Skip to content

remove legacy cache #600

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
merged 1 commit into from
Aug 22, 2023
Merged

remove legacy cache #600

Show file tree
Hide file tree
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
5 changes: 2 additions & 3 deletions pkg/controller/clusterstate/cache/cache.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -47,7 +47,6 @@ import (
"k8s.io/klog/v2"

"github.com/project-codeflare/multi-cluster-app-dispatcher/pkg/controller/clusterstate/api"
"k8s.io/apimachinery/pkg/util/wait"
)

//New returns a Cache implementation.
Expand Down Expand Up @@ -98,10 +97,10 @@ func newClusterStateCache(config *rest.Config) *ClusterStateCache {

func (sc *ClusterStateCache) Run(stopCh <-chan struct{}) {
klog.V(8).Infof("Cluster State Cache started.")
go sc.nodeInformer.Informer().Run(stopCh)
//go sc.nodeInformer.Informer().Run(stopCh)

// Update cache
go wait.Until(sc.updateCache, 0, stopCh)
//go wait.Until(sc.updateCache, 0, stopCh)

}

Expand Down
22 changes: 10 additions & 12 deletions pkg/controller/queuejob/queuejob_controller_ex.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,7 +31,6 @@ import (
"time"

"github.com/eapache/go-resiliency/retrier"
"github.com/gogo/protobuf/proto"
"github.com/hashicorp/go-multierror"
qmutils "github.com/project-codeflare/multi-cluster-app-dispatcher/pkg/quotaplugins/util"

Expand Down Expand Up @@ -71,7 +70,6 @@ import (
"github.com/project-codeflare/multi-cluster-app-dispatcher/pkg/controller/queuejobdispatch"

clusterstateapi "github.com/project-codeflare/multi-cluster-app-dispatcher/pkg/controller/clusterstate/api"
clusterstatecache "github.com/project-codeflare/multi-cluster-app-dispatcher/pkg/controller/clusterstate/cache"
)

// XController the AppWrapper Controller type
Expand Down Expand Up @@ -110,7 +108,7 @@ type XController struct {

//TODO: Do we need this local cache?
// our own local cache, used for computing total amount of resources
cache clusterstatecache.Cache
//cache clusterstatecache.Cache

// is dispatcher or deployer?
isDispatcher bool
Expand Down Expand Up @@ -207,8 +205,8 @@ func NewJobController(config *rest.Config, serverOption *options.ServerOption) *
initQueue: cache.NewFIFO(GetQueueJobKey),
updateQueue: cache.NewFIFO(GetQueueJobKey),
qjqueue: NewSchedulingQueue(),
//TODO: do we still need cache to be initialized?
cache: clusterstatecache.New(config),
//cache is turned-off, issue: https://github.com/project-codeflare/multi-cluster-app-dispatcher/issues/588
//cache: clusterstatecache.New(config),
schedulingAW: nil,
}
//TODO: work on enabling metrics adapter for correct MCAD mode
Expand Down Expand Up @@ -1157,12 +1155,12 @@ func (qjm *XController) ScheduleNext(qj *arbv1.AppWrapper) {
// Assume preemption will remove low priroity AWs in the system, optimistically dispatch such AWs

if aggqj.LessEqual(resources) {
//TODO: should we turn-off histograms?
unallocatedHistogramMap := qjm.cache.GetUnallocatedHistograms()
if !qjm.nodeChecks(unallocatedHistogramMap, qj) {
klog.Infof("[ScheduleNext] [Agent Mode] Optimistic dispatch for AW '%s/%s' requesting aggregated resources %v histogram for point in-time fragmented resources are available in the cluster %s",
qj.Name, qj.Namespace, qjm.GetAggregatedResources(qj), proto.MarshalTextString(unallocatedHistogramMap["gpu"]))
}
//cache is turned-off, refer issue: https://github.com/project-codeflare/multi-cluster-app-dispatcher/issues/588
// unallocatedHistogramMap := qjm.cache.GetUnallocatedHistograms()
// if !qjm.nodeChecks(unallocatedHistogramMap, qj) {
// klog.Infof("[ScheduleNext] [Agent Mode] Optimistic dispatch for AW '%s/%s' requesting aggregated resources %v histogram for point in-time fragmented resources are available in the cluster %s",
// qj.Name, qj.Namespace, qjm.GetAggregatedResources(qj), proto.MarshalTextString(unallocatedHistogramMap["gpu"]))
// }
// Now evaluate quota
fits := true
klog.Infof("[ScheduleNext] [Agent Mode] available resourse successful check for '%s/%s' at %s activeQ=%t Unsched=%t &qj=%p Version=%s Status=%+v.",
Expand Down Expand Up @@ -1464,7 +1462,7 @@ func (cc *XController) Run(stopCh <-chan struct{}) {

cache.WaitForCacheSync(stopCh, cc.appWrapperSynced)

//TODO: do we still need to run cache every second?
//cache is turned off, issue: https://github.com/project-codeflare/multi-cluster-app-dispatcher/issues/588
// update snapshot of ClientStateCache every second
//cc.cache.Run(stopCh)

Expand Down