https://rancher.com/ logo
Join Slack
Powered by
Has anyone played with an external CSI yet? I have...
# harvester
t
Has anyone played with an external CSI yet? I have portworx working for additional volumes. However the loading of an image fails.
m
what errors did you see? are you on 1.5-rc3?
t
yup 1.5. I am rebuilding the node to see if I did something wrong.
m
if that didn't help, first place i will check is the logs of the cdi pods in the 
harvester-system
namespace:
Copy code
k -n harvester-system get po -<http://lcdi.kubevirt.io|lcdi.kubevirt.io>
NAME                              READY   STATUS    RESTARTS        AGE
cdi-apiserver-7cc4c65d8b-tzkw2    1/1     Running   18 (131m ago)   10d
cdi-deployment-7fb79cb949-plphb   1/1     Running   18 (131m ago)   10d
cdi-operator-86f4b76d7b-k689c     1/1     Running   18 (131m ago)   10d
cdi-uploadproxy-dcbf66668-qp6mt   1/1     Running   18 (131m ago)   10d
t
i’ll look
which cdi pod handles the download of images?
b
We're using external CSI for data volumes, but have been having issues with backups.
t
which csi?
b
ceph rbd
t
ah.
I am testing portworx
talking to a pure array
b
Apparently the external csi's are "supported" but not for backups.
So we're also stuck on how to migrate VMs from one cluster to another.
t
are you testing 1.5?
m
@thousands-advantage-10804 the 
cdi-deployment-*
pod usually will provide some info
you can also check the status of the associated `DataVolume`: 
kubectl get datavolumes -A
@bland-article-62755 yeah, backup has been a bit of a pain. tl;dr is backup is not a csi standard, only snapshot is
last when i was involved with the upstream csi data protection wg, they tried to formalized the backup apis but never happened (afaik)
t
I found the issue..
5 Views
Open in Slack
PreviousNext
Powered by