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.

Sign up for GitHub

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

Jump to bottom

safekeeper: improve batching of pipelined AppendRequest #9689

Open
Tracked by #9624
erikgrinaker opened this issue Nov 8, 2024 · 0 comments · May be fixed by #9744
Open
Tracked by #9624

safekeeper: improve batching of pipelined AppendRequest #9689

erikgrinaker opened this issue Nov 8, 2024 · 0 comments · May be fixed by #9744
Assignees
@erikgrinaker
Labels
a/performance Area: relates to performance of the system c/storage/safekeeper Component: storage: safekeeper

Comments

@erikgrinaker
Copy link
Contributor

Currently, each AppendRequest is appended to the WAL. However, Tokio file IO has significant overhead, so many small appends can significantly affect throughput.

When processing pipelined AppendRequest, we should batch the appended WAL to amortize costs. There is no point batching across a segment boundary, since that implies an fsync.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@erikgrinaker erikgrinaker

Labels
a/performance Area: relates to performance of the system c/storage/safekeeper Component: storage: safekeeper
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

safekeeper: batch AppendRequest writes neondatabase/neon
1 participant
@erikgrinaker