A personal snippets to run distributed LibreNMS poller on low-end server resources to monitor hundreds of devices and sensors.
A normal LibreNMS install contains all parts of LibreNMS: - Poller/Discovery workers - RRD (Time series data store) - Database - Webserver (Web UI/API)
Devices monitored with LibreNMS can be grouped together into a
poller_group to pin these devices to a single or a group of designated pollers.
Distributed Polling allows the workers to be spread across additional servers for horizontal scaling. All pollers need to write to the same set of RRD files, shared NFS storage can be used, but I prefer using RRDcached.
It is a required that all pollers can access the central memcached to communicate with each other.
Actually, the distributed poller setup is very dynamic and can be configured to fit your need. Below are my current setup on what you need to consider both from the software layer but also connectivity.
|Node||OS||RAM - CPU||Services||Devices|
|nms82||Ubuntu 18.4||4GB - 4 Core||Web/API, MariaDB, RRDcached, Redis, Memcached, Syslog||< 9|
|nms83||Ubuntu 18.4||4GB - 4 Core||Poller, Oxidized Config Backup||62|
|nms84||Ubuntu 18.4||4GB - 4 Core||Poller, Discovery Module||19|
|bsd-007-c7||CentOS 7||2GB - 4 Core||Poller||128|
All machines above is run under Proxmox Linux Container across different location.
For now, this resource (4 core Xeon and 4GB RAM each node) seem to be overkill to monitor about 200 devices, 1800 ports and 1000 sensors including processor, storages, applications, and disk I/O. The master host (nms82) only use about 1.5% of CPU resource and 600 - 1.200 MiB of RAM.
But you may be need to switch from 7,000 RPM HDD to 15,000 RPM SAS drive or SSD for better performance since disk write is pretty high.
All this setup need a working LibreNMS install, plus:
- rrdtool version 1.7.0 above
- php-memcached module
- a memcached server (nms82)
- a rrdcached server (nms82)
- a redis server (nms82)
- Python 3 python-memcached
- python-dotenv .env loader
- redis-py 3.0+
These can be obtained from OS package manager, or from PyPI with command below (under librenms root directory):
pip3 install -r requirements.txt
Master Host (Web UI, API, MariaDB, Redis, RRDcached, Memcached) (nms82)
This node become one of most important part which contain additional required server running on this node: Web UI / API, Memcached, RRDcached and Redis server.
Web / API Layer
Based on topology above, the web / API layer will respond to Nginx reverse proxy. This is typically Apache but I prefer using Nginx, here some Nginx backend configuration example for LibreNMS.
The pollers, web and API layers should all be able to access the database server directly. Make sure that MariaDB listen to server IP address that can be accessed from other nodes instead of unix socket.
Additionally, you can use stable version of official MariaDB repo instead of distribution repo.
RRD Storage (via RRDCached)
It is advisable to run RRDCached within this setup so that you don't need to share the rrd folder via a remote file share such as NFS. The web service can then generate rrd graphs via RRDCached.
For this example, I'm running RRDCached to allow all pollers and web/api servers to read/write to the rrd files with the rrd directory.
apt install rrdcached
DAEMON=/usr/bin/rrdcached WRITE_THREADS=4 BASE_PATH=/opt/librenms/rrd/ JOURNAL_PATH=/var/lib/rrdcached/journal/ PIDFILE=/var/run/rrdcached.pid SOCKFILE=/var/run/rrdcached.sock SOCKGROUP=librenms DAEMON_GROUP=librenms DAEMON_USER=librenms NETWORK_OPTIONS="-L" BASE_OPTIONS="-B -F -R"
Change ownership of journal path to librenms user:
chown librenms:librenms /var/lib/rrdcached/journal/
Memcache is required for the distributed pollers to be able to register to a central location and record what devices are polled. Memcache can run from any of the servers so long as it is accessible by all pollers.
Example memcached config on ubuntu
# memcached default config file # 2003 - Jay Bonci <[email protected]> # This configuration file is read by the start-memcached script provided as # part of the Debian GNU/Linux distribution. # Run memcached as a daemon. This command is implied, and is not needed for the # daemon to run. See the README.Debian that comes with this package for more # information. -d # Log memcached's output to /var/log/memcached logfile /var/log/memcached.log # Be verbose # -v # Be even more verbose (print client commands as well) # -vv # Start with a cap of 64 megs of memory. It's reasonable, and the daemon default # Note that the daemon will grow to this size, but does not start out holding this much # memory -m 64 # Default connection port is 11211 -p 11211 # Run the daemon as root. The start-memcached will default to running as root if no # -u command is present in this config file -u memcache # Specify which IP address to listen on. The default is to listen on all IP addresses # This parameter is one of the only security measures that memcached has, so make sure # it's listening on a firewalled interface. #-l 127.0.0.1 -l 0.0.0.0 # Limit the number of simultaneous incoming connections. The daemon default is 1024 # -c 1024 # Lock down all paged memory. Consult with the README and homepage before you do this # -k # Return error when memory is exhausted (rather than removing items) # -M # Maximize core file limit # -r # Use a pidfile -P /var/run/memcached/memcached.pid
Note: You also need php-memcached module
Don't forget to configure server firewall so, only selecred server and poller can access the memcached server
Redis instance needed to coordinate the nodes. It's recommended that you do not share the Redis database with any other system. By default, Redis supports up to 16 databases (numbered 0-15). On this topic, I use database number 4 for LibreNMS and use
chris-lea PPA for Redis 3+ on Ubuntu:
add-apt-repository ppa:chris-lea/redis-server apt install redis-server
Edit Ubuntu Redis configuration example: (
bind 0.0.0.0 requirepass YOURSCRETREDISPASSWORD
It's strongly recommended that you deploy a resilient cluster of redis systems, and use redis-sentinel.
Poller Hosts (nms83, nms84, bsd-006-c7)
All poller host need working LibreNMS install. All you need to do after that is reconfigure all poller host config connection (Redis, Database, RRDCached, Memcached) to connect to Master Host (nms82).
Connection Configuration on All Nodes
Once all required dependency is installed and running on master hosts, reconfigure it in the .env and config.php file on each node.
Connection settings are required in
.env file is generated after composer install and APP_KEY and NODE_ID are set. Remember that the APP_KEY value must be the same on all pollers and leave NODE_ID as it.
#APP_KEY= #Required, generated by composer install DB_HOST=[YOUR_DATABASE_HOST] DB_DATABASE=[YOUR_DATABASE_NAME] DB_USERNAME=[YOUR_DATABASE_USER] DB_PASSWORD=[YOUR_DATABASE_PASS] #NODE_ID= #Required, generated by composer install LIBRENMS_USER=librenms # Distributed Polling REDIS_HOST=[YOUR_REDIS_HOST_IP] REDIS_PORT=6379 REDIS_PASSWORD=[YOURSCRETREDISPASSWORD] REDIS_DB=4 CACHE_DRIVER=redis
config.php under LibreNMS directory :
<?php $config['distributed_poller'] = true; $config['distributed_poller_name'] = php_uname('n'); $config['distributed_poller_group'] = '0'; # Set this to your prefered group for each node $config['distributed_poller_memcached_host'] = "[IP_ADDR_OF_MEMCACHED_SERVER]"; $config['distributed_poller_memcached_port'] = 11211; $config['service_poller_workers'] = 24; # Processes spawned for polling $config['service_services_workers'] = 8; # Processes spawned for service polling $config['service_discovery_workers'] = 16; # Processes spawned for discovery //Optional Settings #$config['service_poller_frequency'] = 300; # Seconds between polling attempts #$config['service_services_frequency'] = 300; # Seconds between service polling attempts #$config['service_discovery_frequency'] = 21600; # Seconds between discovery runs #$config['service_billing_frequency'] = 300; # Seconds between billing calculations #$config['service_billing_calculate_frequency'] = 60; # Billing interval #$config['service_poller_down_retry'] = 60; # Seconds between failed polling attempts #$config['service_loglevel'] = 'WARNING'; # Must be one of 'DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL' #$config['service_update_frequency'] = 86400; # Seconds between LibreNMS update checks $config['service_watchdog_enabled'] = true;
LibreNMS systemd Dispatcher Service Installation
A systemd unit file is provided - the sysv and upstart init scripts could also be used with a little modification.
A systemd unit file can be found in
misc/librenms.service. To install run
cp /opt/librenms/misc/librenms.service /etc/systemd/system/librenms.service && systemctl enable --now librenms.service