Skip to content

Fix cachegroup dryrun affinity and warmup command #102

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
zxh326 merged 2 commits into from
Jan 23, 2026
Merged

Fix cachegroup dryrun affinity and warmup command #102

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
93cc0aa
Fix cachegroup dryrun affinity and warmup command
zxh326 Jan 21, 2026
5590fb3
Update internal/controller/cachegroup_controller.go
zxh326 Jan 21, 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
6 changes: 3 additions & 3 deletions internal/controller/cachegroup_controller.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -162,7 +162,7 @@ func (r *CacheGroupReconciler) sync(ctx context.Context, cg *juicefsiov1.CacheGr
}
groupBackUp := r.shouldAddGroupBackupOrNot(cg, actualState, expectState)
podBuilder := builder.NewPodBuilder(cg, secret, node, expectState, groupBackUp)
expectWorker := podBuilder.NewCacheGroupWorker(ctx)
expectWorker := podBuilder.NewCacheGroupWorker(ctx, false)

if actualState != nil && actualState.DeletionTimestamp != nil {
log.Info("actual worker is being deleted, skip updating worker, waiting for next reconciler", "worker", expectWorker.Name)
Expand Down Expand Up @@ -315,7 +315,7 @@ func (r *CacheGroupReconciler) parseExpectStateByScheduler(ctx context.Context,
for _, node := range nodes {
state := expectStates[node.Name]
podBuilder := builder.NewPodBuilder(cg, &corev1.Secret{}, node.Name, state, false)
pod := podBuilder.NewCacheGroupWorker(ctx)
pod := podBuilder.NewCacheGroupWorker(ctx, true)
if ok, err := sim.CanSchedulePodOnNode(ctx, pod, &node); err != nil {
return nil, err
} else {
Expand Down Expand Up @@ -684,7 +684,7 @@ func (r *CacheGroupReconciler) HandleFinalizer(ctx context.Context, cg *juicefsi
}
for node, expectState := range expectStates {
podBuilder := builder.NewPodBuilder(cg, secret, node, expectState, false)
expectWorker := podBuilder.NewCacheGroupWorker(ctx)
expectWorker := podBuilder.NewCacheGroupWorker(ctx, false)
if err := r.cleanWorkerCache(ctx, cg, *expectWorker); err != nil {
log.Error(err, "failed to clean worker cache", "worker", expectWorker.Name)
}
Expand Down
9 changes: 7 additions & 2 deletions pkg/builder/cache_group_pod.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -373,7 +373,7 @@ func (p *PodBuilder) genInitConfigVolumes() {
})
}

func (p *PodBuilder) NewCacheGroupWorker(ctx context.Context) *corev1.Pod {
func (p *PodBuilder) NewCacheGroupWorker(ctx context.Context, dryrun bool) *corev1.Pod {
worker := newBasicPod(p.cg, p.node)
p.genInitConfigVolumes()
p.genCacheDirs()
Expand All @@ -398,7 +398,12 @@ func (p *PodBuilder) NewCacheGroupWorker(ctx context.Context) *corev1.Pod {
if spec.Affinity == nil {
spec.Affinity = &corev1.Affinity{}
}
worker.Spec.Affinity = utils.AddNodeNameNodeAffinity(spec.Affinity.DeepCopy(), p.node)
// dryrun means just generate the pod spec without binding to a specific node
if !dryrun {
worker.Spec.Affinity = utils.AddNodeNameNodeAffinity(spec.Affinity.DeepCopy(), p.node)
} else {
worker.Spec.Affinity = spec.Affinity.DeepCopy()
}
}
}
maps.Copy(worker.Labels, spec.Labels)
Expand Down
6 changes: 3 additions & 3 deletions pkg/builder/job.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,7 +62,7 @@ func (j *JobBuilder) NewWarmUpJob() *batchv1.Job {
Name: common.WarmUpContainerName,
Image: image,
ImagePullPolicy: j.worker.Spec.Containers[0].ImagePullPolicy,
Command: j.getWarmUpCommand(),
Command: j.getWarmUpCommand(image),
Env: j.worker.Spec.Containers[0].Env,
SecurityContext: &corev1.SecurityContext{
Privileged: utils.ToPtr(true),
Expand Down Expand Up @@ -156,7 +156,7 @@ var (
ignoreMountOpts = []string{"foreground", "cache-size", "free-space-ratio", "group-weight", "cache-dir", "group-backup"}
)

func (j *JobBuilder) getWarmUpCommand() []string {
func (j *JobBuilder) getWarmUpCommand(image string) []string {
workerCommad := strings.Split(j.worker.Spec.Containers[0].Command[2], "\n")
authCmd := workerCommad[0]
volName, opts := utils.MustParseWorkerMountCmds(workerCommad[1])
Expand Down Expand Up @@ -241,7 +241,7 @@ func (j *JobBuilder) getWarmUpCommand() []string {
cmds = append(cmds, fmt.Sprintf("--%s", opt))
}
// enable debug mode to get warmup stats log
if utils.WarmupSupportStats(j.wu.Spec.Image) && !hasDebug {
if utils.WarmupSupportStats(image) && !hasDebug {
cmds = append(cmds, "--debug")
}
return []string{
Expand Down