Hey @BergeP,
Thanks for sharing your report on the latest update, and sorry for the frustration. You’re right that nothing changed on your end, and the logs back that up.
The cause is upstream of Roon Server, on your network. Two devices — 192.168.1.140 and 192.168.1.131, are advertising themselves as RAATServer endpoints, accepting the connection from your Nucleus One, and then immediately dropping it. They’ve done this 226,871 times in the ~20 hours of logs we’ve reviewed, at up to 135 times per second. That churn floods Roon Server with device-state updates it has to buffer for your remotes, and rebuilding those buffers is what drives the memory spike that freezes everything. Neither device ever finishes handshaking, so neither shows up as a usable zone, they’re producing nothing but load.
One reassuring thing: your Nucleus One never actually ran out of memory. There’s no sign of the operating system killing Roon for memory exhaustion anywhere in those logs. What’s happening is that the server gets so busy cleaning up after this loop that it stops responding, and the watchdog restarts it. That’s a meaningful difference; it points at the loop, not at your hardware being undersized.
The one thing that would help most, could you identify what’s at 192.168.1.140 and 192.168.1.131? There’s a third at 192.168.1.200 that’s behaving normally, so it may be a sibling of those two.
If you can power those two down and let Roon run for an hour or two, I’d expect the stalls, the spinner and the endpoint dropouts to stop. That single test will confirm the diagnosis. If one of them is a Docker install, it’s also worth checking it’s set to host networking, bridged networking is a very common cause of this exact “connects then instantly drops” pattern.
On the Qobuz albums: that’s a separate problem, and it is on our side, not yours. Your Qobuz sync is running correctly and Roon is receiving your new favourites, I can see your library go from 5,319 to 5,320 albums during the logs. But the step that publishes them into your library is gated behind a periodic full metadata refresh that last completed on 31 July and hasn’t run since. There are currently 70,129 items queued behind it. So the albums are arriving and then sitting in a queue. We’ll need to look into this further.
Two smaller things worth mentioning:
- Your queue was 7,999 tracks at the time of the stall, not 140, that’s Roon’s maximum queue length, and shuffle was on. A queue that size is significantly more expensive for the server to hold and to re-send to your remotes each time one reconnects. Clearing it down won’t fix the root cause, but it will take real pressure off while we sort the rest out.
- Your Samsung T7 SSD is showing as offline to Roon (
DriveNotReady). If that’s meant to be part of your library, let me know and we’ll look at it separately.
Let me know what those two IPs turn out to be and we’ll go from there. 