Kubernetes pod pending when a new volume is attached (EKS) - kubernetes

Question

Let me describe my scenario:
TL;DR
When I create a deployment on Kubernetes with 1 attached volume, everything works perfectly. When I create the same deployment, but with a second volume attached (total: 2 volumes), the pod gets stuck on "Pending" with errors:
pod has unbound PersistentVolumeClaims (repeated 2 times)
0/2 nodes are available: 2 node(s) had no available volume zone.
Already checked that the volumes are created in the correct availability zones.
Detailed description
I have a cluster set up using Amazon EKS, with 2 nodes. I have the following default storage class:
kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
name: gp2
annotations:
storageclass.kubernetes.io/is-default-class: "true"
provisioner: kubernetes.io/aws-ebs
parameters:
type: gp2
reclaimPolicy: Retain
mountOptions:
- debug
And I have a mongodb deployment which needs two volumes, one mounted on /data/db folder, and the other mounted in some random directory I need. Here is an minimal yaml used to create the three components (I commented some lines on purpose):
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
namespace: my-project
creationTimestamp: null
labels:
io.kompose.service: my-project-db-claim0
name: my-project-db-claim0
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 5Gi
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
namespace: my-project
creationTimestamp: null
labels:
io.kompose.service: my-project-db-claim1
name: my-project-db-claim1
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 10Gi
---
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
namespace: my-project
name: my-project-db
spec:
replicas: 1
strategy:
type: Recreate
template:
metadata:
labels:
name: my-db
spec:
containers:
- name: my-project-db-container
image: mongo
imagePullPolicy: Always
resources: {}
volumeMounts:
- mountPath: /my_dir
name: my-project-db-claim0
# - mountPath: /data/db
# name: my-project-db-claim1
ports:
- containerPort: 27017
restartPolicy: Always
volumes:
- name: my-project-db-claim0
persistentVolumeClaim:
claimName: my-project-db-claim0
# - name: my-project-db-claim1
# persistentVolumeClaim:
# claimName: my-project-db-claim1
That yaml works perfectly. The output for the volumes is:
$ kubectl describe pv
Name: pvc-307b755a-039e-11e9-b78d-0a68bcb24bc6
Labels: failure-domain.beta.kubernetes.io/region=us-east-1
failure-domain.beta.kubernetes.io/zone=us-east-1c
Annotations: kubernetes.io/createdby: aws-ebs-dynamic-provisioner
pv.kubernetes.io/bound-by-controller: yes
pv.kubernetes.io/provisioned-by: kubernetes.io/aws-ebs
Finalizers: [kubernetes.io/pv-protection]
StorageClass: gp2
Status: Bound
Claim: my-project/my-project-db-claim0
Reclaim Policy: Delete
Access Modes: RWO
Capacity: 5Gi
Node Affinity: <none>
Message:
Source:
Type: AWSElasticBlockStore (a Persistent Disk resource in AWS)
VolumeID: aws://us-east-1c/vol-xxxxx
FSType: ext4
Partition: 0
ReadOnly: false
Events: <none>
Name: pvc-308d8979-039e-11e9-b78d-0a68bcb24bc6
Labels: failure-domain.beta.kubernetes.io/region=us-east-1
failure-domain.beta.kubernetes.io/zone=us-east-1b
Annotations: kubernetes.io/createdby: aws-ebs-dynamic-provisioner
pv.kubernetes.io/bound-by-controller: yes
pv.kubernetes.io/provisioned-by: kubernetes.io/aws-ebs
Finalizers: [kubernetes.io/pv-protection]
StorageClass: gp2
Status: Bound
Claim: my-project/my-project-db-claim1
Reclaim Policy: Delete
Access Modes: RWO
Capacity: 10Gi
Node Affinity: <none>
Message:
Source:
Type: AWSElasticBlockStore (a Persistent Disk resource in AWS)
VolumeID: aws://us-east-1b/vol-xxxxx
FSType: ext4
Partition: 0
ReadOnly: false
Events: <none>
And the pod output:
$ kubectl describe pods
Name: my-project-db-7d48567b48-slncd
Namespace: my-project
Priority: 0
PriorityClassName: <none>
Node: ip-192-168-212-194.ec2.internal/192.168.212.194
Start Time: Wed, 19 Dec 2018 15:55:58 +0100
Labels: name=my-db
pod-template-hash=3804123604
Annotations: <none>
Status: Running
IP: 192.168.216.33
Controlled By: ReplicaSet/my-project-db-7d48567b48
Containers:
my-project-db-container:
Container ID: docker://cf8222f15e395b02805c628b6addde2d77de2245aed9406a48c7c6f4dccefd4e
Image: mongo
Image ID: docker-pullable://mongo#sha256:0823cc2000223420f88b20d5e19e6bc252fa328c30d8261070e4645b02183c6a
Port: 27017/TCP
Host Port: 0/TCP
State: Running
Started: Wed, 19 Dec 2018 15:56:42 +0100
Ready: True
Restart Count: 0
Environment: <none>
Mounts:
/my_dir from my-project-db-claim0 (rw)
/var/run/secrets/kubernetes.io/serviceaccount from default-token-pf9ks (ro)
Conditions:
Type Status
Initialized True
Ready True
PodScheduled True
Volumes:
my-project-db-claim0:
Type: PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
ClaimName: my-project-db-claim0
ReadOnly: false
default-token-pf9ks:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-pf9ks
Optional: false
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s
node.kubernetes.io/unreachable:NoExecute for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling 7m22s (x5 over 7m23s) default-scheduler pod has unbound PersistentVolumeClaims (repeated 2 times)
Normal Scheduled 7m21s default-scheduler Successfully assigned my-project/my-project-db-7d48567b48-slncd to ip-192-168-212-194.ec2.internal
Normal SuccessfulMountVolume 7m21s kubelet, ip-192-168-212-194.ec2.internal MountVolume.SetUp succeeded for volume "default-token-pf9ks"
Warning FailedAttachVolume 7m13s (x5 over 7m21s) attachdetach-controller AttachVolume.Attach failed for volume "pvc-307b755a-039e-11e9-b78d-0a68bcb24bc6" : "Error attaching EBS volume \"vol-01a863d0aa7c7e342\"" to instance "i-0a7dafbbdfeabc50b" since volume is in "creating" state
Normal SuccessfulAttachVolume 7m1s attachdetach-controller AttachVolume.Attach succeeded for volume "pvc-307b755a-039e-11e9-b78d-0a68bcb24bc6"
Normal SuccessfulMountVolume 6m48s kubelet, ip-192-168-212-194.ec2.internal MountVolume.SetUp succeeded for volume "pvc-307b755a-039e-11e9-b78d-0a68bcb24bc6"
Normal Pulling 6m48s kubelet, ip-192-168-212-194.ec2.internal pulling image "mongo"
Normal Pulled 6m39s kubelet, ip-192-168-212-194.ec2.internal Successfully pulled image "mongo"
Normal Created 6m38s kubelet, ip-192-168-212-194.ec2.internal Created container
Normal Started 6m37s kubelet, ip-192-168-212-194.ec2.internal Started container
Everything is created without any problems. But if I uncomment the lines in the yaml so two volumes are attached to the db deployment, the pv output is the same as earlier, but the pod gets stuck on pending with the following output:
$ kubectl describe pods
Name: my-project-db-b8b8d8bcb-l64d7
Namespace: my-project
Priority: 0
PriorityClassName: <none>
Node: <none>
Labels: name=my-db
pod-template-hash=646484676
Annotations: <none>
Status: Pending
IP:
Controlled By: ReplicaSet/my-project-db-b8b8d8bcb
Containers:
my-project-db-container:
Image: mongo
Port: 27017/TCP
Host Port: 0/TCP
Environment: <none>
Mounts:
/data/db from my-project-db-claim1 (rw)
/my_dir from my-project-db-claim0 (rw)
/var/run/secrets/kubernetes.io/serviceaccount from default-token-pf9ks (ro)
Conditions:
Type Status
PodScheduled False
Volumes:
my-project-db-claim0:
Type: PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
ClaimName: my-project-db-claim0
ReadOnly: false
my-project-db-claim1:
Type: PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
ClaimName: my-project-db-claim1
ReadOnly: false
default-token-pf9ks:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-pf9ks
Optional: false
QoS Class: BestEffort
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute for 300s
node.kubernetes.io/unreachable:NoExecute for 300s
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Warning FailedScheduling 60s (x5 over 60s) default-scheduler pod has unbound PersistentVolumeClaims (repeated 2 times)
Warning FailedScheduling 2s (x16 over 59s) default-scheduler 0/2 nodes are available: 2 node(s) had no available volume zone.
I've already read these two issues:
Dynamic volume provisioning creates EBS volume in the wrong availability zone
PersistentVolume on EBS can be created in availability zones with no nodes (Closed)
But I already checked that the volumes are created in the same zones as the cluster nodes instances. In fact, EKS creates two EBS by default in us-east-1b and us-east-1c zones and those volumes works. The volumes created by the posted yaml are on those regions too.

Answer 1

See this article: https://kubernetes.io/blog/2018/10/11/topology-aware-volume-provisioning-in-kubernetes/
The gist is that you want to update your storageclass to include:
volumeBindingMode: WaitForFirstConsumer
This causes the PV to not be created until the pod is scheduled. It fixed a similar problem for me.

Answer 2

Sounds like it's trying to create a volume in an availability zone where you don't have any volumes on. You can try restricting your StorageClass to the availability zones where you have nodes.
kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
name: gp2
annotations:
storageclass.kubernetes.io/is-default-class: "true"
provisioner: kubernetes.io/aws-ebs
parameters:
type: gp2
reclaimPolicy: Retain
mountOptions:
- debug
allowedTopologies:
- matchLabelExpressions:
- key: failure-domain.beta.kubernetes.io/zone
values:
- us-east-1b
- us-east-1c
This is very similar to this question and this answer except that the issue described is on GCP and in this case it's AWS.

Answer 3

In this case, you should check the availability zone of your worker nodes (EC2 instances).
As a Example :
worker node 1 = eu-central-1b
worker node 2 = eu-central-1c
Then create the volume in including one of an availability zone which mentioned above(do not create the volume with eu-central-1a).
after you create the volume, create your PersistentVolume and PersistentVolumeClaim by attaching a newly created volume to your cluster like below.
apiVersion: v1
kind: PersistentVolume
metadata:
labels:
failure-domain.beta.kubernetes.io/region: eu-central-1
failure-domain.beta.kubernetes.io/zone: eu-central-1b
name: mongo-pv
namespace: default
spec:
accessModes:
- ReadWriteOnce
capacity:
storage: 100Gi
awsElasticBlockStore:
fsType: ext4
volumeID: aws://eu-central-1b/vol-063342ab9be5d2929
storageClassName: gp2
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: mongo-pvc
namespace: default
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 100Gi
storageClassName: gp2
volumeName: mongo-pv