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 | 782,569,866 | 20,991 | alexey-milovidov | > Fix #9969.
Let's add more detailed changelog entry. | 2021-02-20T06:17:39 | 2021-02-20T06:17:39 | {} | MEMBER |
ClickHouse | ClickHouse | 782,615,588 | 20,922 | tavplubix | ClickHouse build check - Build failed without build log. It's inner CI problem. | 2021-02-20T12:06:53 | 2021-02-20T12:06:53 | {} | MEMBER |
ClickHouse | ClickHouse | 782,618,949 | 19,673 | abyss7 | > @abyss7 any feedback on these patches?
I'm in process of reviewing this PR. | 2021-02-20T12:14:32 | 2021-02-20T12:14:32 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,810,965 | 21,015 | templarzq | yes,there are mem leaks if not all of the data are copied. | 2021-02-21T06:57:19 | 2021-02-21T06:57:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,813,897 | 20,479 | ucasfl | > Thank you! It can be slightly incomplete but good to merge.
No thanks! you are too kind!
BTW, which places need to be improved further? | 2021-02-21T07:29:01 | 2021-02-21T07:29:01 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 782,223,041 | 20,898 | alexey-milovidov | Builds failed due to CI failure. | 2021-02-19T17:31:54 | 2021-02-19T17:31:54 | {} | MEMBER |
ClickHouse | ClickHouse | 782,570,527 | 20,882 | azat | >AST fuzzer (debug) — Logical error: 'Trying to get name of not a column: SelectWithUnionQuery'.
#20342
>Marker check Pending — Starting checks (No special conditions apply)
Hanged
>Integration tests (release) — fail: 1, passed: 1123, error: 0
```
test_materialize_mysql_database/test.py::test_select_wit... | 2021-02-20T06:22:47 | 2021-02-20T06:22:47 | {} | MEMBER |
ClickHouse | ClickHouse | 782,571,924 | 20,994 | onine007 | > This is related to roaring bitmaps feature. Need help from @yuzhichang.
>
> By the way, from what version do you upgrade?
from version 20.8.12 | 2021-02-20T06:33:58 | 2021-02-20T06:33:58 | {} | NONE |
ClickHouse | ClickHouse | 782,704,223 | 21,008 | den-crane | > I don't use date to query and the number of stores is to much.
It does not matter. Just use `partition by toYYYYMM(toDate(createDate/1000))`
First of all partitioning is about insert --> create minimum parts during inserting and about data retention.
Partitioning can speed up queries but it's not mandatory.
| 2021-02-20T15:57:31 | 2021-02-20T15:57:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,801,783 | 20,991 | alexey-milovidov | Need to check if Brotli is using uninitialized memory legitimately. | 2021-02-21T05:23:15 | 2021-02-21T05:23:15 | {} | MEMBER |
ClickHouse | ClickHouse | 782,226,562 | 20,943 | alexey-milovidov | Not a bugfix.
Not merged completely. | 2021-02-19T17:38:15 | 2021-02-19T17:38:15 | {} | MEMBER |
ClickHouse | ClickHouse | 782,356,035 | 19,673 | azat | @abyss7 any feedback on these patches? | 2021-02-19T21:14:29 | 2021-02-19T21:14:29 | {} | MEMBER |
ClickHouse | ClickHouse | 782,458,203 | 20,625 | Taschenbergerm | Hi kitaisreal,
we just recently tried on our instance ( when we first encountered the documentation) But when we just gave the tuple argument without the second parameter - we did get a SQL error.
That was the motivation for the fork
Query: CREATE TABLE default.test_arg_max(
arg1 Int8,
date_time DateTime)... | 2021-02-19T23:40:37 | 2021-02-19T23:40:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,552,131 | 20,176 | imfht | @amosbird Thanks! It works! | 2021-02-20T03:42:09 | 2021-02-20T03:42:09 | {} | NONE |
ClickHouse | ClickHouse | 782,559,089 | 19,580 | tisonkun | @alesapin where does the original incompatibility come from? NuRaft itself is compatible with OSX. | 2021-02-20T04:40:34 | 2021-02-20T04:40:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,574,341 | 19,362 | alexey-milovidov | https://clickhouse-test-reports.s3.yandex.net/20989/7474a7e3ca139f1a4e88e83af011b304ebdcaf3c/fuzzer_ubsan/server.log | 2021-02-20T06:52:55 | 2021-02-20T06:52:55 | {} | MEMBER |
ClickHouse | ClickHouse | 782,586,708 | 20,174 | alexey-milovidov | @amosbird Also please run `utils/generate-ya-make/generate-ya-make`. | 2021-02-20T08:31:51 | 2021-02-20T08:31:51 | {} | MEMBER |
ClickHouse | ClickHouse | 782,601,377 | 19,516 | abyss7 | > Why creating copy not changing the poco itself? (we already have a list of own patches to poco).
>
> Not easy to see the diff with the current poco version.
>
> Is there some chances to push that into the poco upstream?
At lease now we can remove our unused patches from Poco - for simpler upstream update in ... | 2021-02-20T10:21:07 | 2021-02-20T10:21:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,829,142 | 21,015 | alexey-milovidov | The question is - how to improve our CI to detect these kind of issues automatically. | 2021-02-21T09:47:24 | 2021-02-21T09:47:24 | {} | MEMBER |
ClickHouse | ClickHouse | 782,211,580 | 20,553 | alexey-milovidov | 1. Let's check `EXPLAIN` output. It should use "read in order" query plan if `Timestamp` column is the first column in primary key. Otherwise large amount of memory will be used for sorting.
2. Maybe lowering block size will help: `SET max_block_size = 8192` or lower.
```
Is there a way to use a pipeline like the ... | 2021-02-19T17:12:19 | 2021-02-19T17:13:02 | {} | MEMBER |
ClickHouse | ClickHouse | 782,216,251 | 20,147 | alexey-milovidov | No idea. | 2021-02-19T17:20:10 | 2021-02-19T17:20:10 | {} | MEMBER |
ClickHouse | ClickHouse | 782,580,795 | 20,971 | alexey-milovidov | @filimonov Let's submit a patch. | 2021-02-20T07:45:00 | 2021-02-20T07:45:00 | {} | MEMBER |
ClickHouse | ClickHouse | 782,330,165 | 20,971 | filimonov | File can be damaged other way (w/o manual interventions).
Dosn't it break the documented invariant?
> If the system detects broken data parts (with the wrong size of files) or unrecognized parts (parts written to the file system but not recorded in ZooKeeper), it moves them to the `detached` subdirectory (they a... | 2021-02-19T20:36:14 | 2021-02-20T07:45:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,725,491 | 20,882 | azat | >@zhang2014 @BohuTANG Could you please take a look at the failure of MaterializeMySQL test?
In the meantime let's preserve mysql logs - #21016 | 2021-02-20T18:08:58 | 2021-02-20T18:08:58 | {} | MEMBER |
ClickHouse | ClickHouse | 782,214,955 | 20,607 | alexey-milovidov | Incomplete. | 2021-02-19T17:17:55 | 2021-02-19T17:17:55 | {} | MEMBER |
ClickHouse | ClickHouse | 782,227,361 | 20,917 | alexey-milovidov | Some builds failed with
> Build failed without build log. It's inner CI problem. Maybe something wrong with docker or host. Just wait for build restart. | 2021-02-19T17:39:36 | 2021-02-19T17:39:36 | {} | MEMBER |
ClickHouse | ClickHouse | 782,562,099 | 12,758 | darkleaf | How does codec work with an array? Codec is applied to two neighboring values in the same array, isn’t it? Or is it applied to values with the same index in to neighboring arrays? | 2021-02-20T05:09:57 | 2021-02-20T05:09:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,740,578 | 21,020 | alexey-milovidov | BTW query profiler can profile itself (for example Real time profiler can measure how frequently CPU time profiler was active). This is somewhat useful property. | 2021-02-20T19:56:07 | 2021-02-20T19:56:07 | {} | MEMBER |
ClickHouse | ClickHouse | 782,223,911 | 20,895 | alexey-milovidov | The issue only happens in misconfiguration. No need to backport. | 2021-02-19T17:33:26 | 2021-02-19T17:33:26 | {} | MEMBER |
ClickHouse | ClickHouse | 782,227,126 | 20,906 | alexey-milovidov | Not merged correctly. | 2021-02-19T17:39:14 | 2021-02-19T17:39:14 | {} | MEMBER |
ClickHouse | ClickHouse | 782,477,026 | 20,964 | den-crane | Upgrade all replicas to 20.5.5.74
Check that all replicas have the same CH version.
If error still happen then after the error "DB::Exception: Part xxxx (state Committed)" grep clickhouse-server.log for xxxxx
`grep 20210219_1152105_1152111_1 /var/log/clickhouse-server/clickhouse-server.log `
at all servers.
... | 2021-02-20T00:29:40 | 2021-02-20T00:29:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,580,599 | 20,625 | alexey-milovidov | @Taschenbergerm Maybe it's only in newer versions? | 2021-02-20T07:43:27 | 2021-02-20T07:43:27 | {} | MEMBER |
ClickHouse | ClickHouse | 782,669,247 | 20,553 | tephrocactus | Setting max_block_size to 8192 helps to reduce memory usage. Here is a query for 2 month period:
```
Row 1:
──────
is_initial_query: 1
user: default
query_id: d0e4e907-f2df-4d44-a8d5-5765aa7735d3
elapsed: 889.8091651
is_cancelled: 0
read_rows: 0... | 2021-02-20T13:56:36 | 2021-02-20T14:08:00 | {} | NONE |
ClickHouse | ClickHouse | 782,705,640 | 20,964 | pronvis | Ok, sorry.
### Partition grep
`k3`:
```
root@k3-video:/home/ansible# zgrep 20210220_1350044_1350049_1 /var/log/clickhouse-server/clickhouse-server.log*
/var/log/clickhouse-server/clickhouse-server.log.0.gz:2021.02.20 17:03:08.688374 [ 738 ] {} <Debug> fluentbit.log: Fetching part 20210220_1350044_1350049_1 fro... | 2021-02-20T16:02:08 | 2021-02-20T16:02:08 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,731,928 | 21,015 | alexey-milovidov | I have missed #19407, maybe @CurtizJ can help to review the fix? | 2021-02-20T18:50:50 | 2021-02-20T18:50:50 | {} | MEMBER |
ClickHouse | ClickHouse | 782,771,498 | 20,991 | kssenii | > https://clickhouse-test-reports.s3.yandex.net/20991/0d88366b2775bdcb60ae3eb18bc9fcb2ce7eef01/functional_stateless_tests_(memory)/stderr.log
For now I figured out that this msan failure can be fixed by changing compression level (and then the query finishes without any error returned), but about 3% of last rows get... | 2021-02-21T00:32:02 | 2021-02-21T00:34:07 | {} | MEMBER |
ClickHouse | ClickHouse | 782,217,763 | 20,866 | alexey-milovidov | Integration tests - spurious failure. | 2021-02-19T17:22:50 | 2021-02-19T17:22:50 | {} | MEMBER |
ClickHouse | ClickHouse | 782,278,832 | 16,481 | azat | >@azat Another uncaught exception from dtor. Please take a look. https://clickhouse-test-reports.s3.yandex.net/16481/5d150cce4778dd14f58dcff67435bdec1efa155b/stress_test_(thread).html#fail1
Here is another patch - #20981 | 2021-02-19T19:03:55 | 2021-02-19T19:03:55 | {} | MEMBER |
ClickHouse | ClickHouse | 782,577,732 | 20,490 | vdimir | @traceon And why using `system.query_log` is not suitable? | 2021-02-20T07:20:57 | 2021-02-20T07:20:57 | {} | MEMBER |
ClickHouse | ClickHouse | 782,579,920 | 19,516 | alexey-milovidov | No chance.
The whole purpose of this patch is to replace iostreams with Read/WriteBuffers. | 2021-02-20T07:38:23 | 2021-02-20T07:40:01 | {} | MEMBER |
ClickHouse | ClickHouse | 782,749,650 | 13,390 | vitlibar | I'm inclined to try another approach to this problem. Instead of creating a new nonstandard syntax `WITH GRANT OPTION(ALL EXCEPT admin, superuser)` I think I'll make the `GRANT` command keep information which user gave that permission and then `REVOKE` will remove the permission after checking the current user is still... | 2021-02-20T21:07:55 | 2021-02-20T21:45:36 | {} | MEMBER |
ClickHouse | ClickHouse | 782,823,457 | 21,015 | azat | >I will temporarily revert #19407 so we can make the next release without a fear.
Sounds reasonable (and just FYI that PR introduces pretty huge leak, if aggregate functions is in use, so this looks like a major problem) | 2021-02-21T09:03:21 | 2021-02-21T09:03:21 | {} | MEMBER |
ClickHouse | ClickHouse | 782,829,811 | 11,819 | leosunli | hi @alexey-milovidov
the version 20.8.12.2 has still this problem. This version has involved #11746. Could you help check this problem？ Thank you.
**Cmake version**
cmake version 3.19.3
**Ninja version**
1.8.2
**Compiler name and version**
9.3.0
**Full cmake and/or ninja output**
[1811/1887] Linking C... | 2021-02-21T09:52:08 | 2021-02-21T09:55:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,218,140 | 20,884 | alexey-milovidov | No UNION DISTINCT in this version. | 2021-02-19T17:23:30 | 2021-02-19T17:23:30 | {} | MEMBER |
ClickHouse | ClickHouse | 782,458,211 | 20,553 | tephrocactus | Thank you for an answer.
Here is the output of EXPLAIN for similar query against a distributed table:
```EXPLAIN
SELECT *
FROM events
WHERE (Timestamp >= 1613573815452) AND (Timestamp <= 1613576700442)
ORDER BY Timestamp DESC
LIMIT 250
Query id: b605856a-d07b-49ce-b544-4bf33c2eb278
┌─explain──────────... | 2021-02-19T23:40:38 | 2021-02-19T23:44:06 | {} | NONE |
ClickHouse | ClickHouse | 782,585,993 | 20,978 | alexey-milovidov | > 01017_uniqCombined_memory_usage
The query succeeded but the server error '241' was expected.
Looks like we also use less memory.
| 2021-02-20T08:26:53 | 2021-02-20T08:26:53 | {} | MEMBER |
ClickHouse | ClickHouse | 782,802,440 | 20,479 | ucasfl | For saving your time, I reopen this pr and update code with your patch^_^. @alexey-milovidov | 2021-02-21T05:26:47 | 2021-02-21T05:26:47 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 782,204,579 | 17,508 | alexey-milovidov | Continuation: #20972 | 2021-02-19T17:00:40 | 2021-02-19T17:00:40 | {} | MEMBER |
ClickHouse | ClickHouse | 782,215,282 | 20,644 | alexey-milovidov | Not fully merged. | 2021-02-19T17:18:30 | 2021-02-19T17:18:30 | {} | MEMBER |
ClickHouse | ClickHouse | 782,227,806 | 20,925 | alexey-milovidov | Minor. | 2021-02-19T17:40:22 | 2021-02-19T17:40:22 | {} | MEMBER |
ClickHouse | ClickHouse | 782,567,321 | 9,545 | amosbird | Do we have estimated number of overheads when enabling `metrics_perf_events_enabled`? | 2021-02-20T05:58:55 | 2021-02-20T05:58:55 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 782,695,102 | 20,964 | den-crane | >clickhouse-server.err.log
err.log has ZERO information about part creation.
I asked clickhouse-server.log
| 2021-02-20T15:04:43 | 2021-02-20T15:04:43 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,791,933 | 20,882 | zhang2014 | > @zhang2014 @BohuTANG Could you please take a look at the failure of MaterializeMySQL test?
>> In the meantime let's preserve mysql logs - #21016
It looks like `SYSTEM STOP MERGE` was not implemented. I'm trying to fix it. | 2021-02-21T03:56:17 | 2021-02-21T03:56:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,817,714 | 21,022 | alexey-milovidov | Where it was broken? | 2021-02-21T08:08:04 | 2021-02-21T08:08:04 | {} | MEMBER |
ClickHouse | ClickHouse | 782,520,416 | 20,954 | talent2333 | > There is data for that table on default disk. And default disk is not a part of used policy.
>
> You need either to move that data (manually) to one of disks which are the part of that policy, either add the default disk to that policy.
Thanks your reply.But the path of default disk is same to the path of disk_... | 2021-02-20T01:56:57 | 2021-02-20T01:56:57 | {} | NONE |
ClickHouse | ClickHouse | 782,576,988 | 20,254 | siradjev | > > It seems there is a need for ruleset to identify "proper stage" correctly.
>
> Only the outmost `group by` will trigger WithMergeableState. All subqueries only deal with local tables (proxy).
This may lead to excessive data redistribution/east-west network traffic in cases, when inner "group by" generate high... | 2021-02-20T07:13:53 | 2021-02-20T07:37:06 | {} | NONE |
ClickHouse | ClickHouse | 782,581,032 | 20,971 | alexey-milovidov | > File can be damaged other way (w/o manual interventions).
If table is replicated - the part will be repaired from replica.
If table is not replicated - the damaged part cannot be repaired automatically if there are no covered parts, and it's in doubt you can do anything. | 2021-02-20T07:46:53 | 2021-02-20T07:47:29 | {} | MEMBER |
ClickHouse | ClickHouse | 782,591,531 | 8,441 | alexey-milovidov | After #19516 the issue is fixed completely:
Master:
```
$ for i in {1..100}; do time curl "http://localhost:8123/?input_format_parallel_parsing=0&query=INSERT%20INTO%20log%20format%20CSV" --data-binary @test.csv; done
real 0m1,621s
user 0m0,169s
sys 0m1,079s
real 0m1,590s
user 0m0,154s
... | 2021-02-20T09:04:30 | 2021-02-20T09:04:30 | {} | MEMBER |
ClickHouse | ClickHouse | 782,214,809 | 20,606 | alexey-milovidov | Incomplete. | 2021-02-19T17:17:41 | 2021-02-19T17:17:41 | {} | MEMBER |
ClickHouse | ClickHouse | 782,220,524 | 17,761 | den-crane | hmm,
20.11.5.18
Received exception from server (version 20.11.5):
Code: 50. DB::Exception: Received from localhost:9000. DB::Exception: Unknown type Decimal(20, 6).
20.12.6.29
Received exception from server (version 20.12.6):
Code: 50. DB::Exception: Received from localhost:9000. DB::Exception: Unknown type D... | 2021-02-19T17:27:43 | 2021-02-19T17:38:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,228,148 | 20,952 | alexey-milovidov | Merged incorrectly. | 2021-02-19T17:40:58 | 2021-02-19T17:40:58 | {} | MEMBER |
ClickHouse | ClickHouse | 782,235,064 | 20,608 | Exus | > Maybe you can download clickhouse with ASAN, that will likely provide more details for such memory corruptions...
I've installed clickhouse-*_21.2.4.18+asan_amd64.deb
> Can you compare with 'baremetal' setup?
Yes, I've installed ch on them too
So far - no crashes
> What CH version do you use with ... | 2021-02-19T17:52:04 | 2021-02-19T17:52:04 | {} | NONE |
ClickHouse | ClickHouse | 782,559,178 | 16,639 | ucasfl | This task has been done! @alexey-milovidov | 2021-02-20T04:41:21 | 2021-02-20T04:41:21 | {
"hooray": 1
} | COLLABORATOR |
ClickHouse | ClickHouse | 782,566,981 | 20,553 | alexey-milovidov | Ok. The query plan is correct.
Now let's go to
> 2. Maybe lowering block size will help: SET max_block_size = 8192 or lower.
Maybe there are very large values in the table? | 2021-02-20T05:55:39 | 2021-02-20T05:55:39 | {} | MEMBER |
ClickHouse | ClickHouse | 782,570,613 | 12,758 | alexey-milovidov | > Codec is applied to two neighboring values in the same array, isn’t it?
Almost this. It is applied to the sequence of all elements of arrays if placed one after the other.
For example, if we have two rows with [1, 2, 3], [1, 2], the codec will be applied to the sequence 1, 2, 3, 1, 2. | 2021-02-20T06:23:24 | 2021-02-20T06:23:24 | {
"+1": 4
} | MEMBER |
ClickHouse | ClickHouse | 782,579,098 | 19,516 | filimonov | Why creating copy not changing the poco itself? (we already have a list of own patches to poco).
Not easy to see the diff with the current poco version.
Is there some chances to push that into the poco upstream? | 2021-02-20T07:32:35 | 2021-02-20T07:32:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,585,466 | 20,978 | alexey-milovidov | 0.3% improvement in average across all existing perf tests :) | 2021-02-20T08:23:08 | 2021-02-20T08:23:08 | {} | MEMBER |
ClickHouse | ClickHouse | 782,590,600 | 20,991 | alexey-milovidov | https://clickhouse-test-reports.s3.yandex.net/20991/0d88366b2775bdcb60ae3eb18bc9fcb2ce7eef01/functional_stateless_tests_(memory)/stderr.log | 2021-02-20T08:57:28 | 2021-02-20T08:57:28 | {} | MEMBER |
ClickHouse | ClickHouse | 782,591,700 | 19,516 | alexey-milovidov | @filimonov See https://github.com/ClickHouse/ClickHouse/issues/8441#issuecomment-782591531 | 2021-02-20T09:06:03 | 2021-02-20T09:06:03 | {} | MEMBER |
ClickHouse | ClickHouse | 782,651,508 | 20,946 | tavplubix | Stress test (address) - #21004
Stress test (thread) - something strange happened... | 2021-02-20T13:25:44 | 2021-02-20T13:25:44 | {} | MEMBER |
ClickHouse | ClickHouse | 782,806,389 | 20,479 | alexey-milovidov | Thank you! It can be slightly incomplete but good to merge. | 2021-02-21T06:06:36 | 2021-02-21T06:06:36 | {} | MEMBER |
ClickHouse | ClickHouse | 782,818,093 | 21,022 | azat | >Where it was broken?
I did not look deeply, but I guess the reason is that status file was preserved after `while killall clickhouse-server; do :; done` right now, while before it was removed | 2021-02-21T08:11:38 | 2021-02-21T08:11:38 | {} | MEMBER |
ClickHouse | ClickHouse | 782,548,613 | 20,994 | onine007 | CREATE TABLE my_sdap.crowd_bitmap_result_replicated
(
`instance_id` UInt32,
`crowd_id` UInt32,
`status` UInt8 DEFAULT 1,
`dt` UInt32,
`measure` AggregateFunction(groupBitmap, UInt32),
`created_at` DateTime DEFAULT now(),
`crowd_cardinality` UInt32 MATERIALIZED ifNull(bitmapCardinalit... | 2021-02-20T03:18:58 | 2021-02-20T03:24:29 | {} | NONE |
ClickHouse | ClickHouse | 782,577,318 | 20,585 | alesapin | Tests look OK, but for sure I'll make several runs (and cleanup code a little bit). | 2021-02-20T07:16:57 | 2021-02-20T07:16:57 | {} | MEMBER |
ClickHouse | ClickHouse | 782,580,326 | 20,254 | siradjev | > There is no data redistribution at all. The semantic of distributed table in subqueries becomes different -- it's a local table and will only return data on the local shard.
Do you mean it would be some sort of "distributed_group_by_no_merge"-like behavior on inner levels? Sounds reasonable too. | 2021-02-20T07:41:39 | 2021-02-20T07:41:39 | {} | NONE |
ClickHouse | ClickHouse | 782,582,036 | 20,254 | amosbird | > Do you mean it would be some sort of "distributed_group_by_no_merge"-like behavior on inner levels?
yes | 2021-02-20T07:54:28 | 2021-02-20T07:54:28 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 782,680,330 | 20,964 | pronvis | ### Upgrade
All 3 replicas have same version: `20.5.5.74`
### Partition grep
`k3`:
```
root@k3-video:/home/ansible# grep 20210220_1350044_1350049_1 /var/log/clickhouse-server/clickhouse-server.err.log
root@k3-video:/home/ansible#
```
`k5`:
```
root@k5-video:/home/ansible# grep 20210220_1350044_1350049_... | 2021-02-20T14:12:58 | 2021-02-20T14:14:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,806,460 | 20,599 | alexey-milovidov | Need to figure out why some queries slowed down (from 0 ms to 1 ms). | 2021-02-21T06:07:22 | 2021-02-21T06:07:22 | {} | MEMBER |
ClickHouse | ClickHouse | 782,829,718 | 21,015 | alexey-milovidov | #21024 | 2021-02-21T09:51:32 | 2021-02-21T09:51:32 | {} | MEMBER |
ClickHouse | ClickHouse | 782,567,274 | 20,994 | alexey-milovidov | This is related to roaring bitmaps feature. Need help from @yuzhichang.
By the way, from what version do you upgrade? | 2021-02-20T05:58:28 | 2021-02-20T05:58:28 | {} | MEMBER |
ClickHouse | ClickHouse | 782,570,376 | 12,758 | alexey-milovidov | See https://github.com/ClickHouse/ClickHouse/issues/12551 and https://github.com/ClickHouse/ClickHouse/pull/15089 | 2021-02-20T06:21:42 | 2021-02-20T06:21:42 | {} | MEMBER |
ClickHouse | ClickHouse | 782,585,347 | 20,978 | alexey-milovidov | It makes sense: 10% improvement of `group_by_fixed_keys` test. | 2021-02-20T08:22:24 | 2021-02-20T08:22:24 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 782,629,750 | 20,625 | kitaisreal | @Taschenbergerm, support for tuple as single argument was added in newer versions in this pull request https://github.com/ClickHouse/ClickHouse/pull/17359. Documentation is fine.
CC: @alexey-milovidov | 2021-02-20T12:29:50 | 2021-02-20T12:29:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,582,428 | 20,980 | alexey-milovidov | > Stress test (thread) — Killed by signal (in clickhouse-server.log)
#20981 | 2021-02-20T07:57:47 | 2021-02-20T07:57:47 | {} | MEMBER |
ClickHouse | ClickHouse | 782,713,913 | 20,964 | den-crane | не хватает все равно логов, надо все файлы целиком.
я думаю есть какой-то рейс, сервера видят новые парты, но не видят задания на мерж этих партов.
возможно проблема из-за того что зукиперы тоже на этих хостах и поэтому тормозят.
k7 смержил парт ( все _0 парты тоже на нем )
k5 увидел новый парт _1 и скачал
k5 ... | 2021-02-20T16:52:28 | 2021-02-20T16:52:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,813,614 | 21,015 | alexey-milovidov | I will temporarily revert #19407 so we can make the next release without a fear. | 2021-02-21T07:25:40 | 2021-02-21T07:25:40 | {} | MEMBER |
ClickHouse | ClickHouse | 782,218,184 | 20,885 | alexey-milovidov | No UNION DISTINCT in this version. | 2021-02-19T17:23:35 | 2021-02-19T17:23:35 | {} | MEMBER |
ClickHouse | ClickHouse | 782,224,889 | 20,916 | alexey-milovidov | Not merged completely. | 2021-02-19T17:35:12 | 2021-02-19T17:35:12 | {} | MEMBER |
ClickHouse | ClickHouse | 782,574,379 | 20,989 | alexey-milovidov | https://github.com/ClickHouse/ClickHouse/issues/19362 | 2021-02-20T06:53:16 | 2021-02-20T06:53:16 | {} | MEMBER |
ClickHouse | ClickHouse | 782,575,769 | 20,882 | alexey-milovidov | @zhang2014 @BohuTANG Could you please take a look at the failure of MaterializeMySQL test? | 2021-02-20T07:03:21 | 2021-02-20T07:03:21 | {} | MEMBER |
ClickHouse | ClickHouse | 782,579,128 | 20,254 | amosbird | There is no data redistribution at all. The semantic of distributed table in subqueries becomes different -- it's a local table and will only return data on the local shard. | 2021-02-20T07:32:48 | 2021-02-20T07:32:48 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 782,580,063 | 19,516 | alexey-milovidov | Also we are looking into perspective to also rewrite TCPServer. | 2021-02-20T07:39:28 | 2021-02-20T07:39:28 | {} | MEMBER |
ClickHouse | ClickHouse | 782,610,561 | 20,500 | adiletkabylbekov | @filimonov so as i expected there is no problem with query structure or data structure.
I found same issue on `ALTER` for table with `ReplicatedMergeTree()` engine (not MV),
So assume we have table with `Enum` column type:
```
CREATE TABLE akabylbekov.enum_alter_main
(
`date` Date,
`enum_type` Enum8('n... | 2021-02-20T11:30:02 | 2021-02-20T11:33:22 | {} | NONE |
ClickHouse | ClickHouse | 782,642,351 | 16,193 | tavplubix | Functional stateless tests (release, DatabaseReplicated) — fail: 0, passed: 2595, skipped: 203
Functional stateful tests (release, DatabaseReplicated) — fail: 0, passed: 109, skipped: 1
Integration tests - OK | 2021-02-20T12:56:44 | 2021-02-20T12:56:44 | {} | MEMBER |
ClickHouse | ClickHouse | 782,225,636 | 20,908 | alexey-milovidov | Some people argued that this bugfix broke their scenarios. | 2021-02-19T17:36:29 | 2021-02-19T17:36:29 | {} | MEMBER |
ClickHouse | ClickHouse | 782,254,121 | 20,208 | traceon | @alexey-milovidov Well, that never was a secret. :)
Here is the documentation draft, that (always) reflects the current (potentially unmerged yet) functionality around LDAP and Kerberos: [External authentication mechanism support in ClickHouse ](https://gist.github.com/traceon/d922c8452e8065b86667168578dc5c12) | 2021-02-19T18:24:32 | 2021-02-19T18:24:32 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 782,588,496 | 20,978 | alexey-milovidov | It because operator new is not counted in builds with sanitizers. | 2021-02-20T08:42:57 | 2021-02-20T08:42:57 | {} | MEMBER |
ClickHouse | ClickHouse | 785,805,388 | 21,041 | OmarBazaraa | After some investigation, we found that the problem occurs when trying to connect to MongoDB using SSL connection. | 2021-02-25T10:52:12 | 2021-02-25T10:52:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 785,893,101 | 21,186 | den-crane | CH does not support table aliases in views yet.
```
Instead of : v.imei,v.visitstartdate,v.visitenddate FROM dbpolaris1.visit AS v
use : dbpolaris1.visit.imei,dbpolaris1.visit.visitstartdate,dbpolaris1.visit.visitenddate FROM dbpolaris1.visit
``` | 2021-02-25T13:25:56 | 2021-02-25T13:26:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 786,899,325 | 21,079 | azat | >Integration tests flaky check (asan) — Flaky tests found: fail: 1, passed: 5, error: 1
Should be addressed in #21264 (after nightly stress testing I cannot reproduce the failure anymore after applying patch from the referenced PR) | 2021-02-26T21:19:13 | 2021-02-27T07:42:38 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.