Frequently Asked Questions

This section lists some of the commonly asked questions related to Magma operation.

Access Gateway

How can I check the status of eNodeB(s) connected on a particular gateway?

Run sudo enodebd_cli.py get_all_status to check eNodeB(s) status.
NOTE: Output of this command will only show eNodeB(s) configured with TR069 protocol.

How can I see different interfaces of gateway?

Run sudo ip addr to see all configured interfaces.
Commonly, interface eth0 is used for internet connectivity and eth1 is used for RAN connectivity.
You may run ping google.com -I eth0 to check AGW connectivity with internet.
Also, you may run ping <eNodeB IP> -I eth1 to check AGW connectivity with eNodeB. You can get eNodeB IP from enodeb_cli as mentioned in above question.

How can I see attached subscribers list on a particular gateway and their browsing status?

Run sudo mobility_cli.py get_subscriber_table, to see all attached subscribers.
Note down the IP obtained by the user from above command, then run sudo pipelined_cli.py debug display_flows | grep <USER IP> for packet flow info of that user.

How can I check services running on a gateway and their status?

Check all services: sudo service magma@* status.
Check specific service: sudo service magma@<SERVICE_NAME> status. e.g. sudo service magma@enodebd status.

How can I collect logs for a particular service?

Run sudo journalctl -fu magma@<SERVICE NAME>, e.g. for enodeb service, run sudo journalctl -fu magma@enodebd.

How can I check the status of the OVS module on a gateway?

Run sudo ovs-vsctl show to check.

How can I capture packets for troubleshooting?

You can use general purpose tcpdump command to capture traffic at any of your interfaces. e.g. run sudo tcpdump -i eth1 -w /tmp/capture.pcap to capture packet at eth1 interface.

What are the common troubleshooting logs in gateway?

syslog: /var/log/syslogs.
MME logs: /var/log/mme.log.
eNodeB logs: /var/log/enodebd.log.

Where can I find the configuration of a gateway?

File /var/opt/magma/configs/gateway.mconfig is the configuration streamed down from Orchestrator.
Any custom static configs will also be in directory /var/opt/magma/configs with a .yml extension.
Default static configs are located at /etc/magma/configs.
systemd unit files are located at /etc/system/systemd.

How can I restart gateway services?

There are 2 options to restart gateway services. First from CLI and other from NMS UI.
To restart services from CLI, login to AGW node, then perform below steps:
- Run sudo service magma@* stop to stop services.
- Then, run sudo service magma@magmad restart to restart services.
To restart services from NMS UI, perform below steps:
- Login to NMS UI and go to Gateways option from left hand pane menu.
- Then, click on edit option against the gateway for which you wish to reboot services.
- From the popup window, select tab COMMANDS. Then, check section Reboot.
- Click on Restart Services to restart all AGW services.

How can I reboot the gateway?

There are 2 options to reboot gateway. First from CLI and other from NMS UI.
To reboot AGW from CLI, login to AGW node, then run sudo init 6 to reboot whole AGW node.
To reboot AGW from NMS UI, perform below steps:
- Login to NMS UI and go to Gateways option from left hand pane menu.
- Then, click on edit option against the gateway which you wish to reboot.
- From the popup window, select tab COMMANDS. Then, check section Reboot.
- Click on Reboot to reboot the AGW node.

Does the AGW support IPv6?

IPv6 and dual stack IPv4v6 session support is implemented for Magma v1.7 and later.

How to change hardware UUID of AGW?

Hardware UUID is in /etc/snowflake.
Delete file /etc/snowflake and restart magma@magmad systemd service, which will create new one.
NOTE If you reinstall magma, the UUID hardware ID will change.

How to switch between info/debug mode in mme logs?

Modify /var/opt/magma/configs/mme.yml and add the line "log_level: DEBUG",
Restart the mme service with sudo service magma@mme restart.
NOTE Expect impact of service while doing above procedure.

What's the command to flush the redis state?

The command in AGWredis-cli -p 6380 FLUSHALL will flush the redis state

Where can I get more information about services running in Magma?

https://github.com/magma/magma/tree/master/docs/readmes/lte

How can I detach/purge a subscriber from AGW?

The command s6a_service_cli.py delete <IMSI>,<IMSI> will detach the list of subscriber from MME.
Detached subscribers will need to reattach to the network.

How to change MME IP address?

Edit the IP address of eth1 interface: sudo vim /etc/network/interfaces.d/eth1
Then restart network manager: sudo systemctl restart networking
And check IP again: ip a s eth1

How can I check if a given subscriber is idle or connected?

SSH into the the AGW
Fetch the subscriber's state using its IMSI:
- state_cli.py parse IMSI716770120257790:MME
Looking at the ecm_state configuration, the value:
- 0 - Indicates the user is IDLE
- 1 - Indicates the user is CONNECTED

Orchestrator

How can I check production pods in Orchestrator running on Kubernetes?

For all pods: kubectl --namespace orc8r get pods.
Logs of a particular pod: kubectl --namespace orc8r logs -f <CONTROLLER PODNAME>.
NOTE: In above commands, please use the actual namespace after --namespace you have chosen to install orc8r in.

How can I test Orchestrator connection with Gateway?

Where can I find API endpoint?

If you have followed the install guide, it'll be at https://api.youdomain.com/swagger/v1/ui/.
You will need to load the admin_operator.pfx certificate into your keychain/browser, otherwise, API access will be blocked.

How can I use Swagger UI to trigger API request?

Open Swagger UI, then got to interested section, e.g. EnodeBs.
Then click on API trigger action button e.g. GET, PUT, DELETE etc.
Click on Try it out button on right hand side.
Put in the required inputs and click Execute.

How can I check the services running in Orchestrator?

List the running pods with kubectl --namespace orc8r get pods
Grab the name of orc8r-controller pods, they are in the format orc8r-controller-xxxxxxxxxx-yyyyy
Collect the state of services on each pod: kubectl --namespace orc8r exec orc8r-controller-xxxxxxxxxx-yyyyy bash -- supervisorctl status

How to verify which services are running in Orc8r-controller?

Get the controller pods: kubectl --namespace orc8r get pods.
Execute the command in the pod: kubectl --namespace orc8r exec <CONTROLLER PODNAME> supervisorctl status.

NMS

What is an NMS (Network Management System)?

NMS is a simple UI based solution provided to manage, configuring, and monitor the network effectively. You can monitor various network metrics, configure and monitor alerts, subscribers, gateways, eNodeBs etc.

How can I check various network metrics over NMS?

How can I configure an alert on NMS?

How can I make sure Gateway is properly connected with NMS?

Login to NMS and select Gateway from left hand side menu bar.
You can see all the Gateways configured in NMS.
Check for the Green light besides any particular Gateway, if it is green that means it is properly connected.

How can I restart eNodeB from NMS?

Login to NMS UI and go to Gateways option from left hand pane menu.
Then, click on edit option against the gateway with which eNodeB is connected.
From the popup window, select tab COMMANDS. Then, check section Reboot eNodeB.
Enter eNodeB serial ID of eNodeB and click on Reboot to reboot that particular eNodeB.
Note:Only those eNodeBs can be rebooted in this manner which are configured with TR069 protocol.

How to enable events and log aggregation?

Network Management → Equipment → Gateways → Select the Gateway → Config → Aggregations → Edit → Enable Log and Even Aggregations.
Make sure fluentd_address and fluentd_port is added in control_proxy.yml on AGW.

How to enable pre-defined alerts on NMS?

Administrative Tools → Alerts → Select Network → Sync Alert.

Miscellaneous

How to disable dhcp in AGW for eNB?

Network Management → Equipment → Gateways → Select the Gateway → Config → RAN → Edit → Disable eNodeB DHCP service.

What is a core dump?

When AGW services crash with a segmentation fault, a core dump is generated.
Core dumps are generated in format core-<timestamp>-<process_name>[-<PID>] under /tmp folder.
Use GDB tool to analyse them as cd /tmp/<core-directory>/ gunzip <core gzip file> gdb /usr/local/bin/<process name> <unzipped core file>.