ORAN Slicing Experiment Error

27 views
Skip to first unread message

Nick Accurso

unread,
Apr 17, 2023, 2:26:35 PM4/17/23
to Powder Users
Hello POWDER team,

I am running into an error when trying to execute the ORAN slicing experiment outlined in the "O-RAN" profile. I am just beginning to use the POWDER platform, so apologies if there is a very simple solution.

I am able to launch the EPC and eNB without issue, and can launch a UE which successfully connects with the eNB. It also appears I am able to onboard and run the nexRAN xApp without issue. However, I am unable to communicate with the xApp using the provided HTTP GET request provided in the profile, as I get a "connection refused" error. A screenshot of this error is provided here.

I should also note that I have tried this with both dawn and e-release RIC releases. Thank you in advance for your help.

Best,
Nick AccursoScreenshot (14).png

David M. Johnson

unread,
Apr 21, 2023, 9:46:20 AM4/21/23
to powder...@googlegroups.com
On 4/17/23 12:26, Nick Accurso wrote:
> Hello POWDER team,
>
> I am running into an error when trying to execute the ORAN slicing
> experiment outlined in the "O-RAN" profile. I am just beginning to use the
> POWDER platform, so apologies if there is a very simple solution.
>
> I am able to launch the EPC and eNB without issue, and can launch a UE
> which successfully connects with the eNB. It also appears I am able to
> onboard and run the nexRAN xApp without issue. However, I am unable to
> communicate with the xApp using the provided HTTP GET request provided in
> the profile, as I get a "connection refused" error. A screenshot of this
> error is provided here.

Sorry for the delay responding. I recently modified the xapp to pend
the initialization of its admin northbound interface until after the
xapp is registered with the RIC, and it now implements a separate
RESTful HTTP interface (the standard RIC endpoints, healthchecks, etc)
which *is* enabled immediately at startup
(https://gitlab.flux.utah.edu/powderrenewpublic/nexran/-/commit/5438bebba15b863e62301dfd0ec78b36bac56866).
Therefore, it is possible that the admin NBI is not yet up (it can
take several seconds for the xApp's RMR routes and connections to RMR
endpoints to be configured). You can do

export NEXRAN_RIC_URL=http://`kubectl -n ricxapp get
services/service-ricxapp-nexran-http -o
jsonpath='{.spec.clusterIP}:{.spec.ports[0].port}'`/ric/v1
curl -i $NEXRAN_RIC_URL/health/ready

and you will get a 200 OK when the xApp is fully attached to the RMR
messaging layer (and therefore ready to interact with the RIC).

If that doesn't help, you will want to take a look at the pod status and
logfile. `kubectl -n ricxapp get pods -l app=ricxapp-nexran`, `kubectl
-n ricxapp logs -f -l app=ricxapp-nexran`. If there are any ERROR or
WARNING messages, or if the pod is in a crash loop, please send along
the tail of the logs.

> I should also note that I have tried this with both dawn and e-release RIC
> releases. Thank you in advance for your help.

I think at this point, your best bet is the f-release option. There is
a g-release option, but it includes the upgrade to influx 3, which
nexran isn't fully ready for yet. I expect the g-release to be
available in the coming weeks.

> Nick Accurso[image: Screenshot (14).png]

David

Nick Accurso

unread,
Apr 24, 2023, 1:17:17 PM4/24/23
to Powder Users
Hello David,

Thank you for your reply. I see that the f-release option is working correctly, so we will proceed with this.

Thanks,
Nick
Reply all
Reply to author
Forward
0 new messages