Skip to content

fix(lfs): extend GitHub token refresh buffer + add LFS fetch timeout #322

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

Merged
jrobotham-square merged 3 commits into from
Jun 1, 2026
Merged

fix(lfs): extend GitHub token refresh buffer + add LFS fetch timeout #322

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
9575ac8
fix(lfs): prevent stale GitHub token from breaking long LFS fetches
jrobotham-square Jun 1, 2026
1cb340a
snapshot/lfs: kill the whole process group on fetch timeout
jrobotham-square Jun 1, 2026
b99f263
review: tighten lfs process-group comment
jrobotham-square Jun 1, 2026
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
5 changes: 4 additions & 1 deletion internal/githubapp/config.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,9 +18,12 @@ type TokenCacheConfig struct {
}

// DefaultTokenCacheConfig returns default token cache configuration.
// RefreshBuffer must exceed the longest subprocess that bakes a token into
// its environment, so a token can't expire mid-call (GitHub App tokens have
// a fixed 1 h TTL).
func DefaultTokenCacheConfig() TokenCacheConfig {
return TokenCacheConfig{
RefreshBuffer: 5 * time.Minute,
RefreshBuffer: 30 * time.Minute,
JWTExpiration: 10 * time.Minute,
}
}
22 changes: 19 additions & 3 deletions internal/strategy/git/snapshot.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,7 @@ import (
"path/filepath"
"strings"
"sync"
"syscall"
"time"

"github.com/alecthomas/errors"
Expand All @@ -23,6 +24,8 @@ import (
"github.com/block/cachew/internal/snapshot"
)

const lfsFetchTimeout = 25 * time.Minute

func snapshotDirForURL(mirrorRoot, upstreamURL string) (string, error) {
repoPath, err := gitclone.RepoPathFromURL(upstreamURL)
if err != nil {
Expand Down Expand Up @@ -915,15 +918,28 @@ func (s *Strategy) generateAndUploadLFSSnapshot(ctx context.Context, repo *gitcl
}

// Fetch only the LFS objects referenced by HEAD (the default branch).
// Timeout must stay below githubapp.RefreshBuffer (30m) so the baked-in
// token can't expire mid-fetch and trigger a retry storm.
fetchStart := time.Now()
fetchCmd, err := repo.GitCommand(ctx, "-C", workDir, "lfs", "fetch", "origin", "HEAD")
fetchCtx, cancel := context.WithTimeout(ctx, lfsFetchTimeout)
fetchCmd, err := repo.GitCommand(fetchCtx, "-C", workDir, "lfs", "fetch", "origin", "HEAD")
Comment thread
jrobotham-square marked this conversation as resolved.
if err != nil {
cancel()
s.metrics.recordLFSPhase(ctx, upstream, "fetch", "error", time.Since(fetchStart))
return errors.Wrap(err, "create git lfs fetch command")
}
if output, err := fetchCmd.CombinedOutput(); err != nil {
// git-lfs spawns transfer helpers that inherit our pipes; without
// killing the whole group, CombinedOutput stays blocked after the
// timeout fires.
fetchCmd.SysProcAttr = &syscall.SysProcAttr{Setpgid: true}
fetchCmd.Cancel = func() error {
return syscall.Kill(-fetchCmd.Process.Pid, syscall.SIGKILL)
}
output, fetchErr := fetchCmd.CombinedOutput()
cancel()
if fetchErr != nil {
s.metrics.recordLFSPhase(ctx, upstream, "fetch", "error", time.Since(fetchStart))
return errors.Wrapf(err, "git lfs fetch: %s", string(output))
return errors.Wrapf(fetchErr, "git lfs fetch: %s", string(output))
}
s.metrics.recordLFSPhase(ctx, upstream, "fetch", "success", time.Since(fetchStart))

Expand Down