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

Disk Read / write spam 100% disk usage #937

Open
WebKenth opened this issue Sep 23, 2024 · 6 comments
Open

Disk Read / write spam 100% disk usage #937

WebKenth opened this issue Sep 23, 2024 · 6 comments
Assignees

Comments

@WebKenth
Copy link

WebKenth commented Sep 23, 2024

image
image

image
To the left the blue bar shows when i closed ray

i have no idea how to reproduce but Ray has been spamming my disk for the past hour

It's happened a few times before but i am unsure how to reproduce, usually it happens after my mac goes to sleep and then wakes up after a few hours to continue work (think go home at 16:00 and start work at 8:00)

My usual fix is just to quit ray and restart it but it would be helpful if the app didn't screw up my SSD health, since SSD's only have a max number of writes, it's pretty annoying that an app can cause 100% disk writes

As shown on the image Ray managed to write 4,5 TB of data in the last 3 hours...
while it isn't obvious from the screenshot it's Ray, i assure you it was showing up in the tab as the top contender for writing to disk, i then promptly shut it down and took the screenshot

I am unsure what/why Ray needs to write to the disk at all? could it not just be in memory? ( imo it's rare to have the need to reopen ray at a later stage and see past dumbs )

@SwithFr
Copy link

SwithFr commented Sep 23, 2024

Same issue on Ubuntu 24.04.1

image

@tancou
Copy link

tancou commented Sep 24, 2024

I am experiencing the same issue. I am only using Ray with a remote SSH server.

I code on a macbook pro i9 , version 13.5
The ssh remote machine is a Ubuntu 22.04.5 LTS
Ray : Version 2.8.1 (2.8.1)

Occasionally (roughly every two days), Ray spikes to 100% CPU usage on a single core without any apparent reason. I also put my laptop into sleep mode when I travel.

@SlyDave
Copy link

SlyDave commented Oct 7, 2024

I'm getting this too, on Windows 11

hundreds of disk access per second
furthermore, this triggers Windows Defender to scan every single write/open

Image

@ah017
Copy link

ah017 commented Nov 13, 2024

Image

Also experiencing this on Windows 11 Pro 23H2 with a remote SSH server connection to an Ubuntu VM. High disk usage remains after disconnecting from the server and stops when I close Ray. I put the laptop in sleep/hibernate mode - will test this to see if this is causing the issue. I'll also test whether this happens when connected/disconnected, sleeping/not-sleeping.

@binaryfire
Copy link

I just bought the lifetime license and noticed the same issue on Windows 11. I'm only using Ray with a remote SSH server. Is there a fix in the works? Otherwise I'll see if I can get a refund.

@sebastienhenau
Copy link

Apologies for the late response to this issue, as this issue has been here for some time without any response from our side. The issues you are having with Ray is not the experience we aim to deliver. We understand the frustration around the performance of Ray and it's definitely something we are focusing on. At the moment we are working towards Ray version 3.0. Which contains a full refactor of the application from the ground up. It will have new look and a better user experience. The performance of the application is, as we are working towards the 3.0 release, mandatory and something we are keeping an eye on. We aim to improve on it in the next major release. We'll communicate through our channels when the next version will be ready for launch! Thank you all for your patience concerning this issue.

@sebastienhenau sebastienhenau self-assigned this Nov 29, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

7 participants