-
Notifications
You must be signed in to change notification settings - Fork 929
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.
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
maya-apiserver hitting panic / runtime error when POSTing volume #3064
Comments
It might be worth noting that I'd been running OpenEBS before this without having enabled iscsid on the host (it's enabled when this error happens) - could be something to do with artifacts created under that condition. |
Checking on this. |
Here are the YAML resources for the StorageClasses / StoragePoolClaims: apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: openebs-bulk
annotations:
openebs.io/cas-type: cstor
cas.openebs.io/config: |
- name: StoragePoolClaim
value: "cstor-bulk-pool"
- name: ReplicaCount
value: "1"
provisioner: openebs.io/provisioner-iscsi
---
apiVersion: openebs.io/v1alpha1
kind: StoragePoolClaim
metadata:
name: cstor-bulk-pool
namespace: openebs
annotations:
cas.openebs.io/config: |
- name: PoolResourceRequests
value: |-
memory: 100Mi
- name: PoolResourceLimits
value: |-
memory: 4Gi
spec:
name: cstor-bulk-pool
type: disk
poolSpec:
poolType: striped
blockDevices:
blockDeviceList:
- blockdevice-9d50e42f15fe83269356445eea81b1e2
---
apiVersion: openebs.io/v1alpha1
kind: StoragePoolClaim
metadata:
name: cstor-work-pool
namespace: openebs
annotations:
cas.openebs.io/config: |
- name: PoolResourceRequests
value: |-
memory: 100Mi
- name: PoolResourceLimits
value: |-
memory: 4Gi
spec:
name: cstor-work-pool
type: disk
poolSpec:
poolType: striped
blockDevices:
blockDeviceList:
- blockdevice-50bfc93f4a03ff39b6219fed5965fe7b
---
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: openebs-work
annotations:
openebs.io/cas-type: cstor
cas.openebs.io/config: |
- name: StoragePoolClaim
value: "cstor-work-pool"
- name: ReplicaCount
value: "1"
provisioner: openebs.io/provisioner-iscsi And here's the blocking PVC in question: apiVersion: v1
items:
- apiVersion: v1
kind: PersistentVolumeClaim
metadata:
annotations:
volume.beta.kubernetes.io/storage-provisioner: openebs.io/provisioner-iscsi
creationTimestamp: "2020-06-15T22:13:22Z"
finalizers:
- kubernetes.io/pvc-protection
labels:
app: prometheus
prometheus: prometheus-operator-prometheus
managedFields:
- apiVersion: v1
fieldsType: FieldsV1
fieldsV1:
f:metadata:
f:annotations:
.: {}
f:volume.beta.kubernetes.io/storage-provisioner: {}
f:labels:
.: {}
f:app: {}
f:prometheus: {}
f:spec:
f:accessModes: {}
f:resources:
f:requests:
.: {}
f:storage: {}
f:storageClassName: {}
f:volumeMode: {}
f:status:
f:phase: {}
manager: kube-controller-manager
operation: Update
time: "2020-06-15T22:13:25Z"
name: prometheus-prometheus-operator-prometheus-db-prometheus-prometheus-operator-prometheus-0
namespace: prometheus
resourceVersion: "787930"
selfLink: /api/v1/namespaces/prometheus/persistentvolumeclaims/prometheus-prometheus-operator-prometheus-db-prometheus-prometheus-operator-prometheus-0
uid: 46cf6742-d3ee-481c-ae06-1a1cf3890666
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 50Gi
storageClassName: openebs-bulk
volumeMode: Filesystem
status:
phase: Pending
kind: List
metadata:
resourceVersion: ""
selfLink: "" |
There are two issues:
|
I tried deleting and re-creating the storage pools / classes, and now I'm seeing errors like this from the apiserver:
|
For the Prometheus chart, can you try to set the metadata.name to allow for < 63 chars. Related issue: helm/charts#13170 |
Regarding the deletion and re-creation, are the block devices in unclaimed state. |
The block devices are in the "Released" state; how would I transition them to Unclaimed? Delete them and restart the ndm Pod? |
they should transition from Released to Unclaimed automatically. There is a job launched to clear any traces of older cstor pool prior to marking them as unclaimed. the ndm-operator logs should help to check why they are still in the "Released" state. |
In my case, I have assigned value of
My pod
|
I used a
|
This looks a lot like #1688: I'll see if Kubic isn't missing UPDATE: yeop
|
Okay, I installed e2fsprogs and rebooted, and for a brief, beautiful moment, I think I actually did see the pod get scheduled. Then at 09:30Z, almost on the dot, the kube API server died. Here are the kubelet logs from when that happened: I can't tell if the time was a coincidence, or if there's something in here the suggests what might have gone wrong. After rebooting, the API server started falling into a crash loop again. Here are the logs for one recent run, containing a few panic invocations that might explain why it's getting overwhelmed ( |
Well, I've decided I'm going to try bringing the cluster back up again from scratch and see if that somehow fixes it. If not, I'll do more digging and see if I can trace where the problem's coming from. For now, I'll go ahead and consider the original issue addressed - thanks! |
The following workaround is working for me.. My changes are below.
Added In AlertManager PVC spec.
In above snippet, added below entries
In Prometheus PVC spec.
In above snippet, added below entries
After executing with above modification, able to provision both pods. Output
PVC:
SVC:
The above entries will help to restrict number of characters for Pods and PVCs. |
resolved, thanks. |
Description
I just installed OpenEBS, and I've been tracking down a bug in my cluster where pods aren't getting scheduled because their volumes aren't getting provisioned. The trail goes cold at the logs for openebs-apiserver:
Your Environment
kubectl get nodes
:kubectl get pods --all-namespaces
:kubectl get services -A
:kubectl get sc
:kubectl get pv -A
:No resources found
kubectl get pvc -A
:/etc/os-release
): openSUSE MicroOS (Kubic)uname -a
):Linux studtop 5.7.1-1-default #1 SMP Wed Jun 10 11:53:46 UTC 2020 (6a549f6) x86_64 x86_64 x86_64 GNU/Linux
The text was updated successfully, but these errors were encountered: