[Spce-user] EC2 AMI network issues

[Michael Prokop]

Hi,

* Nogie Ekunwe [Mon Feb 19, 2024 at 12:06:09PM -0600]:
> > On Feb 19, 2024, at 11:00 AM:
> > * Edo [Sat Feb 17, 2024 at 03:22:35AM -0600]:

> >> I have initiated EC2 instances (2) and discovered that the sipwise ami (
> >> ami-06ef81cd274aafae9
> >> <https://eu-west-1.console.aws.amazon.com/ec2/home?region=eu-west-1#ImageDetails:imageId=ami-06ef81cd274aafae9>)
> >> does not start or configure the network card when used to start an ec2
> >> instance. I tried 2 sipwise ami's, both have the same symptom. Am I missing
> >> something?

> > Thanks for reporting, we'll try to reproduce and look into this.
> > JFTR, which instance type did you use?

> I used the most current AMI and also the most current LTS AMI

We just released a new AMI for mr12.2.1 which fixes the reported
issue. For now you can find it as AMI ami-0bd948c18a837ccba in
eu-west-1 region.

After launching it, you need to briefly adjust the
/etc/ngcp-config/network.yml though, by removing the duplicate
interface configuration (it will contain configuration for ens5,
while the running network interface is neth0).

Fix it either manually via:

  sudo $EDITOR /etc/ngcp-config/network.yml
  sudo ngcpcfg apply 'fix network configuration'

... or on the command line this should do the trick:

  sudo ngcpcfg del /etc/ngcp-config/network.yml hosts.self.ens5
  sudo ngcpcfg del /etc/ngcp-config/network.yml hosts.self.interfaces
  sudo ngcpcfg set /etc/ngcp-config/network.yml hosts.self.interfaces="['lo', 'neth0']"
  sudo ngcpcfg apply 'fix network configuration'

We will work on fine-tuning this, so that workaround shouldn't be
needed in the future.

regards
-mika-