Issue while deploying openness worker-node

[Jaladi-Devika]

Hi,

Have ran ./deploy_ne.sh nodes. But failed with below error. Worker-node is pingable and able do ssh as well.

Can you please help me how to resolve this issue?

Thanks & Regards, Devika

[i-kwilk]

Hi Davika, Can you send the configuration that you are having by sending all git changes that you have made? Thanks, Krzysztof

[Jaladi-Devika]

HI,

below are the configuration have followed

1. set the hostnames On controller: $ hostnamectl set-hostname controller On worker_node: $ hostnamectl set-hostname node01

2. hostname devined in /etc/hosts On master node: 127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4 controller ::1 localhost localhost.localdomain localhost6 localhost6.localdomain6 controller On worker node: 127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4 node01 ::1 localhost localhost.localdomain localhost6 localhost6.localdomain6 node01

3. Did below thing in inventory.ini [all] controller ansible_ssh_user=root ansible_host=192.168.10.91 node01 ansible_ssh_user=root ansible_host=192.168.10.92

   [controller_group] controller

   [edgenode_group] node01

   [edgenode_vca_group]

   [ptp_master] controller

   [ptp_slave_group] node01

4. ssh-keygen

5. ssh-copy-id root@192.168.10.92

6. Did below changes in group_vars/all/10-default.yml Network Time Protocol (NTP) Enable machine's time synchronization with NTP server ntp_enable: true Servers to be used by NTP instead of the default ones (e.g. 0.centos.pool.ntp.org) ntp_servers: ["0.centos.pool.ntp.org"]

   kubernetes_cnis:

   • kubeovn
   • flannel

7. then ran ./deploy_ne.sh nodes

[i-kwilk]

Hello, It seems that OVS didn't start. OVS is a part of kube-ovn, in our setup, it is set to use DPDK by default. DPDK reqires hugepages, therefore you need to disable dpdk:

group_vars/all/10-default.yml

kubeovn_dpdk: false

or add some hugepages for example:

Enable and configure 1G HugePages:

group_vars/controller_group/10-default.yml

hugepage_size: "1G"
hugepage_amount: "16"

group_vars/edgenode_group/10-default.yml

hugepage_size: "1G"
hugepage_amount: "16"

group_vars/all/10-default.yml

kubeovn_dpdk: true kubeovn_dpdk_socket_mem: "1024,1024" kubeovn_dpdk_hugepage_size: "1Gi" kubeovn_dpdk_hugepages: "12Gi" kubeovn_dpdk_resources_requests: "2Gi" kubeovn_dpdk_resources_limits: "2Gi"

[i-kwilk]

Hi @Jaladi-Devika

Does the configuration help with the issue?

[Jaladi-Devika]

Hi @i-kwilk ,

Got some other issue below is the one. Looks like this is know issue. https://github.com/open-ness/openness-experience-kits/issues/49

TASK [kubernetes/cni/kubeovn/common : build OVS-DPDK image (this may take some time...)] ***** task path: /home/sysadmin/Devika/openness-experience-kits/roles/kubernetes/cni/kubeovn/common/tasks/main.yml:89 fatal: [controller]: FAILED! => { "changed": false }

MSG:

Thanks, Devika

[i-kwilk]

Hi @Jaladi-Devika

Right, this is a known issue reported #49 as some RPMs are removed, we are working on a fix with the team.

Do you agree to close that one?

Regards, Krzysztof

[Jaladi-Devika]

Hi @i-kwilk ,

When can we expect the fix will be ready in 20.06?

Thanks & Regards, Devika