[subscriberdb] Subscribers not showing up after IMSI deconcealing fix (12567)

[vroon2703]

Your Environment

• Version: master
• Affected Component: Access Gateway
• Affected Subcomponent: SubscriberDB
• Deployment Environment: Vagrant (AGW)

Describe the Issue

Possibly related commit: #12567

After adding a new subscriber via swagger and running subscriber_cli.py list on the gateway VM returns no subscribers. The debug logs indicate the following error:

ERROR:root:Fetch suci profiles error! [StatusCode.UNKNOWN] handler panic: runtime error: invalid memory address or nil pointer dereference;

To Reproduce

1. Add a subscriber using Swagger (/lte/{network_id}/subscribers)
2. On the AGW VM, run subscriber_cli.py list
3. Observe that the added subscriber does not show up.
4. Run sudo journalctl -u magma@subscriberdb to see the logs

Expected behavior
IMSI of the added subscriber shows up.

Screenshots
GET shows that IMSI has been added.

subscriber_cli.py list does not print any IMSIs

Error message shows up: fetch suci profiles error!

[Akshayp77]

On latest master code base, we have added a new subscriber(IMSI001010000000004) through swagger API and it is getting reflected only after 5 minutes in the subscriberdb. Please refer attached log file and snapshots.
sys-subdb.log

[sebathomas]

I see the error on my laptop as well and I'm not sure it's caused by the changes from #12567. It seems like it's somewhere in the orc8r, and the PR mostly touched agw code.
It might be the change of the proto file though.

The orc8r subscriberdb panics because of a nil pointer access in this line:

magma/lte/cloud/go/services/subscriberdb/load.go

Line 246 in f9812c6

suciProfiles := ngcConfig.(*lte_models.NetworkNgcConfigs).SuciProfiles

[sebathomas]

Okay, the nil check is broken there, you have to do the type conversion before the check. Will add a PR to fix that so subscriberdb returns an error instead of panicking.

LoadSuciProtos will still fail because ngcConfig is null though.

See #13065.

[panyogesh]

Closing the ticket with reference to #13081