Skip to content

Make worker async cache minimum thread configurable#17179

Open
humengyu2012 wants to merge 2 commits intoAlluxio:master-2.xfrom
humengyu2012:min-async-cache-thread
Open

Make worker async cache minimum thread configurable#17179
humengyu2012 wants to merge 2 commits intoAlluxio:master-2.xfrom
humengyu2012:min-async-cache-thread

Conversation

@humengyu2012
Copy link
Copy Markdown
Contributor

@humengyu2012 humengyu2012 commented Apr 2, 2023

What changes are proposed in this pull request?

Make worker async cache minimum thread configurable.

Why are the changes needed?

We want to concurrently cache a file of 30G size(250 blocks), but we found that the concurrency of the file cache is always 4 due to 250 < 512(the default value of queue size). If the number of blocks we cache at the same time is always less than 512, then we will never get a concurrency greater than 4. This is too slow, we should make the minimum threads configurable.

image

Does this PR introduce any user facing changes?

No

@alluxio-bot alluxio-bot added the API Change Changes covering public API label Apr 2, 2023
@ChunxuTang
Copy link
Copy Markdown
Member

@humengyu2012 Thanks for the contribution and your insightful findings!

After reviewing the PR, I'm not sure when the worker async cache minimum thread is very small (e.g. =1) or large (>>4), whether there'll be an impact on other read/write operations.

@beinan When you're available, would you mind taking a quick review of this PR? This might be helpful for our Dora worker and local cache development.

Copy link
Copy Markdown
Contributor

@maobaolong maobaolong left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It LGTM. @ChunxuTang The default behavior is not changed if we merge this PR, and we can specific the core pool size other than 4.

@github-actions
Copy link
Copy Markdown

This pull request has been automatically marked as stale because it has not had recent activity. It will be closed in two weeks if no further activity occurs. Thank you for your contributions.

@github-actions github-actions bot added the stale The PR/Issue does not have recent activities and will be closed automatically label Jul 25, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

API Change Changes covering public API stale The PR/Issue does not have recent activities and will be closed automatically

5 participants