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
Memory usage grows but never shrinks when running find
command
#674
Comments
Additionally we are now seeing issues with the mount getting OOM killed
We see the same behavior as this #630 |
Additional debug info
|
Behavior I have noticed, when under heavy use it seems the memory usage does fluctuate up and down, but as soon as the job that is running has completed it remains with the memory usage it was at. Its currently sitting at |
Hi @daltschu22, thank you for reporting the issue. Would you be able to share a few more details to help us reproduce it? In particular:
|
Same issue for me, mem usage can went up to 12-19Gb and never goes down. |
Thanks @passaro Find is simply Current object count for the bucket I was testing is 5030736. |
Hey @daltschu22, we'd like to investigate a bit more here. We weren't able to reproduce the issue when we last took a look here (sorry for not sharing this) but maybe there's some specifics that will help. What does a typical key look like, can you share one? For example...
|
find
command
The references to For the Reading from files is where we expect higher memory usage. Mountpoint prefetches a subsection of the file and keeps that ready to be read by applications to allow for performant sequential reads. Each file handle will store that prefetched portion of the file. If you have a use case that is opening a large number of files for reading concurrently, we'd like to learn more to help us prioritize mechanisms or configurations that can best manage memory usage during reading. Managing memory usage more effectively when reading from files is something we've been thinking about. When we have more information, we'll share more details on GitHub. |
Thanks Danny, to be clear I don't think this is a I understand what you are saying, it is curious that just metadata operations is triggering this. We have moved away from relying on mountpoint for now until it becomes more mature, but let me see if I can get a test instance spun up to replicate this. Thanks for digging into this! |
I can get more detailed info, but as an example one of the buckets used to test is 5+ Petabytes in size, with 100+ million objects. |
Mountpoint for Amazon S3 version
1.3.1
AWS Region
us-east-1
Describe the running environment
t3.xlarge EC2
Rocky Linux 8.8
x86
Mounting bucket cross account
Mountpoint options
What happened?
Mountpoint memory usage starts small. Around 0.1 when watched like so
Running a find across the bucket sees the memory usage slowly tick upwards. Stopping the find ends with mountpoint consuming
5.6
continuously and not releasing it back.This is also observed on another host where the usage is unclear, we see 18.3 percent memory usage, but no active processes using the mount. This number does not decrease over time.
Relevant log output
No response
The text was updated successfully, but these errors were encountered: