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

The pd leader occupies a large amount of cpu #8800

Open
Linzai99 opened this issue Nov 12, 2024 · 3 comments
Open

The pd leader occupies a large amount of cpu #8800

Linzai99 opened this issue Nov 12, 2024 · 3 comments
Labels
type/bug The issue is confirmed as a bug.

Comments

@Linzai99
Copy link

Bug Report

use tikv-client get region info from pd server, GetRegion QPS 66k, leader CPU 3600%, Goroutine count 244k

What did you do?

pressure test on online tikv culster

What did you expect to see?

Load balancing and low resource usage

What did you see instead?

GetRegion QPS 66k, leader CPU 3600%, Goroutine count 244k

What version of PD are you using (pd-server -V)?

pd-server v8.1.0,tikv-client v2.0.7

@Linzai99 Linzai99 added the type/bug The issue is confirmed as a bug. label Nov 12, 2024
@rleungx
Copy link
Member

rleungx commented Nov 12, 2024

Is there any profile?

@Linzai99
Copy link
Author

Is there any profile?

get profile and goroutine in the zip file
profile.zip

@rleungx
Copy link
Member

rleungx commented Nov 18, 2024

I think most of the CPU is occupied by handling gRPC requests since right now the GetRegion is an unary call. We are working on changing it to a streaming call, which could be help for this case.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
type/bug The issue is confirmed as a bug.
Projects
None yet
Development

No branches or pull requests

2 participants