Configuring Pod Autoscaling and NodePools
- Configure autoscaling and HorizontalPodAutoscaler
- Add a node pool and configure taints on the nodes for Pod anti-affinity
- Configure an exception for the node taint by adding a toleration to a Pod’s manifest
Connect to the lab GKE cluster and deploy a sample workload
In this task, you connect to the lab GKE cluster and create a deployment manifest for a set of Pods within the cluster.
Welcome to Cloud Shell! Type "help" to get started.
Your Cloud Platform project in this session is set to qwiklabs-gcp-04-7ea9bf2c579b.
Use “gcloud config set project [PROJECT_ID]” to change to a different project.
student_00_715f72a34b73@cloudshell:~ (qwiklabs-gcp-04-7ea9bf2c579b)$ export my_zone=us-central1-a
export my_cluster=standard-cluster-1
student_00_715f72a34b73@cloudshell:~ (qwiklabs-gcp-04-7ea9bf2c579b)$ source <(kubectl completion bash)
student_00_715f72a34b73@cloudshell:~ (qwiklabs-gcp-04-7ea9bf2c579b)$ gcloud container clusters get-credentials $my_cluster --zone $my_zone
Fetching cluster endpoint and auth data.
kubeconfig entry generated for standard-cluster-1.
student_00_715f72a34b73@cloudshell:~ (qwiklabs-gcp-04-7ea9bf2c579b)$ git clone https://github.com/GoogleCloudPlatform/training-data-analyst
Cloning into 'training-data-analyst'...
remote: Enumerating objects: 64994, done.
remote: Counting objects: 100% (131/131), done.
remote: Compressing objects: 100% (87/87), done.
remote: Total 64994 (delta 60), reused 99 (delta 41), pack-reused 64863
Receiving objects: 100% (64994/64994), 697.25 MiB | 14.96 MiB/s, done.
Resolving deltas: 100% (41502/41502), done.
Updating files: 100% (12864/12864), done.
student_00_715f72a34b73@cloudshell:~ (qwiklabs-gcp-04-7ea9bf2c579b)$ ln -s ~/training-data-analyst/courses/ak8s/v1.1 ~/ak8s
student_00_715f72a34b73@cloudshell:~ (qwiklabs-gcp-04-7ea9bf2c579b)$ cd ~/ak8s/Autoscaling/
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ ls
loadgen.yaml web-tolerations.yaml web.yaml
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ cat web.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: web
spec:
replicas: 1
selector:
matchLabels:
run: web
template:
metadata:
labels:
run: web
spec:
containers:
- image: gcr.io/google-samples/hello-app:1.0
name: web
ports:
- containerPort: 8080
protocol: TCP
resources:
# You must specify requests for CPU to autoscale
# based on CPU utilization
requests:
cpu: "250m"
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl create -f web.yaml --save-config
deployment.apps/web created
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl expose deployment web --target-port=8080 --type=NodePort
service/web exposed
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get service web
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
web NodePort 34.118.231.200 <none> 8080:31289/TCP 7s
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
Configure autoscaling on the cluster
In this task, you configure the cluster to automatically scale the sample application that you deployed earlier.
To configure your sample application for autoscaling (and to set the maximum number of replicas to four and the minimum to one, with a CPU utilization target of 1%), execute the following command.
When you use kubectl autoscale, you specify a maximum and minimum number of replicas for your application, as well as a CPU utilization target.
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get deployment
NAME READY UP-TO-DATE AVAILABLE AGE
web 1/1 1 1 61s
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl autoscale deployment web --max 4 --min 1 --cpu-percent 1
horizontalpodautoscaler.autoscaling/web autoscaled
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get deployment
NAME READY UP-TO-DATE AVAILABLE AGE
web 1/1 1 1 106s
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
Inspect the HorizontalPodAutoscaler object
The kubectl autoscale command you used in the previous task creates a HorizontalPodAutoscaler object that targets a specified resource, called the scale target, and scales it as needed.
The autoscaler periodically adjusts the number of replicas of the scale target to match the average CPU utilization that you specify when creating the autoscaler.
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get hpa
NAME REFERENCE TARGETS MINPODS MAXPODS REPLICAS AGE
web Deployment/web 0%/1% 1 4 1 56s
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl describe horizontalpodautoscaler web
Name: web
Namespace: default
Labels: <none>
Annotations: <none>
CreationTimestamp: Thu, 20 Jun 2024 12:30:48 +0000
Reference: Deployment/web
Metrics: ( current / target )
resource cpu on pods (as a percentage of request): 0% (1m) / 1%
Min replicas: 1
Max replicas: 4
Deployment pods: 1 current / 1 desired
Conditions:
Type Status Reason Message
---- ------ ------ -------
AbleToScale True ReadyForNewScale recommended size matches current size
ScalingActive True ValidMetricFound the HPA was able to successfully calculate a replica count from cpu resource utilization (percentage of request)
ScalingLimited False DesiredWithinRange the desired count is within the acceptable range
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedGetResourceMetric 64s horizontal-pod-autoscaler No recommendation
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get horizontalpodautoscaler web -o yaml
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
creationTimestamp: "2024-06-20T12:30:48Z"
name: web
namespace: default
resourceVersion: "474829"
uid: 3d6bbbe4-0241-4a8e-b1d7-3534415eb5c3
spec:
maxReplicas: 4
metrics:
- resource:
name: cpu
target:
averageUtilization: 1
type: Utilization
type: Resource
minReplicas: 1
scaleTargetRef:
apiVersion: apps/v1
kind: Deployment
name: web
status:
conditions:
- lastTransitionTime: "2024-06-20T12:32:18Z"
message: recent recommendations were higher than current one, applying the highest
recent recommendation
reason: ScaleDownStabilized
status: "True"
type: AbleToScale
- lastTransitionTime: "2024-06-20T12:31:07Z"
message: the HPA was able to successfully calculate a replica count from cpu resource
utilization (percentage of request)
reason: ValidMetricFound
status: "True"
type: ScalingActive
- lastTransitionTime: "2024-06-20T12:31:03Z"
message: the desired count is within the acceptable range
reason: DesiredWithinRange
status: "False"
type: ScalingLimited
currentMetrics:
- resource:
current:
averageUtilization: 0
averageValue: "0"
value: "0"
name: cpu
type: Resource
currentReplicas: 1
desiredReplicas: 1
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
Test the autoscale configuration
You need to create a heavy load on the web application to force it to scale out. You create a configuration file that defines a deployment of four containers that run an infinite loop of HTTP queries against the sample application web server.
You create the load on your web application by deploying the loadgen application using the loadgen.yaml file that has been provided for you.
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ cat loadgen.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: loadgen
spec:
replicas: 4
selector:
matchLabels:
app: loadgen
template:
metadata:
labels:
app: loadgen
spec:
containers:
- name: loadgen
image: k8s.gcr.io/busybox
args:
- /bin/sh
- -c
- while true; do wget -q -O- http://web:8080; donestudent_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl apply -f loadgen.yaml
deployment.apps/loadgen created
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get deploy
NAME READY UP-TO-DATE AVAILABLE AGE
loadgen 4/4 4 4 11s
web 1/1 1 1 4m41s
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get hpa
NAME REFERENCE TARGETS MINPODS MAXPODS REPLICAS AGE
web Deployment/web 58%/1% 1 4 4 4m11s
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
Once the loadgen Pod starts to generate traffic, the web deployment CPU utilization begins to increase. In the example output, the targets are now at 58% CPU utilization compared to the 1% CPU threshold.
To stop the load on the web application, scale the loadgen deployment to zero replicas:
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl scale deployment loadgen --replicas 0
deployment.apps/loadgen scaled
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get deploy
NAME READY UP-TO-DATE AVAILABLE AGE
loadgen 0/0 0 0 2m10s
web 4/4 4 4 6m40s
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
Get the list of deployments to verify that the web application has scaled down to the minimum value of 1 replica that you configured when you deployed the autoscaler:
Manage node pools
In this task, you create a new pool of nodes using preemptible instances, and then you constrain the web deployment to run only on the preemptible nodes.
Add a node pool
- To deploy a new node pool with three preemptible VM instances, execute the following command:
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ gcloud container node-pools create "temp-pool-1" \
--cluster=$my_cluster --zone=$my_zone \
--num-nodes "2" --node-labels=temp=true --preemptible
Creating node pool temp-pool-1...done.
Created [https://container.googleapis.com/v1/projects/qwiklabs-gcp-04-7ea9bf2c579b/zones/us-central1-a/clusters/standard-cluster-1/nodePools/temp-pool-1].
NAME: temp-pool-1
MACHINE_TYPE: e2-medium
DISK_SIZE_GB: 100
NODE_VERSION: 1.29.4-gke.1043002
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get nodes
NAME STATUS ROLES AGE VERSION
gke-standard-cluster-1-temp-pool-1-c546e5da-d86h Ready <none> 73s v1.29.4-gke.1043002
gke-standard-cluster-1-temp-pool-1-c546e5da-dd9n Ready <none> 74s v1.29.4-gke.1043002
gke-standard-cluster-standard-cluster-f4dffa0b-pqsm Ready <none> 13h v1.29.4-gke.1043002
gke-standard-cluster-standard-cluster-f4dffa0b-vh6n Ready <none> 13h v1.29.4-gke.1043002
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get nodes -l temp=true
NAME STATUS ROLES AGE VERSION
gke-standard-cluster-1-temp-pool-1-c546e5da-d86h Ready <none> 99s v1.29.4-gke.1043002
gke-standard-cluster-1-temp-pool-1-c546e5da-dd9n Ready <none> 100s v1.29.4-gke.1043002
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
All the nodes that you added have the temp=true label because you set that label when you created the node-pool. This label makes it easier to locate and configure these nodes.
Control scheduling with taints and tolerations
To prevent the scheduler from running a Pod on the temporary nodes, you add a taint to each of the nodes in the temp pool. Taints are implemented as a key-value pair with an effect (such as NoExecute) that determines whether Pods can run on a certain node. Only nodes that are configured to tolerate the key-value of the taint are scheduled to run on these nodes.
- To add a taint to each of the newly created nodes, execute the following command.
You can use the temp=true label to apply this change across all the new nodes simultaneously:
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl taint node -l temp=true nodetype=preemptible:NoExecute
node/gke-standard-cluster-1-temp-pool-1-c546e5da-d86h tainted
node/gke-standard-cluster-1-temp-pool-1-c546e5da-dd9n tainted
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
To allow application Pods to execute on these tainted nodes, you must add a tolerations key to the deployment configuration.
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ cat web.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: web
spec:
replicas: 1
selector:
matchLabels:
run: web
template:
metadata:
labels:
run: web
spec:
tolerations:
- key: "nodetype"
operator: Equal
value: "preemptible"
nodeSelector:
temp: "true"
containers:
- image: gcr.io/google-samples/hello-app:1.0
name: web
ports:
- containerPort: 8080
protocol: TCP
resources:
# You must specify requests for CPU to autoscale
# based on CPU utilization
requests:
cpu: "250m"
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl apply -f web.yaml
deployment.apps/web configured
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get pods
NAME READY STATUS RESTARTS AGE
web-68c78798d4-2pbkw 1/1 Running 0 44s
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl describe pods -l run=web
Name: web-68c78798d4-2pbkw
Namespace: default
Priority: 0
Service Account: default
Node: gke-standard-cluster-1-temp-pool-1-c546e5da-d86h/10.10.0.6
Start Time: Thu, 20 Jun 2024 12:47:29 +0000
Labels: pod-template-hash=68c78798d4
run=web
Annotations: <none>
Status: Running
IP: 10.140.2.4
IPs:
IP: 10.140.2.4
Controlled By: ReplicaSet/web-68c78798d4
Containers:
web:
Container ID: containerd://49fd98301dae5d7719b537ca139e3dafc179316adf04ad8e464d6f688a49a427
Image: gcr.io/google-samples/hello-app:1.0
Image ID: gcr.io/google-samples/hello-app@sha256:b1455e1c4fcc5ea1023c9e3b584cd84b64eb920e332feff690a2829696e379e7
Port: 8080/TCP
Host Port: 0/TCP
State: Running
Started: Thu, 20 Jun 2024 12:47:31 +0000
Ready: True
Restart Count: 0
Requests:
cpu: 250m
Environment: <none>
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-qlxhj (ro)
Conditions:
Type Status
PodReadyToStartContainers True
Initialized True
Ready True
ContainersReady True
PodScheduled True
Volumes:
kube-api-access-qlxhj:
Type: Projected (a volume that contains injected data from multiple sources)
TokenExpirationSeconds: 3607
ConfigMapName: kube-root-ca.crt
ConfigMapOptional: <nil>
DownwardAPI: true
QoS Class: Burstable
Node-Selectors: temp=true
Tolerations: node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
nodetype=preemptible
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 54s default-scheduler Successfully assigned default/web-68c78798d4-2pbkw to gke-standard-cluster-1-temp-pool-1-c546e5da-d86h
Normal Pulling 54s kubelet Pulling image "gcr.io/google-samples/hello-app:1.0"
Normal Pulled 53s kubelet Successfully pulled image "gcr.io/google-samples/hello-app:1.0" in 1.326s (1.326s including waiting)
Normal Created 53s kubelet Created container web
Normal Started 52s kubelet Started container web
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
A Tolerations section with nodetype=preemptible in the list should appear near the bottom of the (truncated) output.
The output confirms that the Pods will tolerate the taint value on the new preemptible nodes, and thus that they can be scheduled to execute on those nodes.
To force the web application to scale out again, scale the loadgen deployment back to four replicas:
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl scale deployment loadgen --replicas 4
deployment.apps/loadgen scaled
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
loadgen-65c9cf7485-4889r 1/1 Running 0 21s 10.140.1.11 gke-standard-cluster-standard-cluster-f4dffa0b-pqsm <none> <none>
loadgen-65c9cf7485-7xzhf 1/1 Running 0 21s 10.140.1.10 gke-standard-cluster-standard-cluster-f4dffa0b-pqsm <none> <none>
loadgen-65c9cf7485-krbkd 1/1 Running 0 20s 10.140.0.18 gke-standard-cluster-standard-cluster-f4dffa0b-vh6n <none> <none>
loadgen-65c9cf7485-m8j8h 1/1 Running 0 21s 10.140.0.17 gke-standard-cluster-standard-cluster-f4dffa0b-vh6n <none> <none>
web-68c78798d4-2cj74 1/1 Running 0 3s 10.140.3.6 gke-standard-cluster-1-temp-pool-1-c546e5da-dd9n <none> <none>
web-68c78798d4-2pbkw 1/1 Running 0 2m24s 10.140.2.4 gke-standard-cluster-1-temp-pool-1-c546e5da-d86h <none> <none>
web-68c78798d4-flln5 1/1 Running 0 3s 10.140.2.5 gke-standard-cluster-1-temp-pool-1-c546e5da-d86h <none> <none>
web-68c78798d4-jfdnl 1/1 Running 0 3s 10.140.3.5 gke-standard-cluster-1-temp-pool-1-c546e5da-dd9n <none> <none>
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
You could scale just the web application directly but using the loadgen app will allow you to see how the different taint, toleration and nodeSelector settings that apply to the web and loadgen applications affect which nodes they are scheduled on.
This shows that the loadgen app is running only on default-pool nodes while the web app is running only the preemptible nodes in temp-pool-1.
The taint setting prevents Pods from running on the preemptible nodes so the loadgen application only runs on the default pool. The toleration setting allows the web application to run on the preemptible nodes and the nodeSelector forces the web application Pods to run on those nodes.
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$ history
1 export my_zone=us-central1-a
2 export my_cluster=standard-cluster-1
3 source <(kubectl completion bash)
4 gcloud container clusters get-credentials $my_cluster --zone $my_zone
5 git clone https://github.com/GoogleCloudPlatform/training-data-analyst
6 ln -s ~/training-data-analyst/courses/ak8s/v1.1 ~/ak8s
7 cd ~/ak8s/Autoscaling/
8 ls
9 cat web.yaml
10 kubectl create -f web.yaml --save-config
11 kubectl expose deployment web --target-port=8080 --type=NodePort
12 kubectl get service web
13 kubectl get deployment
14 kubectl autoscale deployment web --max 4 --min 1 --cpu-percent 1
15 kubectl get deployment
16 kubectl get hpa
17 kubectl describe horizontalpodautoscaler web
18 kubectl get horizontalpodautoscaler web -o yaml
19 ls
20 cat loadgen.yaml
21 kubectl apply -f loadgen.yaml
22 kubectl get deploy
23 kubectl get hpa
24 kubectl scale deployment loadgen --replicas 0
25 kubectl get deploy
26 kubectl get hpa
27 kubectl get deploy
28 kubectl get hpa
29 kubectl get deploy
30 kubectl get hpa
31 gcloud container node-pools create "temp-pool-1" --cluster=$my_cluster --zone=$my_zone --num-nodes "2" --node-labels=temp=true --preemptible
32 kubectl get hpa
33 kubectl get nodes
34 kubectl get nodes -l temp=true
35 kubectl taint node -l temp=true nodetype=preemptible:NoExecute
36 nano web.yaml
37 cat web
38 cat web.yaml
39 vi web.yaml
40 cat web.yaml
41 rm web.yaml
42 vi web.yaml
43 cat web.yaml
44 kubectl apply -f web.yaml
45 kubectl get pods
46 kubectl describe pods -l run=web
47 kubectl scale deployment loadgen --replicas 4
48 kubectl get pods -o wide
49 history
student_00_715f72a34b73@cloudshell:~/ak8s/Autoscaling (qwiklabs-gcp-04-7ea9bf2c579b)$
