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 | 713,898,005 | 16,043 | alexey-milovidov | Self-signed certificates work Ok for clickhouse-client if it's configured properly:
```
$ cat ~/.clickhouse-client/config.xml
<config>
<openSSL>
<client>
<loadDefaultCAFile>true</loadDefaultCAFile>
<caConfig>/usr/local/share/ca-certificates/Yandex/YandexInternalRootCA.crt</caConfig>
... | 2020-10-21T21:51:19 | 2020-10-21T22:00:35 | {} | MEMBER |
ClickHouse | ClickHouse | 713,978,264 | 12,257 | alexey-milovidov | @filimonov You can write a test for this issue to prevent it happening again.
It's rather easy, see `01507_clickhouse_server_start_with_embedded_config.sh`. | 2020-10-21T23:30:15 | 2020-10-21T23:30:15 | {} | MEMBER |
ClickHouse | ClickHouse | 713,526,523 | 15,592 | alexey-milovidov | @filimonov Does not look like. | 2020-10-21T12:15:57 | 2020-10-21T12:15:57 | {} | MEMBER |
ClickHouse | ClickHouse | 713,584,402 | 8,643 | SaltTan | Related #10278 | 2020-10-21T13:41:02 | 2020-10-21T13:41:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,651,263 | 16,237 | moweonlee | Thank you always @den-crane.
That is exactly what I need to know.
It works very fine. | 2020-10-21T15:15:47 | 2020-10-21T15:15:47 | {} | NONE |
ClickHouse | ClickHouse | 713,930,892 | 14,736 | alexey-milovidov | @den-crane Don't think it was by design. | 2020-10-21T22:59:19 | 2020-10-21T22:59:19 | {} | MEMBER |
ClickHouse | ClickHouse | 713,539,391 | 10,417 | alexey-milovidov | Fix: https://github.com/ClickHouse/ClickHouse/pull/15646 | 2020-10-21T12:39:31 | 2020-10-21T12:39:31 | {
"heart": 1
} | MEMBER |
ClickHouse | ClickHouse | 713,554,346 | 12,450 | alesapin | I'll close this issue, if you encounter this error again fill free to reopen. | 2020-10-21T13:05:09 | 2020-10-21T13:05:09 | {} | MEMBER |
ClickHouse | ClickHouse | 713,574,989 | 16,224 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/4831 | 2020-10-21T13:29:56 | 2020-10-21T13:29:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,878,657 | 16,245 | alexey-milovidov | Duplicate #13281 | 2020-10-21T21:08:49 | 2020-10-21T21:08:49 | {} | MEMBER |
ClickHouse | ClickHouse | 713,978,975 | 12,143 | alexey-milovidov | Hold until #15174 | 2020-10-21T23:31:04 | 2020-10-21T23:31:04 | {} | MEMBER |
ClickHouse | ClickHouse | 713,475,776 | 12,169 | filimonov | https://github.com/ClickHouse/ClickHouse/pull/11296 | 2020-10-21T10:36:09 | 2020-10-21T10:36:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,479,606 | 16,132 | zhang2014 | > Integration tests (asan) — Some exception occurred :(
> Integration tests (thread) — Some exception occurred :(
Unknown reason. | 2020-10-21T10:43:41 | 2020-10-21T10:43:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,522,737 | 12,550 | abyss7 | Probably added regression for queries like:
`SELECT * FROM table WHERE has(['a', 'b'], low_cardinality_string_column)`
With exception:
> Types of array and 2nd argument of function "has" must be identical up to nullability, cardinality, numeric types, or Enum and numeric type. Passed: Array(String) and LowCard... | 2020-10-21T12:09:08 | 2020-10-21T12:14:53 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,829,358 | 16,189 | tavplubix | AST fuzzer - reported #16245
Stress test (thread) — Sanitizer assert - fixed #16244 | 2020-10-21T19:38:14 | 2020-10-21T19:38:14 | {} | MEMBER |
ClickHouse | ClickHouse | 713,848,158 | 15,938 | CurtizJ | You can find examples of performance tests in tests/performance. Test is a .xml file with bunch of queries for measuring time of their executing and possible some additional queries for preparation. Also you can set specific settings for tests. A good [example](https://github.com/ClickHouse/ClickHouse/blob/master/tests... | 2020-10-21T20:15:54 | 2020-10-21T20:25:15 | {} | MEMBER |
ClickHouse | ClickHouse | 713,923,503 | 15,359 | alexey-milovidov | Yes, it will have the same performance as queries with FINAL. | 2020-10-21T22:53:09 | 2020-10-21T22:53:09 | {} | MEMBER |
ClickHouse | ClickHouse | 713,514,766 | 16,192 | alexey-milovidov | > Just a side note - it also seems that clickhouse-local can start merging parts of the attached tables on its own, so it may be potentially unsafe to use it to introspect the data of the offline server.
Discussed in Telegram that it's Ok to keep this behaviour. | 2020-10-21T11:54:03 | 2020-10-21T11:54:03 | {} | MEMBER |
ClickHouse | ClickHouse | 713,574,876 | 16,224 | den-crane | I think this parameter is applied to CH server, when CH server is a client itself and executes a query against another server.
And it's not obvious how client should/can use this. Because if ch-client pulls value of this parameter from a server and use it then CH server also have to pull this when it plays as a clie... | 2020-10-21T13:29:44 | 2020-10-21T13:29:44 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,903,196 | 16,220 | SaltTan | Reminds me of #13331 | 2020-10-21T22:03:42 | 2020-10-21T22:03:42 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,910,502 | 15,806 | CurtizJ | Yes, I will try to review your PR by the end of the week.
BTW, please, add tests with other table engines, at least with `MergeTree`, because storage `Memory` doesn't use serialization/deserialization from `IDataType` and it's not checked. I've created a `MergeTree` table with type `Map` and got segfault on insert.
... | 2020-10-21T22:21:41 | 2020-10-21T22:21:41 | {} | MEMBER |
ClickHouse | ClickHouse | 713,922,632 | 15,414 | alexey-milovidov | The error "Too many simultaneous queries" are happening for one of the following reasons:
1. Hardware errors: replica remain accessible but suddenly was slowdown.
2. Software errors: lack of rate limiting, too heavy queries, a scipt with a loop...
Reason 2 is more common.
The proposed solution will help for c... | 2020-10-21T22:50:21 | 2020-10-21T22:50:21 | {} | MEMBER |
ClickHouse | ClickHouse | 713,970,929 | 12,408 | alexey-milovidov | Fixed in version 20.10-prestable by enabling compact data parts. | 2020-10-21T23:25:32 | 2020-10-21T23:25:32 | {} | MEMBER |
ClickHouse | ClickHouse | 713,504,795 | 16,206 | alexey-milovidov | Integration test `test_multiple_disks` - bad test, @excitoon | 2020-10-21T11:33:59 | 2020-10-21T11:33:59 | {} | MEMBER |
ClickHouse | ClickHouse | 713,511,596 | 16,004 | alexey-milovidov | It's appeared to be important bugfix. | 2020-10-21T11:47:55 | 2020-10-21T11:47:55 | {} | MEMBER |
ClickHouse | ClickHouse | 713,526,127 | 16,221 | alexey-milovidov | We already have it - `timeout_before_checking_execution_speed`. | 2020-10-21T12:15:15 | 2020-10-21T12:15:15 | {} | MEMBER |
ClickHouse | ClickHouse | 713,541,816 | 12,550 | alex-zaitsev | Fixed in https://github.com/ClickHouse/ClickHouse/pull/16038 ?
@myrrc, could you check?
> Probably added regression for queries like:
>
> `SELECT * FROM table WHERE has(['a', 'b'], low_cardinality_string_column)`
| 2020-10-21T12:43:58 | 2020-10-21T12:43:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,706,063 | 16,208 | den-crane |
on_cluster -- is just a helper. It executes this CREATE MAT.... at all nodes.
then CREATE MAT.... POPULATE is executed at node N1 it does POPULATE over all data
then CREATE MAT.... POPULATE is executed at node N2 (which is replica for N1) and it does POPULATE over all the same data as N1
and you get all duplic... | 2020-10-21T16:39:38 | 2020-10-21T16:39:38 | {
"heart": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,919,782 | 15,935 | alexey-milovidov | Fixed. | 2020-10-21T22:42:17 | 2020-10-21T22:42:17 | {} | MEMBER |
ClickHouse | ClickHouse | 713,986,009 | 12,068 | alexey-milovidov | @BohuTANG looks like easy to fix. Let's do it
(it's enough to ping in `readImpl`, not before every value) | 2020-10-21T23:35:37 | 2020-10-21T23:35:37 | {} | MEMBER |
ClickHouse | ClickHouse | 713,989,454 | 12,068 | alexey-milovidov | Or maybe it's already fixed by https://github.com/ClickHouse/ClickHouse/pull/15299? | 2020-10-21T23:37:41 | 2020-10-21T23:37:41 | {} | MEMBER |
ClickHouse | ClickHouse | 713,482,176 | 10,417 | alijared | Why can't the region be parsed as a parameter to the table function? | 2020-10-21T10:48:56 | 2020-10-21T10:48:56 | {} | NONE |
ClickHouse | ClickHouse | 713,532,415 | 16,215 | alesapin | > Didn't understand the changes.
Just fewer ties in each concurrent query. It was flaky since creation. | 2020-10-21T12:26:58 | 2020-10-21T12:26:58 | {} | MEMBER |
ClickHouse | ClickHouse | 713,537,252 | 15,592 | sundy-li | > Does it also solve #16045 ?
No, `restarting_thread` is not started when the table is read-only during startup.
To solve that we should also have the metric code to `StorageReplicatedMergeTree.cpp`. | 2020-10-21T12:35:48 | 2020-10-21T12:35:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,892,607 | 16,190 | filimonov | Question - does clickhouse relies somehow on locales? How can it be tested? Any other potential issues which we should recheck? | 2020-10-21T21:39:03 | 2020-10-21T21:39:03 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,918,622 | 15,998 | alexey-milovidov | > Maybe it's worth having that as a flag in Kafka table settings too? Smth like 'kafka_allow_selecting_data'?
Yes, it's reasonable, let's implement it. | 2020-10-21T22:39:56 | 2020-10-21T22:39:56 | {} | MEMBER |
ClickHouse | ClickHouse | 713,949,090 | 13,057 | alexey-milovidov | @vitlibar This is a small tech debt, need to fix. | 2020-10-21T23:12:05 | 2020-10-21T23:12:05 | {} | MEMBER |
ClickHouse | ClickHouse | 713,575,650 | 16,224 | den-crane | for a client it should be set at a client
clickhouse-client --receive_timeout=6000 | 2020-10-21T13:30:49 | 2020-10-21T13:30:49 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,577,314 | 16,222 | den-crane | duplicate https://github.com/ClickHouse/ClickHouse/issues/15784 | 2020-10-21T13:32:41 | 2020-10-21T13:32:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,627,792 | 16,238 | alexey-milovidov | It's the best place where we can also add Thread Fuzzer... | 2020-10-21T14:42:16 | 2020-10-21T14:42:16 | {} | MEMBER |
ClickHouse | ClickHouse | 713,882,438 | 16,199 | azat | >We don't run compatibility check for unbundled build.
Yep, I know, but unbundled build should not have these requirements anyway, no? | 2020-10-21T21:16:23 | 2020-10-21T21:16:23 | {} | MEMBER |
ClickHouse | ClickHouse | 713,930,702 | 14,736 | alexey-milovidov | Fixed in https://github.com/ClickHouse/ClickHouse/pull/15646 | 2020-10-21T22:58:58 | 2020-10-21T22:58:58 | {} | MEMBER |
ClickHouse | ClickHouse | 713,939,693 | 14,018 | alexey-milovidov | We have cleaned up some places where DNS response can be cached for longer time than needed.
Does the issue reproduce with version 20.9? | 2020-10-21T23:06:18 | 2020-10-21T23:06:27 | {} | MEMBER |
ClickHouse | ClickHouse | 713,741,135 | 15,938 | Avogar | > It will be useful to add a perf test and also to check the difference in performance with case, when all partitions are merged.
@CurtizJ I've never written perf tests before, can you explain what should I do in performance test? Just make a query with `SELECT ... FINAL`? And how to check the difference in performa... | 2020-10-21T17:39:44 | 2020-10-21T17:39:44 | {} | MEMBER |
ClickHouse | ClickHouse | 713,834,614 | 16,238 | alesapin | What's wrong with tests? | 2020-10-21T19:49:10 | 2020-10-21T19:49:10 | {} | MEMBER |
ClickHouse | ClickHouse | 713,877,740 | 16,199 | alexey-milovidov | We don't run compatibility check for unbundled build. | 2020-10-21T21:06:51 | 2020-10-21T21:06:51 | {} | MEMBER |
ClickHouse | ClickHouse | 713,892,528 | 16,199 | alexey-milovidov | Yes, unbundled build is incompatible by design. | 2020-10-21T21:38:50 | 2020-10-21T21:38:50 | {} | MEMBER |
ClickHouse | ClickHouse | 713,915,366 | 16,221 | alexey-milovidov | Currently hold because the implementation is non-trivial. | 2020-10-21T22:34:42 | 2020-10-21T22:34:42 | {} | MEMBER |
ClickHouse | ClickHouse | 713,916,396 | 16,037 | alexey-milovidov | > Does that mean at least for some time we might see duplicates?
No, there is a lock (ephemeral node that will prevent insertion of duplicates and prevent merging with the new part) that is acquired at step 2 and released at step 4. | 2020-10-21T22:37:27 | 2020-10-21T22:37:27 | {} | MEMBER |
ClickHouse | ClickHouse | 713,953,923 | 13,012 | alexey-milovidov | @vitlibar @nvartolomei
How about to always allow to change settings within constraints?
So, negative constraints will be set up as usual
and for positive constraints you just specify `<readonly>1</readonly>`
and then specify your constraints to allow something. | 2020-10-21T23:14:57 | 2020-10-21T23:14:57 | {} | MEMBER |
ClickHouse | ClickHouse | 713,954,735 | 13,012 | alexey-milovidov | But it will prevent doing non-readonly queries. | 2020-10-21T23:15:29 | 2020-10-21T23:15:29 | {} | MEMBER |
ClickHouse | ClickHouse | 713,511,582 | 16,001 | alexey-milovidov | It's appeared to be important bugfix. | 2020-10-21T11:47:52 | 2020-10-21T11:47:52 | {} | MEMBER |
ClickHouse | ClickHouse | 713,512,345 | 16,004 | alexey-milovidov | Don't backport due to conflicts. | 2020-10-21T11:49:19 | 2020-10-21T11:49:19 | {} | MEMBER |
ClickHouse | ClickHouse | 713,560,665 | 15,646 | tooptoop4 | @excitoon are IAM role supported for private buckets? | 2020-10-21T13:15:10 | 2020-10-21T13:15:10 | {} | NONE |
ClickHouse | ClickHouse | 713,627,593 | 16,236 | trollhe | ```
SET max_memory_usage = 20000000000000
```
| 2020-10-21T14:42:00 | 2020-10-21T14:42:00 | {} | NONE |
ClickHouse | ClickHouse | 713,981,109 | 16,192 | alexey-milovidov | @filimonov Maybe also worth to fix https://github.com/ClickHouse/ClickHouse/issues/12123? | 2020-10-21T23:32:21 | 2020-10-21T23:32:21 | {} | MEMBER |
ClickHouse | ClickHouse | 713,501,230 | 15,592 | filimonov | Does it also solve https://github.com/ClickHouse/ClickHouse/issues/16045 ? | 2020-10-21T11:26:52 | 2020-10-21T11:26:52 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,511,547 | 16,002 | alexey-milovidov | It's appeared to be important bugfix. | 2020-10-21T11:47:47 | 2020-10-21T11:47:47 | {} | MEMBER |
ClickHouse | ClickHouse | 713,536,275 | 16,223 | zhang2014 | CC: @tavplubix
I's not sure if the will be useful for ClickHouse(master branch).
BTW: Maybe we should provide a binary for `utils/` in CI(they are very useful). this will help some cannot build ClickHouse users | 2020-10-21T12:34:03 | 2020-10-21T12:34:50 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,539,330 | 10,998 | alexey-milovidov | Fix: https://github.com/ClickHouse/ClickHouse/pull/15646 | 2020-10-21T12:39:25 | 2020-10-21T12:39:25 | {} | MEMBER |
ClickHouse | ClickHouse | 713,640,974 | 16,237 | den-crane | You missed an alias
groupUniqArray(arrayJoin( users) ) **user**
MV uses column names to make an insert into destination table. | 2020-10-21T15:00:59 | 2020-10-21T15:02:01 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,715,034 | 16,221 | filimonov | Yes, but the idea is to collect different measurements (not from the query start only)
Analogy from network: you can have a big ping, bug great speed. So you can wait few seconds till the first byte, but after you get that the download goes fast. If you will measure the speed since sending the request you will have... | 2020-10-21T16:54:34 | 2020-10-21T16:54:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,904,173 | 15,173 | den-crane | > And it happens way too often. In system.merges I can see that the part being dropped is also being merged at the same time. It seems to me pointless to continue merging a part when it is no longer needed. And according to the logs, CH tries to stop the merges but looks like it fails to do so, which causes timeouts.
... | 2020-10-21T22:06:11 | 2020-10-21T22:06:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,975,166 | 12,281 | alexey-milovidov | Still outdated, contains incomplete info. | 2020-10-21T23:28:20 | 2020-10-21T23:28:20 | {} | MEMBER |
ClickHouse | ClickHouse | 713,519,360 | 15,856 | nikitamikhaylov | Merged in neighboring PR. | 2020-10-21T12:03:00 | 2020-10-21T12:03:00 | {} | MEMBER |
ClickHouse | ClickHouse | 713,649,856 | 16,236 | trollhe | > > inflate failed: buffer error
>
> I think it's a bug. How is your file compressed?
file by zlib compression,is *.gz | 2020-10-21T15:13:40 | 2020-10-21T15:13:40 | {} | NONE |
ClickHouse | ClickHouse | 713,735,783 | 16,192 | filimonov | Stress-test failure - can't diagnose that (it seems like clickhouse-server.log was rotated, and the important info is in the rotated / gzipped file). | 2020-10-21T17:30:10 | 2020-10-21T17:30:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,741,316 | 16,192 | akuzm | > > Just a side note - it also seems that clickhouse-local can start merging parts of the attached tables on its own, so it may be potentially unsafe to use it to introspect the data of the offline server.
>
> Discussed in Telegram that it's Ok to keep this behaviour.
`clickhouse-local` running background merges ... | 2020-10-21T17:40:02 | 2020-10-21T17:40:02 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,860,188 | 16,043 | alexey-milovidov | @filimonov
> I think we may need to adjust that to avoid having 2 copies.
We don't use GRPC's BoringSSL, this recursive submodule can be safely removed. | 2020-10-21T20:31:15 | 2020-10-21T20:31:15 | {} | MEMBER |
ClickHouse | ClickHouse | 713,963,719 | 12,557 | alexey-milovidov | For example:
\- flush every second if data size is at least one megabyte (to control max frequency under reasonable data size) but if not - flush every ten seconds (to control max latency) or every 10 megabytes (to limit memory usage). | 2020-10-21T23:20:59 | 2020-10-21T23:21:09 | {} | MEMBER |
ClickHouse | ClickHouse | 713,573,622 | 16,208 | den-crane | You have executed POPULATE twice. At each replica. It doubled the data.
You need to run POPULATE only at one replica. | 2020-10-21T13:28:22 | 2020-10-21T13:28:22 | {
"heart": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,827,492 | 16,190 | filimonov | Well i've made 2 images this way:
* filimonovq/clickhouse-server:20.8.4.11-busybox
* filimonovq/clickhouse-server:20.9.3.45-busybox
They work and pass some basic tests. The main problems - lack of bash (so if you pass some scripts in you need to have proper shebang), and lack of any package manager (in contrast ... | 2020-10-21T19:34:21 | 2020-10-21T20:12:56 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,901,752 | 16,043 | alexey-milovidov | The patch #8238 has to be reapplied, there are still issues with BoringSSL and asynchronous unwinding with LLVM's libunwind. | 2020-10-21T22:00:19 | 2020-10-21T22:00:19 | {} | MEMBER |
ClickHouse | ClickHouse | 713,583,501 | 16,222 | zvonand | > duplicate #15784
Not quite, the #15784 is about out-of normal-range in general, and THIS issue is ONLY about 1969-12-31, it has almost nothing in common with #15784 | 2020-10-21T13:39:37 | 2020-10-21T13:39:37 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,612,501 | 16,236 | den-crane | https://clickhouse.tech/docs/en/operations/settings/query-complexity/#settings_max_memory_usage | 2020-10-21T14:20:17 | 2020-10-21T14:20:17 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,613,602 | 16,236 | den-crane | @trollhe What CH version do use use?
It could be https://github.com/ClickHouse/ClickHouse/issues/7650 | 2020-10-21T14:21:57 | 2020-10-21T14:21:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,617,783 | 16,236 | trollhe | > @trollhe What CH version do use use?
> It could be #7650
the Clickhouse version is 20.9.3 | 2020-10-21T14:28:02 | 2020-10-21T14:28:02 | {} | NONE |
ClickHouse | ClickHouse | 713,835,891 | 15,236 | alesapin | All tests passed | 2020-10-21T19:51:39 | 2020-10-21T19:51:39 | {
"+1": 1
} | MEMBER |
ClickHouse | ClickHouse | 713,845,415 | 16,121 | azat | >I like the idea, I have also tried to make something like that but then found it difficult.
Ok, great (also it worth add a message when it is synced with RSS to see how it will differs later)
It still in draft due to gettid usage (that fact that performance does not shows the regression does not mean that there ... | 2020-10-21T20:10:28 | 2020-10-21T20:10:28 | {} | MEMBER |
ClickHouse | ClickHouse | 713,895,970 | 16,190 | alexey-milovidov | > Question - does clickhouse relies somehow on locales?
No.
clickhouse-client may work bad if `C` locale is used instead of `UTF-8`, but it's not the issue of clickhouse-client itself but the environment.
There is a risk that some external libraries depend on locales, but in 99% of cases it is erroneous, deve... | 2020-10-21T21:46:52 | 2020-10-21T21:47:15 | {} | MEMBER |
ClickHouse | ClickHouse | 713,495,648 | 16,221 | filimonov | Loosely related https://github.com/ClickHouse/ClickHouse/issues/12169 | 2020-10-21T11:15:19 | 2020-10-21T11:15:19 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,516,957 | 16,190 | alexey-milovidov | LGTM.
Motivation (vague):
- I've met some people that tend to say that using Ubuntu for Docker images is bad practice;
- when we base our images on Ubuntu, container scanners report about old software unrelated to ClickHouse.
We will merge as soon as there will be any evidence that it's working.
(It will not r... | 2020-10-21T11:58:27 | 2020-10-21T11:58:27 | {} | MEMBER |
ClickHouse | ClickHouse | 713,627,281 | 16,236 | trollhe | > https://clickhouse.tech/docs/en/operations/settings/query-complexity/#settings_max_memory_usage
thanks den-crane, my setting max_memory_usage,bug is error:
```
↗ Progress: 8.34 million rows, 1.51 GB (1.88 million rows/s., 341.09 MB/s.)
Received exception from server (version 20.9.3):
Code: 354. DB::Exception:... | 2020-10-21T14:41:33 | 2020-10-21T14:41:33 | {} | NONE |
ClickHouse | ClickHouse | 713,643,109 | 16,191 | alesapin | Still doesn't help:
```
create 1 23.091315507888794 INSERT INTO mt_100_parts SELECT number, rand() % 10000, rand() FROM numbers_mt(100000000)
create 0 23.331719875335693 INSERT INTO mt_100_parts SELECT number, rand() % 10000, rand() FROM numbers_mt(100000000)
create 1 42.9458713531494... | 2020-10-21T15:03:55 | 2020-10-21T15:03:55 | {} | MEMBER |
ClickHouse | ClickHouse | 713,644,916 | 16,236 | den-crane | >inflate failed: buffer error
I think it's a bug. How is your file compressed? | 2020-10-21T15:06:35 | 2020-10-21T15:06:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,667,381 | 16,214 | minaevmike | @den-crane Yes, sure. I've modified my initial message. | 2020-10-21T15:39:28 | 2020-10-21T15:39:28 | {} | NONE |
ClickHouse | ClickHouse | 713,719,168 | 16,216 | den-crane | it's aliases issue + join_use_nulls issue
```sql
SELECT if(isNull(fn), ss, fs) as s
FROM ( SELECT number fn, 'first' AS fs FROM numbers(2) ) AS first
FULL OUTER JOIN
( SELECT number sn, 'second' AS ss FROM numbers(2, 2)) AS second on fn=sn
SETTINGS join_use_nulls = 1
┌─s─────┐
│ first │
│ first │
└───────┘... | 2020-10-21T17:01:25 | 2020-10-21T17:01:25 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,872,203 | 16,121 | alexey-milovidov | You can also compare addresses of dummy thread_local variable (with the address of the same var from main thread). | 2020-10-21T20:55:48 | 2020-10-21T20:56:11 | {} | MEMBER |
ClickHouse | ClickHouse | 713,898,238 | 15,367 | CurtizJ | In master. | 2020-10-21T21:51:48 | 2020-10-21T21:51:48 | {} | MEMBER |
ClickHouse | ClickHouse | 713,942,912 | 13,894 | alexey-milovidov | Should be resolved with https://github.com/ClickHouse/ClickHouse/pull/16123. | 2020-10-21T23:08:13 | 2020-10-21T23:08:13 | {} | MEMBER |
ClickHouse | ClickHouse | 713,583,693 | 16,220 | den-crane | Can you show cluster `cubes` description?
And `select *, _shard_num from test_dist` | 2020-10-21T13:39:56 | 2020-10-21T13:40:10 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,591,119 | 16,214 | den-crane | And a data sample?
insert into test_1 values(1,'a'),(2,'b'); ? | 2020-10-21T13:51:11 | 2020-10-21T13:51:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,724,393 | 16,231 | den-crane | ```
20.3.19.4
SELECT
sumMapIf([1], [1], nullIf(number, 3) > 0) AS col1,
countIf(1, nullIf(number, 3) > 0) AS col2,
sumIf(1, nullIf(number, 3) > 0) AS col3
FROM numbers(1, 5)
┌─col1──────┬─col2─┬─col3─┐
│ ([1],[4]) │ 4 │ 4 │
└───────────┴──────┴──────┘
20.4.9.110
┌─col1──────┬─col2─┬─c... | 2020-10-21T17:10:31 | 2020-10-21T17:21:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,830,438 | 16,199 | azat | >The issue does not happen in the official build in CI.
>We are using clang-11, not 10.
>It fails compatibility check.
Yes for all of the above, however as it turns out it is not clang 10, but clang in unbundled build with libstdc++
>Maybe add a substitute from Musl?
Interesting idea | 2020-10-21T19:40:25 | 2020-10-21T19:40:25 | {} | MEMBER |
ClickHouse | ClickHouse | 713,837,492 | 16,190 | filimonov | It looks like alpine version also ok, I've pushed the image here:
* filimonovq/clickhouse-server:20.9.3.45-alpine
it's about 2Mb bigger than busybox as expected, but has a lot of advantages (better with security, package manager, will be the natural choice if we will ever switch to musl).
patch to make it wor... | 2020-10-21T19:55:03 | 2020-10-21T20:13:29 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 713,921,145 | 15,475 | alexey-milovidov | Looks like it's use after free, already fixed in https://github.com/ClickHouse/ClickHouse/pull/15984 | 2020-10-21T22:45:51 | 2020-10-21T22:45:51 | {} | MEMBER |
ClickHouse | ClickHouse | 713,959,368 | 12,615 | alexey-milovidov | @filimonov Could you please verify this issue? | 2020-10-21T23:18:28 | 2020-10-21T23:18:28 | {} | MEMBER |
ClickHouse | ClickHouse | 706,236,657 | 15,756 | alesapin | Test failures don't seem to be related to changes. | 2020-10-09T15:07:37 | 2020-10-09T15:07:37 | {} | MEMBER |
ClickHouse | ClickHouse | 706,586,540 | 15,531 | alexey-milovidov | The idea appeared to be flawed, sorry. | 2020-10-10T17:46:17 | 2020-10-10T17:46:17 | {} | MEMBER |
ClickHouse | ClickHouse | 706,611,161 | 15,497 | alexey-milovidov | The maximum unit argument is not checked for correctness. | 2020-10-10T21:03:34 | 2020-10-10T21:03:34 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.