This message was deleted.
# rke2a
adamant-kite-43734
07/28/2023, 2:52 AMThis message was deleted.
c
creamy-pencil-82913
07/28/2023, 3:02 AMCheck the logs on the nodes. Specifically the cloud provider pod logs. Vsphere config is probably missing something.
s
swift-sunset-4572
07/28/2023, 3:09 AMOkay understood , let me have a look at it and see if i can find anything will share it here.
swift-sunset-4572
07/28/2023, 3:33 AMEverything seems to be in a pending state
swift-sunset-4572
07/28/2023, 3:36 AMwhen i try to describe it i get this
swift-sunset-4572
07/28/2023, 3:36 AMDo i have to have worker nodes seperately ?
swift-sunset-4572
07/28/2023, 3:42 AMmore into exploring
when i check the logs on
kubectl logs rancher-vsphere-cpi-cloud-controller-manager-89wzf -n kube-system Copy code
E0728 03:32:36.545852 1 nodemanager.go:185] shakeOutNodeIDLookup failed. Err=nodeID is empty
I0728 03:32:36.545922 1 search.go:76] WhichVCandDCByNodeID nodeID: externalc1-pool1-08946451-8nzdb
E0728 03:34:08.754796 1 connection.go:177] Failed to create new client. err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out
E0728 03:34:08.754901 1 connection.go:63] Failed to create govmomi client. err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out
E0728 03:34:08.754941 1 connectionmanager.go:147] Cannot connect to vCenter with err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed o ut
E0728 03:36:19.826776 1 connection.go:177] Failed to create new client. err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out
E0728 03:36:19.826934 1 connection.go:63] Failed to create govmomi client. err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out
E0728 03:36:19.826976 1 connectionmanager.go:147] Cannot connect to vCenter with err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed o ut
E0728 03:38:30.898832 1 connection.go:177] Failed to create new client. err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out
E0728 03:38:30.899014 1 connection.go:63] Failed to create govmomi client. err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out
E0728 03:38:30.899058 1 connectionmanager.go:147] Cannot connect to vCenter with err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed o ut
E0728 03:38:31.899770 1 search.go:118] WhichVCandDCByNodeID error vc:Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out
E0728 03:38:31.899916 1 nodemanager.go:185] shakeOutNodeIDLookup failed. Err=Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out
E0728 03:38:31.899989 1 node_controller.go:229] error syncing 'externalc1-pool1-08946451-8nzdb': failed to get provider ID for node externalc1-pool1-08946451-8nzdb at cloudpro vider: failed to get instance ID from cloud provider: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out, requeuing
I0728 03:38:31.900141 1 node_controller.go:415] Initializing node externalc1-pool1-08946451-8nzdb with cloud provider
I0728 03:38:31.900287 1 search.go:76] WhichVCandDCByNodeID nodeID: externalc1-pool1-08946451-8nzdb
E0728 03:40:41.970929 1 connection.go:177] Failed to create new client. err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out
E0728 03:40:41.970966 1 connection.go:63] Failed to create govmomi client. err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: connect: connection timed out
E0728 03:40:41.970973 1 connectionmanager.go:147] Cannot connect to vCenter with err: Post "<https://vc.prgx.com:443/sdk>": dial tcp 10.0.77.254:443: conneswift-sunset-4572
07/28/2023, 3:45 AMI think this might be the reason its not able to remove the toleration that is causing the pods to go into pending state
c
creamy-pencil-82913
07/28/2023, 4:34 AMYes that would do it. Why can't it connect to your vcenter server?
s
swift-sunset-4572
07/28/2023, 4:37 AMwell thats because the network admins have not opened this url for the ips that are being used by the downstream clusters. so i am not even able to curl on this url from the downstream nodes.
this becomes a prerequisite for the cpi and csi to have this url accessible from the whole ip pool thats being utilised to create the downstream nodes and use vsphere as a storage . Right ?
c
creamy-pencil-82913
07/28/2023, 4:54 AMYes, the cloud and storage controllers need to talk to vsphere. All of the clusters - both the rancher cluster, and the clusters it provisions - that you want to integrate into vsphere, need to be able to talk to vsphere.
💯 1
s
swift-sunset-4572
07/28/2023, 5:57 AMPoint noted. Will test and update once i will get the desired access. 😇
swift-sunset-4572
07/29/2023, 2:53 AMHi, @creamy-pencil-82913
I was not provided with the access to the vcenter url , my network team said that its not possible to provide access on all nodes ( master/ worker) of all the clusters .
Is there way where i can deploy cpi and csi only on master nodes ?
c
creamy-pencil-82913
07/29/2023, 3:24 AMYou can configure node selectors in the chart values
s
swift-sunset-4572
07/29/2023, 3:26 AMHow many nodes do you suggest it should be deployed on ? If i deploy it on lets say 3 nodes would that be suffice for the whole cluster
Plus if i am provisioning a cluster and choosing a cloud provider as vsphere there i cannot use the node selector for cpi and csi
c
creamy-pencil-82913
07/29/2023, 5:40 PMYou would need to deploy HelmChartConfig resources to customize the cpi and csi chart values
s
swift-sunset-4572
07/29/2023, 6:07 PMSimilar to The process steps on this link will work right ?
https://github.com/rancher/rke2/issues/663
swift-sunset-4572
08/01/2023, 3:50 PM@rich-garage-1614
241 Views