> librenms.sql","title":"Backing Up"},{"location":"tutorials/librenms/deploy/","text":"Deploying I wrote a script to deploy libreNMS with the configuration that SCN uses. The repo is here Software requirements Only tested on debian and ubuntu. Not sure what else it works on but it could work on other linux distros Steps Install docker if it is not installed already Install docker compose if it is not installed already Instal unzip if it is not installed already Check out this repo If you want to restore a previous install, provide a sqldump named librenms.sql flat in this checked out repo. There is a helper script called get_database_from_currently_running_server.sh to get the database off of the non dockerized install (needs ssh access to the server) If you want to restore the graphs too, you can provide a file named rrd.zip flat in the checked out repo which is just the rrd folder ziped up. There is a helper script called get_rrd_zip_from_currently_running_server.sh to get the rrd zip from the non dockerized install (needs ssh access to the server) Run ./deploy.sh builds the librenms image builds the database image with/without the backup Starts the service using docker compose . This creates 2 shared volumes in the compose directory The librenms folder is for the librenms docker images to share configuration data including rrd files The db volume is the database unzips the rrd folder in the rrd directory of the shared librenms volume The UI will run on port 8000","title":"Deploying"},{"location":"tutorials/librenms/deploy/#deploying","text":"I wrote a script to deploy libreNMS with the configuration that SCN uses. The repo is here","title":"Deploying"},{"location":"tutorials/librenms/deploy/#software-requirements","text":"Only tested on debian and ubuntu. Not sure what else it works on but it could work on other linux distros","title":"Software requirements"},{"location":"tutorials/librenms/deploy/#steps","text":"Install docker if it is not installed already Install docker compose if it is not installed already Instal unzip if it is not installed already Check out this repo If you want to restore a previous install, provide a sqldump named librenms.sql flat in this checked out repo. There is a helper script called get_database_from_currently_running_server.sh to get the database off of the non dockerized install (needs ssh access to the server) If you want to restore the graphs too, you can provide a file named rrd.zip flat in the checked out repo which is just the rrd folder ziped up. There is a helper script called get_rrd_zip_from_currently_running_server.sh to get the rrd zip from the non dockerized install (needs ssh access to the server) Run ./deploy.sh builds the librenms image builds the database image with/without the backup Starts the service using docker compose . This creates 2 shared volumes in the compose directory The librenms folder is for the librenms docker images to share configuration data including rrd files The db volume is the database unzips the rrd folder in the rrd directory of the shared librenms volume The UI will run on port 8000","title":"Steps"},{"location":"tutorials/librenms/upgrade/","text":"Upgrading Upgrading can be segmented into 2 parts. The container OSses and the service OS Each container runs an OS - The Maria db container is based on ubuntu, so you can just do sudo apt update and sudo apt upgrade when executing bash on the container - The Redis container for some reason only has an ash executable installed on the container. Also it runs on alpine so it can be updated using apk update and apk upgrade - The libreNMS and the libreNMS dispatcher container is instantiaated on the same container which is on alpine, and uses bash. Update as usual for alpine installs. Service I think that libreNMS vends a script called daily.sh (librenms docs here ) that is added to the cron, but cron is not running on docker containers since each container only runs one process. Even if we manually run daily.sh, there are errors. I think that the docs also give a manually manual way to update, by doing a git clone, but since the librenms files were not pulled using git, we cant use this way. I tried using rsync to overwrite the old files with the new files, but there are some issues. The nuclear option can be used, which is remove the containers, build new updated ones, and start that, but this will include a small outage Go to the compose directory and run sudo docker compose down to stop and remove all the containers. We store data (rrd files and the database) in a docker volume inside the compose directory anyway so we should not need to worry about removing containers removing any data Go to the librenms_image directory, change the version of the image to the latest version here run sudo docker build . -t scn-librenms , which should build the new image go to the db_image directory and update the Dockerfile's version to the latest version here run sudo docker build . -t scn_mariadb_librenms go to the compose directory and update the compose.yml file to use the latest redis release here Then you should be able to start the service with a sudo docker compose -f compose/compose.yml up -d The service should come up as it was before. If it does not, you may have to do a ./lnms migrate","title":"Upgrading"},{"location":"tutorials/librenms/upgrade/#upgrading","text":"Upgrading can be segmented into 2 parts. The container OSses and the service","title":"Upgrading"},{"location":"tutorials/librenms/upgrade/#os","text":"Each container runs an OS - The Maria db container is based on ubuntu, so you can just do sudo apt update and sudo apt upgrade when executing bash on the container - The Redis container for some reason only has an ash executable installed on the container. Also it runs on alpine so it can be updated using apk update and apk upgrade - The libreNMS and the libreNMS dispatcher container is instantiaated on the same container which is on alpine, and uses bash. Update as usual for alpine installs.","title":"OS"},{"location":"tutorials/librenms/upgrade/#service","text":"I think that libreNMS vends a script called daily.sh (librenms docs here ) that is added to the cron, but cron is not running on docker containers since each container only runs one process. Even if we manually run daily.sh, there are errors. I think that the docs also give a manually manual way to update, by doing a git clone, but since the librenms files were not pulled using git, we cant use this way. I tried using rsync to overwrite the old files with the new files, but there are some issues. The nuclear option can be used, which is remove the containers, build new updated ones, and start that, but this will include a small outage Go to the compose directory and run sudo docker compose down to stop and remove all the containers. We store data (rrd files and the database) in a docker volume inside the compose directory anyway so we should not need to worry about removing containers removing any data Go to the librenms_image directory, change the version of the image to the latest version here run sudo docker build . -t scn-librenms , which should build the new image go to the db_image directory and update the Dockerfile's version to the latest version here run sudo docker build . -t scn_mariadb_librenms go to the compose directory and update the compose.yml file to use the latest redis release here Then you should be able to start the service with a sudo docker compose -f compose/compose.yml up -d The service should come up as it was before. If it does not, you may have to do a ./lnms migrate","title":"Service"}]}
\ No newline at end of file
diff --git a/tutorials/epc-setup/index.html b/tutorials/epc-setup/index.html
index 87e2cd5..53d3f5c 100644
--- a/tutorials/epc-setup/index.html
+++ b/tutorials/epc-setup/index.html
@@ -167,7 +167,7 @@ Operating System Support
In SCN we will typically perform these installation steps using a fresh install of Ubuntu 22.04 on an x86-64-based computer; however, any operating system that open5gs supports should work.
Note: When you're installing Ubuntu, we suggest choosing the "minimal install" option that doesn’t install extra unnecessary software. In prior installs this has led to version conflicts.
Software Components
-As of August 2021, in the Open5GS software package, the LTE-specific components (which run on Ubuntu as systemd services) are as follows:
+As of November 2024, in the Open5GS software package, the LTE-specific components (which run on Ubuntu as systemd services) are as follows:
- MME - Mobility Management Entity:
open5gs-mmed.service
- HSS - Home Subscriber Server:
open5gs-hssd.service
@@ -187,14 +187,17 @@ Step 1: Open5GS Install Notes
To do this, first install MongoDB as described in the Quickstart guide. Then follow instructions under the "Ubuntu" section to install Open5GS via apt.
Note: If installing over a ssh connection, we recommend using tmux or another program in case you get disconnected from the session in the process.
-Remember to follow all the network configuration steps in the Open5GS Quickstart documentation. For SCN's Ubuntu machines, this means:
-- Allowing IP forwarding on your machine, e.g. via the following command:
+Remember to follow all the network configuration steps in the Open5GS Quickstart documentation. For SCN's Ubuntu machines, this means:
+
+- Allowing IP forwarding on your machine, e.g. via the following command:
+
sudo sysctl -w net.ipv4.ip_forward=1
-- Setting up NAT rules using
iptables so that traffic from the eNB can reach the Internetand vice versa
- Using Netplan to configure network interfaces with IP addresses in the desired way.
+- Setting up NAT rules using
iptables so that traffic from the eNB can reach the Internet and vice versa
+The latter two steps are explained in detail below.
Netplan Configuration
A. Recommended
For this recommended configuration, we require an EPC
@@ -205,7 +208,7 @@ A. Recommended
line. A list of network interfaces will appear in the terminal. Find the ones
corresponding to your ethernet ports (their names usually start with “eth,”
“enp,” or “enx”).
-For Ubuntu 20.04, we're currently using the Netplan program to manage our network configuration.
+
For Ubuntu 22.04, we're currently using the Netplan program to manage our network configuration.
Create a file in the /etc/netplan directory (i.e. a folder) named
99-open5gs-config.yaml, and add the following lines, substituting the correct
interface names and subnets for your configuration:
