Re: gluster volume service and endpoint lost

[Tim Dudgeon]

I'm not really sure if that was the case.

I did manage to fix this by manually re-creating the service and 
endpoints using yaml like this;


**

*apiVersion: v1*

*

kind: Service

metadata:

 labels:

   gluster.kubernetes.io/provisioned-for-pvc: pvc-name

 name: glusterfs-dynamic-pvc-name

spec:

 ports:

 - port: 1

   protocol: TCP

   targetPort: 1

 sessionAffinity: None

 type: ClusterIP

*

**

*apiVersion: v1*

*

kind: Endpoints

metadata:

 labels:

   gluster.kubernetes.io/provisioned-for-pvc: pvc-name

 name: glusterfs-dynamic-pvc-name

subsets:

- addresses:

 - ip: 10.0.0.18

 - ip: 10.0.0.24

 - ip: 10.0.0.8

 ports:

 - port: 1

   protocol: TCP

*

I have no idea whether this is the "correct" solution, but it seema to work.
Nor do I have any real idea as to what caused this. Googling suggests 
that this can happen if you delete and re-create the PVC in quick 
succession, but I'm pretty sure that was not the case here.


Tim

On 07/04/2019 23:51, Nikolas Philips wrote:

Hi Tim,
Is it possible that these failing PVC/PVs are mounted on the same 
compute node(s)? It might be, that a node has somehow issues with the 
gluster mount points / docker storage.
If it's only one node or so, you could either try to reset the storage 
(https://docs.openshift.com/container-platform/3.11/admin_guide/manage_nodes.html#managing-nodes-docker-reset) 
oder simply reinstall the node (safer).


Nikolas


Am So., 7. Apr. 2019 um 10:02 Uhr schrieb Tim Dudgeon 
mailto:tdudgeon...@gmail.com>>:


I have a series of GlusterFS PVC that were working fine.

For some of these the corresponding service and endpoint has been
lost.
e.g. when I do a `oc get svc` or `oc get endpoints` some of the
PVC are
not listed and those PVCs cannot then be mounted to a pod with an
error
like this:

> MountVolume.NewMounter initialization failed for volume
> "pvc-3aafc4fa-3e5e-11e9-8522-fa163eca01d7" : endpoints
> "glusterfs-dynamic-xxx" not found

Any thoughts on:

1. what might have caused this?

2. How to re-create the service and endpoints?

Tim

___
users mailing list
users@lists.openshift.redhat.com

http://lists.openshift.redhat.com/openshiftmm/listinfo/users

___
users mailing list
users@lists.openshift.redhat.com
http://lists.openshift.redhat.com/openshiftmm/listinfo/users

Re: gluster volume service and endpoint lost

[Nikolas Philips]

Hi Tim,
Is it possible that these failing PVC/PVs are mounted on the same compute
node(s)? It might be, that a node has somehow issues with the gluster mount
points / docker storage.
If it's only one node or so, you could either try to reset the storage (
https://docs.openshift.com/container-platform/3.11/admin_guide/manage_nodes.html#managing-nodes-docker-reset)
oder simply reinstall the node (safer).

Nikolas


Am So., 7. Apr. 2019 um 10:02 Uhr schrieb Tim Dudgeon :

> I have a series of GlusterFS PVC that were working fine.
>
> For some of these the corresponding service and endpoint has been lost.
> e.g. when I do a `oc get svc` or `oc get endpoints` some of the PVC are
> not listed and those PVCs cannot then be mounted to a pod with an error
> like this:
>
> > MountVolume.NewMounter initialization failed for volume
> > "pvc-3aafc4fa-3e5e-11e9-8522-fa163eca01d7" : endpoints
> > "glusterfs-dynamic-xxx" not found
>
> Any thoughts on:
>
> 1. what might have caused this?
>
> 2. How to re-create the service and endpoints?
>
> Tim
>
> ___
> users mailing list
> users@lists.openshift.redhat.com
> http://lists.openshift.redhat.com/openshiftmm/listinfo/users
>
___
users mailing list
users@lists.openshift.redhat.com
http://lists.openshift.redhat.com/openshiftmm/listinfo/users