altinity / clickhouse Goto Github PK

Version: altinity/clickhouse-server:22.3.8.40.altinitystable (docker)
Issue related to problem: ClickHouse#35357
Probably fixed in: ClickHouse@7d61fa5
Source: https://t.me/clickhouse_ru/282520

Any chance of being merged into a stable release?

https://github.com/Altinity/ClickHouse/runs/7269277280?check_suite_focus=true

We need Altinity rpm for RHEL8 / CentOS8

Hi,

I am new to Clickhouse and I am raising to get an urgent help since I am implementing a big system using Clickhouse.

• CH version - 23.3.8.22.altinitystable
• Machine - centos79
• CPU - 8
• Nodes - 20
• Memory - 32 GB
• Shards - 5

Table Structure- ReplicatedMergeTree

CREATE TABLE table ON CLUSTER '{cluster}' ( col1INT(),col2INT(),col3INT(),col4INT(),timeDateTime,col5FLOAT(64),col6 INT() ) ENGINE = ReplicatedMergeTree('/clickhouse/tables/{cluster}/{shard}/<>/<>', '{replica}') ORDER BY (col1, toYYYYMMDD(time), col2, col3, col4) PARTITION BY (toYYYYMMDD(time)) PRIMARY KEY (col1)

Select Query -

SELECT col1, col2, col3, col4, time, col5, col6 FROM table WHERE col1 = 30 AND col2 = 11 ORDER BY time LIMIT 10 OFFSET 1

I am trying to hit a huge request with this query to CH cluster and the CPU is using around 80% of each nodes and my response time seems to be around 40 ms in each nodes.

I have connected CH to an golang API but my QPS of the API seems very less with max of 5K with big API and CH cluster resources.

I am using an external load balancer to clickhouse tcp client connection with 9000 port.
I feel the response time seems very less and I want to have a 100k qps with the system.
Am I aiming too much with this resources?

Do I need more resources for this usecase.
But I have seen people getting maximum performance with one or two servers.

Please help me to increase the performance or let me know If I am missing something!
Thanks in Advance!

We need Altinity rpm with debug info

Describe the issue
According to the documentation at https://docs.altinity.com/releasenotes/altinity-stable-release-notes/22.8/altinity-stable-22.8.20/, PR #278 should be not related to #51492

I noticed ClickHouse#51492 is not v22.8.20.12.altinitystable or the most recent v23.3.8.22.altinitystable but it is v22.8.21.38-lts and v23.8.2.8-lts in the official CH repo.

Could you please take a look and see if we can get a fresh build out fast with this fix?

Thanks,

Almost all tables replicate between nodes. Stored data is Graphite and Prometheus metrics. Carbon-clickhouse used as a writer.
Graphite-clickhouse used as a reader.
04,05,06 - read/write nodes.
02 - just replica, not used for reads/writes.
mark_cache_size is set to default 5GB.
Dictionaries are not used.
The uncompressed cache is disabled.

Executed on 04: