01</host>
<port>9000</port>
... | 2020-06-03T19:21:24 | 2020-06-03T19:21:24 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,449,055 | 11,417 | den-crane | Oh, I think AggregateFunction still does not support Nulls
https://github.com/ClickHouse/ClickHouse/issues/3402 | 2020-06-03T20:39:09 | 2020-06-03T20:39:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,782,017 | 11,410 | akuzm | ```
00746_sql_fuzzy | FAIL | 59.52
```
Oh my, the fuzz test failed right away in debug. https://github.com/ClickHouse/ClickHouse/issues/11436
https://clickhouse-test-reports.s3.yandex.net/11410/1c33918f07796d172491ff3267f038e362f7b347/functional_stateless_tests_(debug).html#fail1
| 2020-06-04T11:08:59 | 2020-06-04T12:07:21 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 637,589,838 | 11,335 | alexey-milovidov | @qoega
The test
```
test_ttl_move/test.py::test_concurrent_alter_with_ttl_move[concurrently_altering_ttl_mt-MergeTree()]
```
should be reported. | 2020-06-02T14:45:02 | 2020-06-02T14:45:02 | {} | MEMBER |
ClickHouse | ClickHouse | 637,932,007 | 11,396 | den-crane | Do you need data distributed among all nodes or duplicated to all nodes?
For example you have 2 servers and inserted 10 rows. Do you need 5 rows at the first node and another 5 at the second node? Or do you need data duplicated? Both nodes have all 10 rows? | 2020-06-03T03:33:36 | 2020-06-03T03:34:16 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 638,050,459 | 11,134 | alesapin | `test_ttl_move/test.py::test_concurrent_alter_with_ttl_move[concurrently_altering_ttl_mt-MergeTree()]` known flaky test. | 2020-06-03T08:38:50 | 2020-06-03T08:38:50 | {} | MEMBER |
ClickHouse | ClickHouse | 638,124,192 | 7,795 | qoega | @CurtizJ @Akazz Do you have ideas how we can lose less performance? Do you plan to return to this issue? | 2020-06-03T11:05:36 | 2020-06-03T11:05:36 | {} | MEMBER |
ClickHouse | ClickHouse | 638,451,140 | 11,417 | amartin916 | hmm ya, I'm also seeing that, well that's unfortunate. Since that other ticket is open i'm going to close mine. Thanks! | 2020-06-03T20:43:37 | 2020-06-03T20:43:37 | {} | NONE |
ClickHouse | ClickHouse | 638,907,843 | 6,697 | 4ertus2 | Simple samples from #9861
```
SELECT dummy, name
FROM system.one, system.columns
WHERE (SELECT count() FROM system.columns WHERE name != '' AND dummy = 0) > 0
LIMIT 1;
# fail
SELECT o.dummy, name
FROM system.one o, system.columns
WHERE (SELECT count() FROM system.columns WHERE name != '' AND o.dummy = 0) > 0... | 2020-06-04T15:02:44 | 2020-06-04T15:02:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 653,558,201 | 11,913 | alexey-milovidov | There are plans to use ClickHouse with high-latency filesystem (e.g. S3). For that case the users may want to enable 100% compact parts. And if codecs won't work, it's a huge drawback. | 2020-07-03T13:50:37 | 2020-07-03T13:50:37 | {} | MEMBER |
ClickHouse | ClickHouse | 653,684,395 | 12,123 | alexey-milovidov | Because it cannot open a shared library for DNS lookup (part of glibc loaded at runtime). | 2020-07-03T21:52:33 | 2020-07-03T21:52:33 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.