"Error: Connection details are not configured. Run 'maglev login' first."

[AndiBuchmann157]

Sup Team,

after a fresh start aka reinstallation of the Appliance i get the following output on the CLI when trying to do an maglev related commands like "(sudo) maglev package status".

"Error: Connection details are not configured. Run 'maglev login' first."

I assigend (V)IPs to every Interface like the guides say. I am trying to (re-) install the appliance as a standalone node.

Whats the error / matter here?

Thanks in advance,

Andreas

[Tomas de Leon]

maglev login
Cluster:

Run the following command:
maglev login -u admin -p <passwd> -c <dnac-ip>:443

** Where the "dnac-ip" would be the ip address of one of your VIP addresses.  I would use the Clusterlink VIP ip address.

[AndiBuchmann157]

Hi Thomas,

what a coincidence that you are answering..exactly in this moment i found your CX-DNAC-IpAddressWorksheet_v2.pdf.. 

Will try this ASAP when i am back at the lab!

Thank you for now!

[AndiBuchmann157]

hi there,

no good news. i can not login using the command you provided. i tried every physicall ip and every ip with both "admin" and maglev user.

i can ping every physical interface but can not ping any vip.

see screenshot for some more details...

any other ideas?

[Tomas de Leon]

Please provide the following:

** Check & Verify Network Interface IP & VIP Configurations

$ cat /etc/network/interfaces
$ etcdctl get /maglev/config/cluster/cluster_network
$ cat /etc/resolv.conf | grep "^nameserver" | awk '{print $2}'
$ ip a | grep enp

### CHECK FOR IP GATEWAYS & ROUTES
#
$ route -n | grep "enp"

If you want to keep you data private, attach the file with the display output to a private message to me.

Regards

T.

[AndiBuchmann157]

See attached file for your requested output.

I found the following lines which made me kind of curios....:

$ etcdctl get /maglev/config/cluster/cluster_network

Error:  client: etcd cluster is unavailable or misconfigured
error #0: dial tcp 127.0.0.1:2379: getsockopt: connection refused
error #1: dial tcp 127.0.0.1:4001: getsockopt: connection refused

Thanks for your help!!!

[Tomas de Leon]

So at first glance you have not static routes setup.  On a default gateway on the recommended enterprise link.

NETWORK ADAPTER #1 10_Gbit port [enp10s0] - Intra Cluster Link (recommended)
NETWORK ADAPTER #2 1_Gbit port [enp1s0f0] - Management (recommended)
NETWORK ADAPTER #3 1_Gbit port [enp1s0f1] - Cloud Update Connectivity (recommended)
NETWORK ADAPTER #4 10_Gbit port [enp9s0] - Enterprise Network (recommended)

The list above it the recommend (but should be followed) personalities for the Network adapters.  The only network adapter that does not require a static route is the CLUSTERLINK since the DNACs should be in the same VLAN it is not needed.  The other Network Adapters need a default gateway (only 1) and then others Static Routes pointing to there default gateway.  they should also have DNS configured for each.

T.

[AndiBuchmann157]

I did another fresh install now and then followed the CVD Deployment Guide and only assigned a (V)IP to the enp10s0 and to the enp9so interfaces. physically connected are the enp9so and cimc port of the appliance. still no changes..cant login (maglev login on the cli) and no gui apperaring... :(

[Tomas de Leon]

in addition 2 of the NICs are DOWN

4: enp1s0f0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000
5: enp1s0f1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN group default qlen 1000

Please connect them... if they are not connected their VIPs will not come up

[AndiBuchmann157]

on the other hand two nics are up and i still only can ping their pyscial adresses and not the vips.. :/

[AndiBuchmann157 · Accepted solution]

Hi Community,

together with TAC i was able to fix the issue.

Attenion: this was the solution to my error - i am sure there are other reasons which can cause this / a similar error - so please dont take it as a "general" solution.

During the Setup Step #16 you are asked by the DNA-C to enter the "Services Subnet" and the "Cluster Services Subnet".  The format the Subnets needs to be entered is:

123.123.123.123/21

234.234.234.234/21

There is no space allowed between the network address and the subnet mask, like this:

123.123.123.123   /21

234.234.234.234   /21

We found this error in the logs:

FAILED - RETRYING: TASK: post_reboot : master, initializing the cluster using kubeadm (2 retries left).

FAILED - RETRYING: TASK: post_reboot : master, initializing the cluster using kubeadm (1 retries left).

fatal: [localhost]: FAILED! => {"attempts": 30, "changed": true, "cmd": "kubeadm init --skip-preflight-checks --config \"/etc/kubernetes/kubeadm.conf\"", "delta": "0:00:00.144671", "end": "2018-12-12 09:49:10.878848", "failed": true, "failed_when_result": true, "rc": 3, "start": "2018-12-12 09:49:10.734177", "stderr": "[service-subnet: Invalid value: \"10.240.248.0 /21\": couldn't parse subnet, pod-subnet: Invalid value: \"10.240.240.0 /21\": couldn't parse subnet]", "stdout": "[kubeadm] WARNING: kubeadm is in beta, please do not use it for production clusters.

TL;DR:

DNAC Setup failed because DNAC could initialize its cluster services because of a "space" between network adress and subnetmask in setup step 16. Use 123.123.123.123/21 and do not use 123.123.123.123  /21