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 | 739,024,777 | 15,413 | azat | countMatches - #16277 and #17459
countSubstrings - #17347
Looks like can be closed. | 2020-12-04T21:15:18 | 2020-12-04T21:15:18 | {} | MEMBER |
ClickHouse | ClickHouse | 739,404,251 | 17,830 | alexey-milovidov | Strings in binary file are encoded in the following way:
- length of the string in VarUInt format;
- then the bytes of the string itself.
If string in CSV file is unquoted, newline character in CSV is interpreted as record separator and does not belong to the string value.
What you see for special bytes in .bin... | 2020-12-05T20:23:26 | 2020-12-05T20:24:02 | {} | MEMBER |
ClickHouse | ClickHouse | 738,668,212 | 14,735 | slimtom95 | How about MySQL enum type?
Exception: Unknown data type family: enum. Maybe you meant: ['Enum','Enum8'] | 2020-12-04T09:16:05 | 2020-12-04T09:34:17 | {} | NONE |
ClickHouse | ClickHouse | 738,107,120 | 17,709 | alesapin | Test failed in pika
```
E pika.exceptions.ChannelClosed: (404, "NOT_FOUND - no exchange 'select' in vhost '/'")
``` | 2020-12-03T16:08:16 | 2020-12-03T16:08:16 | {} | MEMBER |
ClickHouse | ClickHouse | 738,514,833 | 12,589 | tangxueming | @Slach Thank you for your help.
```sql
#schema (in fact, I use distributed table)
CREATE TABLE pupu_monitor.apm_zipkin_shard
(
`timestamp` UInt64,
`id` String,
`kind` LowCardinality(String),
`name` LowCardinality(String),
`traceId` String,
`duration` UInt64,
`localEndpoint.servi... | 2020-12-04T02:20:03 | 2020-12-04T02:20:03 | {} | NONE |
ClickHouse | ClickHouse | 738,713,335 | 16,867 | BhavyaRajSharma | @nikitamikhaylov not sure how to raise it as PR so raised it as unexpected behavior which is marked as duplicate.
Let us know if there is any progress. | 2020-12-04T10:46:55 | 2020-12-04T10:47:45 | {} | NONE |
ClickHouse | ClickHouse | 739,449,388 | 17,623 | l1t1 | suggest a major version every year, a minor version each month. so that, rename 20.13 to 21.1 | 2020-12-06T03:31:15 | 2020-12-06T03:31:15 | {
"+1": 2
} | NONE |
ClickHouse | ClickHouse | 739,509,045 | 17,838 | sundy-li | BTW, it's a summary metric, why not use `system.parts`, it has more detail `metrics` (per database, table ...).
And it's also a table in memory for statistics, clickhouse exporters (or custom HTTP metrics handler) work good with it. | 2020-12-06T14:17:58 | 2020-12-06T14:18:47 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,517,419 | 13,890 | ucasfl | > Proposal: add settings to MergeTreeSettings. They will have higher priority than the global settings and take effect when they are set.
Hi, How can this be tested? @alexey-milovidov | 2020-12-06T15:20:18 | 2020-12-06T15:20:18 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 738,610,405 | 17,789 | MaxWk | emmmmm...
it's my fault,data shard cannot be randomly assigned here.. | 2020-12-04T07:09:03 | 2020-12-04T07:11:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,692,307 | 17,739 | alexey-milovidov | @spongedu
> Infrastructure engineer. Keen on database technologies. TiDB Committer.
Now you can make a correction:
> Infrastructure engineer. Keen on database technologies. TiDB and ClickHouse Committer. | 2020-12-04T10:02:00 | 2020-12-04T10:02:00 | {} | MEMBER |
ClickHouse | ClickHouse | 738,748,111 | 17,792 | CurtizJ | This feature is already implemented. You can set `TTL GROUP BY` for a table #10537. Data will be aggregated by prefix of the primary key, specified in the `GROUP BY` clause and columns will be updated by aggregate functions, that are specified in the `SET` clause.
One of possible usages:
```sql
CREATE TABLE metr... | 2020-12-04T12:05:54 | 2020-12-04T12:06:35 | {} | MEMBER |
ClickHouse | ClickHouse | 738,800,938 | 17,739 | spongedu | > @spongedu
>
> > Infrastructure engineer. Keen on database technologies. TiDB Committer.
>
> Now you can make a correction:
>
> > Infrastructure engineer. Keen on database technologies. TiDB and ClickHouse Committer.
:) | 2020-12-04T14:03:31 | 2020-12-04T14:03:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,837,843 | 17,801 | hatrena | I also noticed that I also have to run:
```sql
ALTER TABLE ".inner.my_table_replica" MODIFY ORDER BY (date, cityHash64(organization_id), integration_id, offer_id, sub_id)
```
That apparently, fixed the problem. can you please confirm this solution? | 2020-12-04T15:12:45 | 2020-12-04T15:12:45 | {} | NONE |
ClickHouse | ClickHouse | 738,105,882 | 17,731 | den-crane | try to disable WAL.
```
# cat /etc/clickhouse-server/conf.d/in_memory_parts_enable_wal.xml
<yandex>
<merge_tree>
<in_memory_parts_enable_wal>0</in_memory_parts_enable_wal>
</merge_tree>
</yandex>
service clickhouse-server restart
select value from system.merge_tree_settings where name ='i... | 2020-12-03T16:06:07 | 2020-12-03T16:06:07 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,136,334 | 17,679 | nikitamikhaylov | `Not enough space` is known issue. | 2020-12-03T16:55:23 | 2020-12-03T16:55:23 | {} | MEMBER |
ClickHouse | ClickHouse | 738,319,728 | 17,767 | tavplubix | I'm not sure if `std::filesystem:path` is suitable for paths in ZooKeeper, because ZooKeeper is actually not a filesystem. Some paths like `/q/../w`, `./file`, `/dir/` are fine in terms of `std::filesystem`, but all of them are invalid in terms of ZooKeeper.
Btw, which problem are you trying to solve?
| 2020-12-03T21:15:37 | 2020-12-03T21:15:37 | {} | MEMBER |
ClickHouse | ClickHouse | 738,780,507 | 11,110 | filimonov | Yes, you can use imports.
If you use it on server - put those extra schemas in /var/lib/clickhouse/format_schemas folder (all paths in your import will use that as a base folder).
If you use it on client - it will use current working directory as a base folder for your includes. | 2020-12-04T13:21:27 | 2020-12-04T13:21:27 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,809,030 | 17,623 | filimonov | > Currently not in roadmap, we did not collect the use cases,
The usage scenarios maybe smth like
a) minute partitions for data younger than ~15 minutes,
b) 15 minute partioning for data younger than ~few hours
c) hourly partitioning for data younger than ~few days
d) weekly partitioning for data younger than ... | 2020-12-04T14:19:31 | 2020-12-04T14:19:31 | {
"+1": 4
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,880,540 | 17,775 | mzealey | Yup that fixes it, a little better than `AggregateFunction` actually:
```
add column total_new SimpleAggregateFunction(sum, UInt64) CODEC(T64, LZ4)
... total_new SimpleAggregateFunction(sum, UInt64) 0.7236739151593901 7.997841015209072 2574139486 1.73 GiB 19.17 GiB
``` | 2020-12-04T16:31:03 | 2020-12-04T16:31:03 | {} | NONE |
ClickHouse | ClickHouse | 738,934,657 | 17,765 | den-crane | AFAIK it's another problem. Each site_id has own toStartOfDay UTC timestamp | 2020-12-04T18:08:06 | 2020-12-04T18:08:06 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,999,699 | 17,806 | den-crane | @ygj6 check `select table, parts_to_do, is_done from system.mutations` | 2020-12-04T20:22:55 | 2020-12-04T20:22:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,419,953 | 16,787 | kitaisreal | @alexey-milovidov changed submodule to new fast float repository. Also updated float precise parsing implementation to use fast_float by default. | 2020-12-05T21:43:15 | 2020-12-05T21:43:15 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,560,128 | 15,811 | alesapin | `test_distributed_over_live_view/test.py::TestLiveViewOverDistributedSuite::test_distributed_over_live_view_order_by_key[lv_over_distributed_table-node0]` -- don't care, flaky test | 2020-12-06T20:37:55 | 2020-12-06T20:37:55 | {} | MEMBER |
ClickHouse | ClickHouse | 738,113,485 | 17,623 | vkingnew | > nice job! **Window functions**
this function is Strong expectations | 2020-12-03T16:18:32 | 2020-12-03T16:18:32 | {
"+1": 7
} | NONE |
ClickHouse | ClickHouse | 738,521,907 | 12,589 | bobrik | @tangxueming your schema is not optimized for lookup by `id`, because `id` is not a part of the primary key.
Take a look at this: https://github.com/bobrik/jaeger/tree/ivan/clickhouse/plugin/storage/clickhouse#schema | 2020-12-04T02:36:12 | 2020-12-04T02:36:12 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,005,787 | 17,792 | den-crane | Earlier this was possible by clear_column in SummingMT. | 2020-12-04T20:35:59 | 2020-12-04T20:35:59 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,554,771 | 17,831 | alexey-milovidov | Special build check:
> docker: Error response from daemon: Get https://registry-1.docker.io/v2/yandex/clickhouse-deb-builder/manifests/latest: EOF.
Transient error, ignore. | 2020-12-06T19:58:44 | 2020-12-06T19:58:44 | {} | MEMBER |
ClickHouse | ClickHouse | 738,130,825 | 17,446 | RonanMorgan | I dig further with tshark today :
`log.log:Dec 3, 2020 16:10:35.700820027 CET /?query=INSERT+INTO+mytable.visits+%28VISITORCODE%2C+VISITNUMBER%2C+TIMESTARTED%2C+NUMBERPAGES%2C+OS%2C+BROWSER%2C+DEVICETYPE%2C+ADBLOCK%2C+LANGUAGE%2C+LANDINGPAGEURL%2C+VISITDURATION%2C+FIRSTREFERRERURL%2C+TRAFFICTYPE%2C+TIMESINCEPREVIOU... | 2020-12-03T16:46:31 | 2020-12-03T16:47:03 | {} | NONE |
ClickHouse | ClickHouse | 738,182,177 | 17,765 | den-crane | You can use `case` I think
```sql
SELECT case when y='mos' then toTimeZone(toStartOfDay(toDateTime(x, 'Europe/Moscow')), 'UTC')
when y='sant' then toTimeZone(toStartOfDay(toDateTime(x, 'America/Santiago')), 'UTC')
end event_time
from (select now() x, 'mos' y union all select now(), 'sa... | 2020-12-03T18:02:00 | 2020-12-03T18:02:00 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,841,082 | 17,317 | filimonov | Maybe also related to greedy execution of alias columns.
https://github.com/ClickHouse/ClickHouse/issues/2582 | 2020-12-04T15:18:30 | 2020-12-04T15:18:30 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,862,530 | 17,775 | den-crane | Interesting.
It looks like countState stored not as UInt64. Not sure. | 2020-12-04T15:57:26 | 2020-12-04T15:57:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,877,182 | 17,672 | filimonov | There are too many different materialized view, not all of them can help optimizing the queries on the original table.
They can do enrichements, they can have different keys, they can push the data to another cluster, they can have another lifetime.
It it not possible to build something like you want in general ca... | 2020-12-04T16:24:51 | 2020-12-04T16:24:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,316,218 | 17,775 | den-crane | it needs to be addressed somehow
```sql
SELECT countMerge(x)
FROM
(
SELECT initializeAggregation('countState', 100) AS x
)
┌─countMerge(x)─┐
│ 1 │
└───────────────┘
``` | 2020-12-05T16:16:57 | 2020-12-05T16:16:57 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,097,094 | 17,762 | den-crane | @stufently what version did you use before upgrade?
Do you use mutations?
Show create table ... | 2020-12-03T15:53:56 | 2020-12-03T15:54:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,607,475 | 17,789 | MaxWk | Here is my DDL
```
CREATE TABLE logs.test_shard on cluster track_cluster
(
`_id` String,
`time_ns` Int64 DEFAULT 0
)
ENGINE = ReplacingMergeTree(time_ns)
ORDER BY _id
SETTINGS index_granularity = 8192
create table logs.test on cluster track_cluster as logs.test_shard
ENGINE=Distributed(track... | 2020-12-04T07:01:58 | 2020-12-04T07:01:58 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,747,139 | 17,759 | hombit | Looks like `max_threads=1` resolves the issue, but performance is not good | 2020-12-04T12:03:28 | 2020-12-04T12:03:28 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,766,798 | 17,671 | alesapin | 
| 2020-12-04T12:50:02 | 2020-12-04T12:50:02 | {} | MEMBER |
ClickHouse | ClickHouse | 738,844,770 | 17,731 | Farfaday | `find . -name wal* | wc -l ` gives 27
` find . -name wal* | xargs -I {} du -m {} | awk '{sum+=$1}END{print sum}' ` gives 4152 | 2020-12-04T15:25:14 | 2020-12-04T15:25:14 | {} | NONE |
ClickHouse | ClickHouse | 738,853,801 | 17,731 | Farfaday | ```
SELECT
part_type,
sum(rows),
formatReadableSize(sum(data_compressed_bytes)),
formatReadableSize(sum(data_uncompressed_bytes))
FROM system.parts
WHERE active
GROUP BY part_type
┌─part_type─┬───sum(rows)─┬─formatReadableSize(sum(data_compressed_bytes))─┬─formatReadableSize(sum(data_uncompre... | 2020-12-04T15:41:26 | 2020-12-04T15:41:26 | {} | NONE |
ClickHouse | ClickHouse | 739,442,311 | 17,828 | l1t1 | https://clickhouse.tech/docs/en/sql-reference/data-types/decimal/ talks about overflow | 2020-12-06T01:47:17 | 2020-12-06T01:47:17 | {} | NONE |
ClickHouse | ClickHouse | 739,522,783 | 17,844 | den-crane | What do you mean? What is bash mode?
```
echo "select 1" |clickhouse-client
1
``` | 2020-12-06T15:58:18 | 2020-12-06T15:58:18 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,550,205 | 17,656 | bharatnc | @tavplubix , @alesapin this is my current implementation of the `ddl_worker_queue` table. It's now ready for review.
The table structure is as follows:
```text
:) describe system.ddl_worker_queue;
DESCRIBE TABLE system.ddl_worker_queue
Query id: 5c760afb-f63c-4ee9-b359-e4ee38c663f9
┌─name─────┬─type───... | 2020-12-06T19:23:49 | 2020-12-06T19:23:49 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,552,281 | 17,847 | fibersel | @alexey-milovidov , Where can I find unit-tests for compression codecs?File `src/Compression/tests/gtest_compressionCodec.cpp` is not referenced in any target. | 2020-12-06T19:39:31 | 2020-12-06T19:39:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,556,149 | 17,807 | azat | >I think this is the relevant message (set medium_nonsharded) that appears on the nodes that do the join incorrectly:
Indeed, and the reason is that JOIN is done locally on each node, so they will receive `parallel_replicas_count`/`parallel_replica_offset` (this is what is used for `max_parallel_replicas`), and this... | 2020-12-06T20:08:35 | 2020-12-06T20:08:35 | {} | MEMBER |
ClickHouse | ClickHouse | 739,557,918 | 16,723 | alesapin | `00840_long_concurrent_select_and_drop_deadlock` -- test fixed in master | 2020-12-06T20:21:28 | 2020-12-06T20:21:28 | {} | MEMBER |
ClickHouse | ClickHouse | 739,558,444 | 17,847 | alexey-milovidov | @fibersel unit tests are built into single binary `unit_tests_dbms`. | 2020-12-06T20:25:38 | 2020-12-06T20:25:38 | {} | MEMBER |
ClickHouse | ClickHouse | 737,378,140 | 17,623 | filimonov | Also quite often requested:
* Protobuf fixes
* problems with complicated structures in protobuf https://github.com/ClickHouse/ClickHouse/issues/16979 https://github.com/ClickHouse/ClickHouse/issues/15990 https://github.com/ClickHouse/ClickHouse/issues/10469 https://github.com/ClickHouse/ClickHouse/issues/9069 ht... | 2020-12-02T17:26:20 | 2020-12-04T13:48:49 | {
"+1": 1
} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,849,383 | 17,731 | detailyang | @Farfaday can you try this sql ?
```sql
SELECT
part_type,
sum(rows),
formatReadableSize(sum(data_compressed_bytes)),
formatReadableSize(sum(data_uncompressed_bytes))
FROM system.parts
WHERE active
GROUP BY part_type
``` | 2020-12-04T15:33:40 | 2020-12-04T15:33:40 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,849,072 | 17,731 | detailyang | @ctr it looks like your scene doesn't match mine and my wal is too large (my scene is too many tiny requests so i tweak the min_rows_for_compact_part = 8390848)
@CurtizJ what confused me is before restarting the server the total of memory parts is not high usage as the following:
```sql
SELECT
part_type,
... | 2020-12-04T15:33:09 | 2020-12-04T15:35:09 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,927,973 | 17,459 | azat | >Performance — 4 errors, 6 faster, 4 slower, 25 unstable
Failures related to new function is because there is no such in upstream/master | 2020-12-04T17:56:04 | 2020-12-04T17:56:04 | {} | MEMBER |
ClickHouse | ClickHouse | 739,021,553 | 17,805 | azat | >Integration tests flaky check (asan) — No flaky tests: fail: 0, passed: 57, error: 0
This time `test_ttl_move` passed | 2020-12-04T21:08:40 | 2020-12-04T21:08:40 | {} | MEMBER |
ClickHouse | ClickHouse | 739,317,410 | 17,819 | alexey-milovidov | What do you mean? | 2020-12-05T16:28:04 | 2020-12-05T16:28:04 | {} | MEMBER |
ClickHouse | ClickHouse | 738,106,166 | 17,762 | stufently | before version 20.10.2.20
Yes we use mutations for clean old data, but data in part fresh, will delete after 3 month.
Ask about scheme , maybe nda | 2020-12-03T16:06:35 | 2020-12-03T16:06:35 | {} | NONE |
ClickHouse | ClickHouse | 738,335,202 | 17,767 | bharatnc | > se ZooKeeper is actually not a filesystem
@tavplubix I agree that zk is not a filesystem & the general objective here is to just make the path concatenation more standardized using std::filesystem - like how it's being used few other places, for instance: https://github.com/ClickHouse/ClickHouse/pull/14558/files, ... | 2020-12-03T21:38:47 | 2020-12-03T22:04:01 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,618,563 | 17,731 | amorphina | We tried that but it did not help, starting time of clickhouse is unchanged. | 2020-12-04T07:30:45 | 2020-12-04T07:30:45 | {} | NONE |
ClickHouse | ClickHouse | 738,658,321 | 17,775 | mzealey | Ah I was not aware of `SimpleAggregateFunction` but it looks like it will make life simple.
Strangely though a simple count column (default compression settings in both cases) is taking a lot more space:
```
SELECT
table,
column,
type,
sum(column_data_compressed_bytes) / sum(rows) AS used_byt... | 2020-12-04T08:57:38 | 2020-12-04T08:57:38 | {} | NONE |
ClickHouse | ClickHouse | 739,123,140 | 17,790 | l1t1 | run explain sql
and test to query on motorvehicle_local? | 2020-12-05T04:23:41 | 2020-12-05T04:23:41 | {} | NONE |
ClickHouse | ClickHouse | 738,835,228 | 5,538 | ygj6 | @alex-zaitsev, @audricbizouarn, Could you give me some advice please? | 2020-12-04T15:07:57 | 2020-12-04T15:07:57 | {} | NONE |
ClickHouse | ClickHouse | 738,896,306 | 5,568 | filimonov | There is no bug here.
Clickhouse doesn't do anything with certs, it doesn't analyze the cert chain / etc. All that is done by OpenSSL.
Clickhouse just passes a few parameters to OpenSSL. ClickHouse has OpenSSL packaged in, and it has slightly different defaults (coming from ubuntu), and sometimes paths to defau... | 2020-12-04T17:00:33 | 2020-12-04T17:00:33 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,900,626 | 17,732 | filimonov | Also - only innermost subquery is distributed. Try to push most of the work to shard, let it return less data...
```
select brand_id, sumIf(engagement_cnt, <condition satisfied once per docid>), unin(doc_id) from metrics.dws_3_view
group by brand_id
``` | 2020-12-04T17:10:51 | 2020-12-04T17:10:51 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,150,348 | 11,298 | azat | Introduces some `-Wsuggest-destructor-override` warnings
<details>
```
[870/1542] Building CXX object src/Parsers/New/CMakeFiles/clickhouse_parsers_new.dir/AST/ColumnExpr.cpp.o
In file included from ../src/Parsers/New/AST/ColumnExpr.cpp:11:
/src/ch/clickhouse/src/Parsers/New/ClickHouseLexer.h:60:3: warning: '~... | 2020-12-05T09:10:35 | 2020-12-05T09:11:42 | {} | MEMBER |
ClickHouse | ClickHouse | 739,560,438 | 14,587 | alesapin | `01514_distributed_cancel_query_on_error` fixed in master. Fix cherry-picked. | 2020-12-06T20:40:07 | 2020-12-06T20:40:07 | {} | MEMBER |
ClickHouse | ClickHouse | 738,123,183 | 17,763 | altvod | Actually, this is fixed in the current version, so closing it | 2020-12-03T16:34:17 | 2020-12-03T16:34:17 | {} | NONE |
ClickHouse | ClickHouse | 738,538,340 | 12,589 | tangxueming | > @tangxueming your schema is not optimized for lookup by `id`, because `id` is not a part of the primary key.
>
> Take a look at this: https://github.com/bobrik/jaeger/tree/ivan/clickhouse/plugin/storage/clickhouse#schema
@bobrik Thanks.
1.After look at the blog, I maybe need to create a new table which set id... | 2020-12-04T03:17:39 | 2020-12-04T03:17:39 | {} | NONE |
ClickHouse | ClickHouse | 738,802,523 | 17,792 | mzealey | Yes from #14345 it looks like the PK cannot be updated so it wouldn't actually be downsampling unless we did this in the order by clause | 2020-12-04T14:06:40 | 2020-12-04T14:06:40 | {} | NONE |
ClickHouse | ClickHouse | 738,843,124 | 17,731 | detailyang | @CurtizJ Sorry for the late reply. My wal.bin is too large (300GB) see the below:
```bash
root@553bf0dab02a:/var/lib/clickhouse/store# find . -name wal* | wc -l
417
root@553bf0dab02a:/var/lib/clickhouse/store# find . -name wal* | xargs -I {} du -m {} | awk '{sum+=$1}END{print sum}'
300252
```
@amorphina c... | 2020-12-04T15:22:14 | 2020-12-04T15:24:14 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,866,830 | 11,518 | filimonov | You will not see that exception anymore on newer version.
See https://github.com/ClickHouse/ClickHouse/pull/11831
| 2020-12-04T16:05:22 | 2020-12-04T16:05:22 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,087,358 | 17,808 | vzakaznikov | @alexey-milovidov, this is a small PR that adds "How To Debug Why Test Failed" section to the `tests/TestFlows/README.md`. | 2020-12-05T00:17:55 | 2020-12-05T00:17:55 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,800,234 | 17,055 | l1t1 | I got an intel xeon 16c 64 g mem VM.
```
localhost :) select count(*) from numbers(100000)a,numbers(10000)b;
SELECT count(*)
FROM numbers(100000) AS a
, numbers(10000) AS b
Query id: efe0f0d4-2362-451c-8c79-c5b72899884e
┌────count()─┐
│ 1000000000 │
└────────────┘
1 rows in set. Elapsed: 3.821 sec. ... | 2020-12-04T14:01:55 | 2020-12-05T04:18:02 | {} | NONE |
ClickHouse | ClickHouse | 739,523,386 | 17,839 | den-crane | https://clickhouse.tech/docs/en/operations/settings/query-complexity/#settings_max_memory_usage
**In the default configuration file, the maximum is 10 GB.**
Check
`select * from system.settings where name = 'max_memory_usage'`
https://clickhouse.tech/docs/en/operations/settings/query-complexity/#settings-max_... | 2020-12-06T16:02:45 | 2020-12-06T16:02:45 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,120,356 | 17,737 | alesapin | Let me rerun some tests. | 2020-12-03T16:29:57 | 2020-12-03T16:29:57 | {} | MEMBER |
ClickHouse | ClickHouse | 738,292,359 | 17,681 | Akazz | Tests are OK | 2020-12-03T20:36:35 | 2020-12-03T20:36:35 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,691,697 | 17,739 | alexey-milovidov | LGTM.
I checked the functional tests - some of them failed because of low free space in CI.
PS. It's possible to write a test for the change in this PR but it will be too cumbersome for such a small change. | 2020-12-04T10:00:53 | 2020-12-04T10:00:53 | {} | MEMBER |
ClickHouse | ClickHouse | 738,725,829 | 17,731 | CurtizJ | @detailyang Can you show count and sizes of wal files? They are located in root directory of table and have name `wal.bin` for active log and names in format `wal_N_M.bin` for rotated logs. | 2020-12-04T11:12:52 | 2020-12-04T11:12:52 | {} | MEMBER |
ClickHouse | ClickHouse | 738,766,863 | 17,671 | alesapin | Almost win. | 2020-12-04T12:50:11 | 2020-12-04T12:50:11 | {} | MEMBER |
ClickHouse | ClickHouse | 738,902,235 | 17,749 | filimonov | @alexey-milovidov may be we can define a list of compatible aggregatefunctions states and allow casting between them? | 2020-12-04T17:13:29 | 2020-12-04T17:13:41 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,524,810 | 17,839 | lucasguo | Thanks for fast response. Problem resolved. I think get confused with max_server_memory_usage. | 2020-12-06T16:13:53 | 2020-12-06T16:13:53 | {} | NONE |
ClickHouse | ClickHouse | 738,707,519 | 12,609 | hodzanassredin | No answers????
We have eventhub in azure and can connect to it using kafkacat tool. Config:
security.protocol=SASL_SSL
sasl.mechanisms=PLAIN
sasl.username=$ConnectionString
sasl.password=Endpoint=xxx
In clickhouse(yandex cloud) we configured kafka engine the same way as kafkacat, but no luck, we have a lot... | 2020-12-04T10:34:08 | 2020-12-04T10:34:08 | {} | NONE |
ClickHouse | ClickHouse | 738,739,193 | 17,458 | qoega | we should check stress test report https://clickhouse-test-reports.s3.yandex.net/17458/e57ad9cb7e061bbbf35f2caf90a59ea387caed06/stress_test_(thread)/stderr.log | 2020-12-04T11:43:34 | 2020-12-04T11:43:34 | {} | MEMBER |
ClickHouse | ClickHouse | 738,898,212 | 17,446 | RonanMorgan | I did the same test today, this time there was 12 sec between the time when clickhouse received the request (chunked) and the time clickhouse throw the error : is it normal ? | 2020-12-04T17:04:59 | 2020-12-04T17:04:59 | {} | NONE |
ClickHouse | ClickHouse | 738,906,523 | 17,765 | filimonov | Known issue (p.2 https://github.com/ClickHouse/ClickHouse/issues/1298 )
Dynamic timezones switching may be quite costly, and storing timezone names is also costy.
You can just store one more time in client tz. | 2020-12-04T17:20:26 | 2020-12-04T17:20:26 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,151,532 | 17,646 | azat | Ok, I did the following:
- Add level for LockExceptionInThread to make decisions based on the level
- Add block_fault_injections for LockExceptionInThread
- ~~use LockExceptionInThread in StorageBuffer w/o blocking faulty injections~~
- Do not ignore server/user memory limits in StorageBuffer
- keep buffer_ignore_... | 2020-12-05T09:24:04 | 2020-12-05T18:26:13 | {} | MEMBER |
ClickHouse | ClickHouse | 739,474,700 | 17,833 | ucasfl | @CurtizJ | 2020-12-06T08:59:52 | 2020-12-06T08:59:52 | {} | COLLABORATOR |
ClickHouse | ClickHouse | 738,137,253 | 17,678 | nikitamikhaylov | `Not enough space` is known issue. | 2020-12-03T16:56:47 | 2020-12-03T16:56:47 | {} | MEMBER |
ClickHouse | ClickHouse | 738,688,281 | 17,767 | alexey-milovidov | > are fine in terms of std::filesystem, but all of them are invalid in terms of ZooKeeper
AFAIK, unless we canonize the path, it will simply do concatenation with collapsing of slashes - should be safe. | 2020-12-04T09:54:26 | 2020-12-04T09:54:26 | {} | MEMBER |
ClickHouse | ClickHouse | 738,874,115 | 17,775 | den-crane | Alexander Kuzmenkov @akuzm: Yes, there are zero high bytes are not written. If their numbers are mostly less than 128, then you will actually get close to one byte per line. Maybe T64 codec will give a similar effect to the Simple version.
https://github.com/ClickHouse/ClickHouse/blob/master/src/AggregateFunctions/A... | 2020-12-04T16:18:59 | 2020-12-04T17:34:11 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,133,502 | 16,947 | azat | Just a friendly reminder that this one still waiting for a review | 2020-12-05T06:16:23 | 2020-12-05T06:16:23 | {} | MEMBER |
ClickHouse | ClickHouse | 739,331,598 | 17,646 | azat | Actually [failed `01278_min_insert_block_size_rows_for_materialized_views`](https://clickhouse-test-reports.s3.yandex.net/17646/2777cbe2e055f4632b0a1ec866b9a723e9e798e5/fast_test.html#fail1) shows that the memory limits should be ignored, not just exception blocked, since otherwise it will bail on the next allocation
... | 2020-12-05T18:29:36 | 2020-12-05T18:29:36 | {} | MEMBER |
ClickHouse | ClickHouse | 739,558,177 | 16,723 | alesapin | testflows -- looks like random failures | 2020-12-06T20:23:32 | 2020-12-06T20:23:32 | {} | MEMBER |
ClickHouse | ClickHouse | 738,090,026 | 17,731 | den-crane | can you show an output of
```sql
select count() part_count
from system.parts
where active = 1 ;
select count() part_count, formatReadableSize(sum(bytes)) size, sum(rows) rows
from system.parts
where active = 1
group by database,table
order by part_count desc
limit 10;
``` | 2020-12-03T15:45:31 | 2020-12-03T15:45:31 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,095,825 | 17,731 | amorphina | here are the requested outputs on behalf of @Farfaday
```
HOSTNAME :) select count() part_count
:-] from system.parts
:-] where active = 1 ;
SELECT count() AS part_count
FROM system.parts
WHERE active = 1
┌─part_count─┐
│ 1278 │
└────────────┘
1 rows in set. Elapsed: 0.017 sec. Processed 1.28 th... | 2020-12-03T15:52:55 | 2020-12-03T15:54:07 | {} | NONE |
ClickHouse | ClickHouse | 738,106,154 | 17,762 | den-crane | try to disable WAL.
```
# cat /etc/clickhouse-server/conf.d/in_memory_parts_enable_wal.xml
<yandex>
<merge_tree>
<in_memory_parts_enable_wal>0</in_memory_parts_enable_wal>
</merge_tree>
</yandex>
service clickhouse-server restart
select value from system.merge_tree_settings where name ='i... | 2020-12-03T16:06:34 | 2020-12-03T16:06:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 738,330,123 | 17,775 | den-crane | countState -- is a bit overhead, now days you can use simpleAggregateFunction (sum)
I think it almost impossible to solve it.
I see only dumb way like
```sql
SELECT countMerge(x)
FROM
(
SELECT countState(arrayJoin(range(100))) AS x
)
┌─countMerge(x)─┐
│ 100 │
└───────────────┘
```
... | 2020-12-03T21:27:38 | 2020-12-03T21:27:38 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,331,825 | 17,788 | den-crane | Try https://github.com/Vertamedia/chproxy | 2020-12-05T18:31:34 | 2020-12-05T18:31:34 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 739,479,576 | 17,775 | azat | >it needs to be addressed somehow
Right now it works as expected, `initializeAggregation` accept single value
And you cannot initialize state just from a single value (how `uniqCombined` suppose to work?)
Something like `initializeAggregationArray` can be added, but this will not work fast anyway. | 2020-12-06T09:50:35 | 2020-12-06T09:50:35 | {} | MEMBER |
ClickHouse | ClickHouse | 739,513,552 | 17,839 | lucasguo | For reference, here is the full config.xml
> <?xml version="1.0"?>
> <yandex>
> <logger>
> <level>trace</level>
> <log>/var/log/clickhouse-server/clickhouse-server.log</log>
> <errorlog>/var/log/clickhouse-server/clickhouse-server.err.log</errorlog>
> <size>1000M</size>
> ... | 2020-12-06T14:51:49 | 2020-12-06T14:51:49 | {} | NONE |
ClickHouse | ClickHouse | 749,303,629 | 14,411 | compasses | @nikitamikhaylov Got it, thank you very much. | 2020-12-22T02:34:48 | 2020-12-22T02:34:48 | {} | CONTRIBUTOR |
ClickHouse | ClickHouse | 749,391,857 | 18,299 | alesapin | ```
2020.12.22 00:49:54.945305 [ 142 ] {} <Debug> ThreadFuzzer: Sleep probability 0.1
2020.12.22 00:49:54.945333 [ 142 ] {} <Debug> ThreadFuzzer: Yield probability 0.0
2020.12.22 00:49:54.945365 [ 142 ] {} <Debug> ThreadFuzzer: Migrate probability 0.0
2020.12.22 00:49:54.945396 [ 142 ] {} <Debug> ThreadFuzzer: CPU ... | 2020-12-22T07:26:19 | 2020-12-22T07:26:19 | {} | MEMBER |
ClickHouse | ClickHouse | 749,402,489 | 16,664 | alexey-milovidov | @FawnD2 I have made all the required changes for this PR and now we are ready to continue to work on Arrow server. | 2020-12-22T07:56:09 | 2020-12-22T07:56:09 | {} | MEMBER |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.