Closed med-jordan closed 4 months ago
Hello Andre,
we need more information here,
and the message is not about volume exist but operation on the volume already exist , this is a transient message during processing
we need full CSI logs and status/event of PV/PVC/PODS
in all case CSI is a supported component so you can also open a support call directly
Hi Christophe,
all errors i can see in csi logs is this :
2024-03-01T10:09:58.587596308-05:00 stderr F (exit status 8) 2024-03-01T10:09:58.587599155-05:00 stderr F E0301 15:09:58.587357 1 utils.go:31] GRPC error: rpc error: code = Internal desc = iscsi/lvm failure, last err seen: 2024-03-01T10:09:58.587601756-05:00 stderr F iscsi: failed to attach disk: Error: iscsiadm: Could not login to [iface: default, target: iqn.2010-06.com.nutanix:ntnx-k8s-793fec83-fb4c-4f52-88b2-4fc2fe12c188-tgt0, portal: 10.150.138.138,3260]. 2024-03-01T10:09:58.587604143-05:00 stderr F iscsiadm: initiator reported error (8 - connection timed out) 2024-03-01T10:09:58.587608699-05:00 stderr F iscsiadm: Could not log into all portals 2024-03-01T10:09:58.587611136-05:00 stderr F Logging in to [iface: default, target: iqn.2010-06.com.nutanix:ntnx-k8s-793fec83-fb4c-4f52-88b2-4fc2fe12c188-tgt0, portal: 10.150.138.138,3260] 2024-03-01T10:09:58.587613419-05:00 stderr F (exit status 8)
I checked port 3260 to ISCSI Data Services IP but its open ... credentials secret also ok... But i think youre right its better to create a support call directly.
Thanks Andre
is iscsid service started on the nodes ? there is two port to open for iscsi 3205, 3260 even if this is not a good idea to route iscsi traffic in all case there is the entire flow here https://portal.nutanix.com/page/documents/list?type=software&filterKey=software&filterVal=Ports%20and%20Protocols&productType=Nutanix%20Kubernetes%20Engine
can you give also status/event of PV/PVC/PODS
some other questions
and please share support ticket once open
rancher version is 2.8.1 k8s cluster is a rke2 cluster with embedded etcd OS on all nodes is centos stream 9
CentOS-Stream-GenericCloud-9-latest
is RKE2 used manually or with the nutanix rancher node driver ?
i have 3 nodes ( vms on nutanix cluster ) where is installed centos stream 9 and on top of that i have installed the rke2 cluster. on top of that there is the rancher server management interface installed where the rancher node driver is installed and activated.
so im using it with nutanix rancher node driver and its working good ... i can create kubernetes clusters without any problems... only the installed csi storage driver ( installed via helm chart in the store ) have the described behavior.
also have a second env where rancher runs on vm in docker container ... also same behavior.
Good New Christophe - after checking port tcp 7050,3205,2379 and udp 123 with firewall colleauges the issue is resolved.
Port tcp 7050 - Authenticates clients accessing clusters, and provides access to the NKE when a firewall exists between Prism Central and the workstation and provides authentication.
Thank you for your support and BR Andre
ok good news
Hi Gents,
im using latest rancher version and latest nutanix csi snapshot and csi storage. All nodes meet the requirements as described. I can create a pvc and it get status bound and is online in rancher and nutanix prism elements.
When creating a new pod with the pvc then i got the error that the given volume allready exists.
"MountVolume.SetUp failed for volume "pvc-xxxxx-xxxxx-xxxxxx-xxxxx" : rpc error: code = Unavailable desc = An operation with the given Volume ID NutanixVolumes-xxxxx-xxxxxx-xxxxxx-xxxxxxx already exists"
Any ideas how to solve this ?
Thanks and BR Andre