unkinben
e7c9387bcc
Fixes #71 ## Why `FindOrphanedBlobs` returned any blob not currently referenced. Because CAS dedups (the blob row can exist before its artifact/local_files row is written), a concurrent upload reusing an existing hash could have its S3 object deleted mid-flight by the GC. ## Changes - `FindOrphanedBlobs` now takes a `minAge` and only returns blobs with `created_at < now()-minAge`. - The collector passes a 1h `blobGracePeriod`. ## Validation - `go test ./internal/gc/...` and `make e2e` pass. --------- Co-authored-by: BenVincent <benvin@main.unkin.net> Reviewed-on: #86 Co-authored-by: Ben Vincent <ben@unkin.net> Co-committed-by: Ben Vincent <ben@unkin.net>
73 lines
1.7 KiB
Go
73 lines
1.7 KiB
Go
package gc
|
|
|
|
import (
|
|
"context"
|
|
"log/slog"
|
|
"time"
|
|
|
|
"git.unkin.net/unkin/artifactapi/internal/database"
|
|
"git.unkin.net/unkin/artifactapi/internal/storage"
|
|
)
|
|
|
|
// blobGracePeriod is how old an orphaned blob must be before GC will delete
|
|
// it. This avoids racing in-flight dedup uploads that insert the blob row
|
|
// before the referencing artifact/local_files row exists.
|
|
const blobGracePeriod = 1 * time.Hour
|
|
|
|
type Collector struct {
|
|
db *database.DB
|
|
store *storage.S3
|
|
interval time.Duration
|
|
}
|
|
|
|
func New(db *database.DB, store *storage.S3, interval time.Duration) *Collector {
|
|
return &Collector{db: db, store: store, interval: interval}
|
|
}
|
|
|
|
func (c *Collector) Run(ctx context.Context) {
|
|
slog.Info("gc started", "interval", c.interval)
|
|
ticker := time.NewTicker(c.interval)
|
|
defer ticker.Stop()
|
|
|
|
for {
|
|
select {
|
|
case <-ctx.Done():
|
|
slog.Info("gc stopped")
|
|
return
|
|
case <-ticker.C:
|
|
c.sweep(ctx)
|
|
}
|
|
}
|
|
}
|
|
|
|
func (c *Collector) sweep(ctx context.Context) {
|
|
start := time.Now()
|
|
|
|
orphaned, err := c.db.FindOrphanedBlobs(ctx, blobGracePeriod)
|
|
if err != nil {
|
|
slog.Error("gc: find orphaned blobs", "error", err)
|
|
return
|
|
}
|
|
|
|
deleted := 0
|
|
for _, blob := range orphaned {
|
|
if err := c.store.Delete(ctx, blob.S3Key); err != nil {
|
|
slog.Warn("gc: delete s3 object", "key", blob.S3Key, "error", err)
|
|
continue
|
|
}
|
|
if err := c.db.DeleteBlob(ctx, blob.ContentHash); err != nil {
|
|
slog.Warn("gc: delete blob row", "hash", blob.ContentHash, "error", err)
|
|
continue
|
|
}
|
|
deleted++
|
|
}
|
|
|
|
if deleted > 0 || len(orphaned) > 0 {
|
|
slog.Info("gc sweep complete",
|
|
"orphaned_found", len(orphaned),
|
|
"deleted", deleted,
|
|
"duration_ms", time.Since(start).Milliseconds(),
|
|
)
|
|
}
|
|
}
|