Lab 4.5: Workaround-OOM #2 for SDC pod failures

farsli1

Hi everyone,
Some SDC jobs are not completing and pods are stuck in Init:Error status.
I tried the workaround script increasing liveness probe counter as suggested but I still have the same problem.
Could anyone help me to move forward ?
Thanks in advance

vmuthukrishnan

Hi Farah,

The issue is that we cannot pause or restart ONAP VMs. In Dublin a lot of components do not come up automatically when you restart VMs.

We would recommend that you plan for running the VMs for 2 to 3 days to complete all the lab exercises.

Regards
Vivek

vmuthukrishnan

Hi Farah,

When you setup the Socks tunnel use FireFox browser proxy settings as shown below.

Hope you have updated the /etc/hosts file with ONAP host names as per the documentation. Then try accessing the ONAP portal URL

Update the /etc/hosts file in the jump host (gcloud instance)

ONAP-MASTER-INTERNAL/PRIVATEIP portal.api.simpledemo.onap.org
ONAP-MASTER-INTERNAL/PRIVATEIP sdc.api.fe.simpledemo.onap.org
ONAP-MASTER-INTERNAL/PRIVATEIP sdc.fe.simpledemo.onap.org
ONAP-MASTER-INTERNAL/PRIVATEIP vid.api.simpledemo.onap.org
ONAP-MASTER-INTERNAL/PRIVATEIP sdc.api.fe.simpledemo.onap.org
ONAP-MASTER-INTERNAL/PRIVATEIP portal-sdk.simpledemo.onap.org
ONAP-MASTER-INTERNAL/PRIVATEIP policy.api.simpledemo.onap.org
ONAP-MASTER-INTERNAL/PRIVATEIP aai.api.sparky.simpledemo.onap.org
ONAP-MASTER-INTERNAL/PRIVATEIP cli.api.simpledemo.onap.org
ONAP-MASTER-INTERNAL/PRIVATEIP msb.api.discovery.simpledemo.onap.org
ONAP-MASTER-INTERNAL/PRIVATEIP portal-app.onap

Then try accessing the ONAP Portal from FireFox

https://portal.api.simpledemo.onap.org:30225/ONAPPORTAL/login.htm

Hope this help you. Please follow the document to handle mixed contents as well.

Regards
Vivek

vmuthukrishnan

Please verify the following

• /etc/hosts on the ONAP master VM should have the internal IP of the GCP instance. Please DO NOT USE PUBLIC IP of the VM

Regards
Vivek

vmuthukrishnan

Hi Farah,

You can get the GCP VMs internal IP using the below gcloud CLI

$ gcloud compute instances list

Sample output

NAME INTERNAL_IP EXTERNAL_IP STATUS
anod-master 10.128.0.82 35.192.110.225 RUNNING
anod-worker-01 10.128.0.77 104.197.86.144 RUNNING
anod-worker-02 10.128.0.78 35.226.35.157 RUNNING
anod-worker-03 10.128.0.79 104.155.136.142 RUNNING
anod-worker-04 10.128.0.80 34.69.57.206 RUNNING
anod-worker-05 10.128.0.81 35.225.129.60 RUNNING
anod-worker-06 10.128.0.82 35.225.129.61 RUNNING

Regards
Vivek

farsli1

Thanks a lot for your help!
Finally it is working !!
My mistake was using external IP address of the anod-master to update /etc/hosts instead of using internal IP address.
Steps are a little bit confusing, so I am going to summarize steps to access the portal once this kind of error is encountered:

For GCP users:
-use the internal IP addrees of the anod-master as load balancer IP address to update /etc/hosts
-for firefox: use the proxy configurations as proposed by vmuthukrishnan bellow
-start the socks tunnel from your laptop using the following command