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

--parallel-files-upload is not working as expected from end to end #7211

Closed
Lyndon-Li opened this issue Dec 15, 2023 · 1 comment
Closed
Assignees
Labels
Milestone

Comments

@Lyndon-Li
Copy link
Contributor

Create 6 files in the volume
Run a data mover backup and set --parallel-files-upload=6
ParallelFilesUpload has been set correctly in DUCR's dataMoverConfig
Find the node where the data mover is running
Find the node-agent pid in the node
run ls -l /proc/<pid>/fd
There are not 6 files being opened in parallel. In most of time, the opened file number is 4, which is the number of CPU cores

@Lyndon-Li
Copy link
Contributor Author

This is related to below Kopia uploader's behavior:

  • Say the parallel value is set to X, X - 1 files are processed by the worker pool and the left 1 is processed by the current routine which is also traversing the current dir and assigning entries to the other workers
  • If the current routine meets a large file, traversing will be blocked even though there are free workers
  • On the other hand, Kopia uploader has another mechanism to handle large files --- file concatenation, that is, if a file is large enough, file will be divided into parts and uploaded concurrently through the same workers.

However, Velero doesn't enable the file concatenation feature. Let's try to enable it in v1.14 so as to gain the best performance.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

3 participants