Questions Regarding Distributed Poller during Downtime of Poller Master

[Karl_Lim]

Hi I am new in using Librenms and I wanted to know what would happen if the Poller Master ( one with database and has rrd and rrdcached) got down. What would happened to the other pollers ? would they still do checks on the devices that are assigned to them ? If yes then what would happen to the data that they gathered will it be sent to the Poller Master once it is up and running again ?

Topology:

Scenario:
-Poller Master got down (network failure)
-Poller1, Poller2, and Poller3 cannot reach Poller Master

Question:

1. How does Poller1, Poller2, Poller3 tolerate loss of connectivity to master without data loss ?
2. Can Poller1, Poller2, Poller3 still run checks on the Devices assigned to them ?
3. If Yes then where do they store that data they gathered ?
4. If Yes then Once the Poller Master has been up and running would they be able to send that gathered data to the Poler Master to catch up the data loss during the Downtime of the Poller Master ?
5. If No, would you be able to recover that loss data ? What other alternatives would prevent the loss of data during the downtime of Master Poller.

Thank you

Regards,
Karl

[Kevin_Krumm]

Where is your Data base and RRD files stored

[Karl_Lim]

Hi Kevin,

Sorry for the late reply, the Database and RRD files are stored in the Poller Master. Thank you

Regards,
Karl

[Kevin_Krumm]

Distributed poller is meant to balance the load of polling. It’s not meant for redundancy.
If the pollers can’t talk back the database and the RRD files. then no data.

You can setup Galera for database redundancy Galera Database Cluster - LibreNMS Docs

But for the RRD files, I’m not sure how you can make that redundant.