Skip to content

Fix CPU spin when defunct runner is still in use #260

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

Closed
ericcurtin wants to merge 1 commit into from
Closed

Fix CPU spin when defunct runner is still in use #260

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 11 additions & 2 deletions pkg/inference/scheduling/loader.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -283,13 +283,22 @@ func (l *loader) idleCheckDuration() time.Duration {
// Compute the oldest usage time for any idle runner.
var oldest time.Time
for _, runnerInfo := range l.runners {
// Check if this runner is defunct (process has exited)
defunct := false
select {
case <-l.slots[runnerInfo.slot].done:
// Check immediately if a runner is defunct
return 0
defunct = true
default:
}

// Only check unused runners. Defunct runners that are still in use
// will be handled when they become unused, so we don't schedule
// immediate checks for them (which would cause a tight CPU loop).
if l.references[runnerInfo.slot] == 0 {
if defunct {
// Check immediately if an unused runner is defunct
return 0
}
timestamp := l.timestamps[runnerInfo.slot]
if oldest.IsZero() || timestamp.Before(oldest) {
oldest = timestamp
Expand Down