I moved to my new virtual machine on my new XCP-ng virtualization servers I built and I also took my music relay server to the latest version of OS. I am moved off of my c7000 blade server with this update and can shut down this final blade that had my music, playlist, database etc on. The music relay server is what I use to relay my music out via IceCast and ShoutCast to the internet plus securely wrap with ssl certs for a more secure listening stream.

I went from Windows Data Center server 2019 to Windows Data Center 2025 with Mariadb 12.0. I’m running Sam Broadcaster radio automation on that system with those components. That then gets piped to my IceCast and ShoutCast music relay server I’m running Fedora Linux 44 which I just updated to from Fedora 41. I use stunnel to supply the ssl certs to the streams, I may go to nginix or apache proxy later to but the stunnel utility is straightforward so may stick with that.

I had actually built this vm music server about 10 months ago last August in my XCP-ng virtualization environment but had a lot of other things to deal with after and prior to building it that took my time. The VM has been dormant almost that entire time. Upon starting it up recently I realized I had completed the major items on the system last August so I just needed to update the music, pull over the latest database from my physical blade system, work out networking etc. At least that is what I figured.

Always plan for the items you don’t account for. I knew I would have to change my forwarded ports to my new server, I wanted to get my relay server to the latest version, gave the VM more cpu sockets and memory etc.

While updating the ssl certificate on my backup music relay server I used while I updated my main music relay server I noticed a plethora of certificates in the cert directory I hadn’t placed there! After some researching I found this was due to the new openssl handling I need to read up on. I was shocked to see all of those certificates there by default!

I ran into a handful of time consuming issues I didn’t plan like why I suddenly couldn’t login to my music relay server. I found my dhcp server had allocated my relay servers ip address to another device unexpectedly while it was being updated offline from Fedora 41 to 44. The upgrade took barely 20 minutes so this was really strange when I tried to ssh to the server I was rejected and couldn’t login! Finally I went on my router to look at what was connected that ip addres and found the other device, booted it off that ip address, nmap was taking too long to come back to me but would’ve given me the relevant information to resolve that as well .

That brief duplicated ip address caused my stunnel utility to fail without my knowledge of course. This manifested itself through my testing the secure ports to listen to my music on ports 8002 and 8032 failing but playing fine on unsecured ports 8000 and 8030. I used systemd to view the status of stunnel and it was failed so restarted stunnel and then IceCast and ShoutCast and voila I could connect on my secure ports

Then my listeners statistics didn’t update due to a network I gave the VM access to I don’t use often that caused some routing issues so I disabled the NIC to that network. That issue had me checking ports, firewalls, router services until I remembered the behavior of that network as I was testing from a Raspberry Pi that was on that network.

Then the last touch for tonight before badly writing this post was figuring out my ftp issue to update my album artwork which had typos to the picture directories. Observing the resources on the VM with my new more powerful virtualization servers the VM will do just fine as of this moment. I will continue to observe the performance but it’s way below expected cpu, memory and disk utilization.

I’ll try to give a better picture of everything I have done with my new environment when I have time and after I have the new devices racked and the old devices unracked. For now I’m up on my new VM on my new virtualization servers.