Closed Zhenye-Na closed 1 year ago
Added more findings:
It seems like the NRF chart works when:
If there are multiple instances, and the source of them is mix-matching. The NRF helm chart is installed and MongoDB is running, the wait-mongo
container of NRF
could not find the the Mongo pods, with same logs found above
kubectl logs nrf-free5gc-nrf-nrf-79d7d9b788-cmghx -c wait-mongo
nc: bad address 'mongodb:27017'
waiting for the MongoDB
nc: bad address 'mongodb:27017'
waiting for the MongoDB
In AWS EKS Console, I am able to see that mongodb-0
and nrf-free5gc-nrf-nrf-79d7d9b788-cmghx
are installed in the same EC2 instances, resources such as memory, total number of pods or IPs are not running up.
So I would really appreciate it if any tips or findings can be provided.
Thanks!!
Closed this issue since it is not related to specific helm chart implementation.
For anyone who is curious, this issue is related to DNS lookup issue that is similar to https://github.com/aws/aws-cdk/issues/10884
I am using
helm
to installfree5gc-nrf
chart. The MongoDB dependency could be instantiated but nrf is waiting for MongoDB forever. But I verified that mongodb is up. So not sure how does this happen.If I checked the log of
wait-mongo
container ofnrf-free5gc-nrf-nrf-79d7d9b788-cmghx
Meanwhile I checked the log of MongoDB-0 pod