Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fast data parts loading by delaying table startup process #28313

Merged
merged 1 commit into from Aug 30, 2021

Conversation

amosbird
Copy link
Collaborator

I hereby agree to the terms of the CLA available at: https://yandex.ru/legal/cla/?lang=en

Changelog category (leave one):

  • Improvement

Changelog entry (a user-readable short description of the changes that goes to CHANGELOG.md):

Speed up data parts loading by delaying table startup process.

Detailed description / Documentation draft:
.

@robot-clickhouse robot-clickhouse added the pr-improvement Pull request with some product improvements label Aug 29, 2021
@tavplubix
Copy link
Member

Could you please explain why does it make data parts loading faster?

@amosbird
Copy link
Collaborator Author

Could you please explain why does it make data parts loading faster?

After table starting up, there can be many background merges and fetches and these tasks are quite disk intensive, which contend disk resources against the part loading process.

@amosbird
Copy link
Collaborator Author

A test result from production env: 40 min -> 24 min.

@tavplubix
Copy link
Member

ClickHouse special build check - internal CI issue

@tavplubix tavplubix merged commit b47f34a into ClickHouse:master Aug 30, 2021
@compasses
Copy link

@amosbird
Hi from my test, the Loading data parts why use single thread? which cost about 2 hours to load 300K parts.
And i have set the max_part_loading_threads to 48

the iotop result:
image

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
pr-improvement Pull request with some product improvements
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

4 participants