-
-
Notifications
You must be signed in to change notification settings - Fork 43
Closed
Labels
needs-kindIndicates a PR lacks a label and requires one.Indicates a PR lacks a label and requires one.needs-priorityIndicates a PR lacks a label and requires one.Indicates a PR lacks a label and requires one.needs-triageIndicates an issue or PR lacks a label and requires one.Indicates an issue or PR lacks a label and requires one.
Milestone
Description
What would you like to be added:
When serving with ray for multi-host cases, you should be able to set the /dev/shm
size to improve the performance, see the log here:
2025-01-14 03:08:29,081 WARNING services.py:2022 -- WARNING: The object store is using /tmp instead of /dev/shm because /dev/shm has only 67108864 bytes available. This will harm performance! You may be able to free up space by deleting files in /dev/shm. If you are inside a Docker container, you can increase /dev/shm size by passing '--shm-size=5.04gb' to 'docker run' (or add it to the run_options list in a Ray cluster config). Make sure to set this to more than 30% of available RAM.
Ray natively supports a distributed shared memory store for large objects (objects that are bigger than 100KB) to efficiently manage the memory usage of the cluster. When using Linux, Ray tries to use /dev/shm for the shared memory. But if this is too small, Ray automatically fallback to /tmp folder, which usually can cause worse performance than using the shared memory.
Why is this needed:
Completion requirements:
This enhancement requires the following artifacts:
- Design doc
- API change
- Docs update
The artifacts should be linked in subsequent comments.
Metadata
Metadata
Assignees
Labels
needs-kindIndicates a PR lacks a label and requires one.Indicates a PR lacks a label and requires one.needs-priorityIndicates a PR lacks a label and requires one.Indicates a PR lacks a label and requires one.needs-triageIndicates an issue or PR lacks a label and requires one.Indicates an issue or PR lacks a label and requires one.