I know a High Availability cluster can be setup however I want to run a backup Splynx server on another hosting provider. I’ve been reading up and Proxmox is not recommended to run across a VPN as it really wants <15ms extremely stable latency otherwise servers can flap up and down
So instead is there another way to have a backup Splynx server? Even if its just mirroring the server and then if the hosting provider goes down I can manually move the site-to-site VPN to the backup location? But I don’t know if I can run 2 separate Splynx servers or if that will cause issues with licencing etc
Same situation here.
Our workaround (if HA isnt a option) is at the moment:
– Install a second Splynx Server
– Copy the Splynx Data from /var/www/splynx onto your second server, except encrypt.php
– Delete the Splynx dB on the second Server
– Dump the First Server dB regularly (once a hour for example) into a mysql dump on the second Server
– On the second server, run a script checking if the first server goes down (15min works for me). If it does, import the SQL Dump on the second server in the Splynx dB
– Second Server is now active
– See with the Splynx Team about licencing, as far I understood as only one Server is active at a time, the same License is working.
It should also be possible to have the dB on the second server running/active all the time and instead just kill all its communications to the routers (disable cron, setup Firewall) to avoid that two splynx systems are talking to the same router. Haven’t done it yet though.
Heiko
This is by far the most important required feature for splynx to be at par with enterprise OSS/BSS applications on the market. Support for native active high availability.
I once pointed them in my post Required Native Active-Active Freeradius and MySQL Deployment support for Splynx where I suggested they take a look at severalnines - Deploying an Active-Active FreeRadius Cluster with MySQL NDB or Galera to build such a robost system.
I hope this year they do consider such an architecture.