Hi @Geoff_Bishop,
Thanks for your extended patience here - the Qobuz playback issue has largely been resolved, but there may be a few lingering tracks that cause you issues.
Qobuz aside, a fresh Nucleus diagnostic report has shown the following:
The Naim Mu-so (Kitchen, 192.168.0.224) is the root cause of grouped playback failures
The single most important pattern in the logs: the grouped zone Music Room + Kitchen is destroyed and recreated 35+ times on April 7 alone, every 10–30 minutes. Each cycle follows an identical sequence:
- The Mu-so at
192.168.0.224drops off the network (No route to hostor port unreachable) - The grouped zone is torn down
- Roon rebuilds the zone with just the Uniti Star, then tries to add the Mu-so back
- Playback resumes briefly, then the cycle repeats
The [rnet/RnetJsonClient] failed to connect No route to host warnings consistently reference 192.168.0.224 (the Mu-so’s address), and they appear immediately before zone destructions. This Rnet connection is how Roon communicates device state — when it fails, Roon tears down the zone.
Question for you - What switch port does the Mu-so plug into, and what switch? The Mu-so’s network interface or the switch port feeding it may be periodically going down/coming back (EEE/green-mode link drop, faulty cable, or a switch port auto-negotiation issue).
With that:
- Try a different Ethernet port on the switch
- Try a different Ethernet cable to the Mu-so
- Log into the Mu-so's web interface and check if it shows any link/network errors
- Consider assigning the Mu-so a static IP in the router to rule out DHCP lease churn
Alongside this, we also found that the DESKTOP zone (on a Windows PC at 192.168.0.249) is being created and destroyed hundreds of times per day, every day. On April 15 alone: 1,044 create/destroy cycles.
This is almost certainly a PC with Roon installed that goes into sleep/screensaver/power-saving mode repeatedly. While this is not a direct cause of audio dropouts on the Naim endpoints, it generates continuous zone management traffic on the Nucleus, thousands of transport layer events per day, and contributes to both memory pressure and GC load. This is a significant background stressor that should be eliminated.
Thanks, Geoff! ![]()