Hey @david_klipper,
Thanks for all the updates, and I’m sorry to hear you’re still suffering from issues.
If there are any specific steps that you weren’t able to perform, certainly let us know. This would be helpful to know and address.
We were able to review another fresh diagnostic report from your Roon Server, and see that the retry queue was at 4,193 albums in the earlier logs analyzed. In the new logs (April 19–21), it dropped to 3,607 albums — meaning roughly 586 albums actually processed and cleared. Then, on April 21–22 it jumped back up to 4,161 albums. This almost certainly corresponds to the restart that happened at April 21, 6:53 PM.
Every restart re-enqueues the full library and adds newly-failed identifications back to the retry pool. The good news is the queue is slowly shrinking between restarts when left alone.
The spinner will persist until either: (a) those ~4,000+ live recordings either get identified or are explicitly set to “prefer file” metadata via the import settings in Roon, or (b) you let the system run without restarting until the backoff timer finally retries and resolves them.
For your Issue 2: Playback Dropouts
There are four separate snd_rpi_hifiberry_dacplushd clock connections in the logs. You mentioned you have 3 HiFiBerry DAC+ HD boards, the logs show all of them are in the same grouped zone (“Office + Bedroom + Garage + Rendu”), and they are all exhibiting severe clock instability.
The logs show:
- Routine readings of 100–200+ ppm on the DAC+ HD boards
- Frequent spikes of 1,000–6,800 ppm
- Brief extreme spikes of up to 41,585 ppm (this is roughly 150 seconds of drift per hour, quite unstable)
The logs show 10 “OnPlayFeedback Stopped/Paused” events today alone (April 22), with clustered stops around 9:09 AM, 10:45 AM, 11:26 AM, and 11:31 AM. These correlate directly with the long rtt sync warnings on the DAC+ HD endpoints. When the clock drifts far enough, Roon cannot keep the grouped zone synchronized and kills playback.
Some next troubleshooting steps for your dropout issues:
- Go to Settings → Audio, and for each HiFiBerry DAC+ HD, temporarily disable it or remove it from the "Office + Bedroom + Garage + Rendu" grouped zone. Play through just the Sonore ultraRendu alone. See if you run into the same dropouts.
- Check the power supplies on each Raspberry Pi running a DAC+ HD. Clock instability at this severity could potentially be a power issue on a Pi, underpowered or failing PSU.
- Update the OS and HiFiBerry drivers on each Pi. The logs show the SonicTransporter is running kernel
5.6.13-100.sgc.fc30.x86_64which is quite old (Fedora 30 kernel from ~2020). The Pi OS on each endpoint should also be current. HiFiBerry has released driver updates that improve clock stability. - Re-add the DAC+ HD boards one at a time after checking power. This will tell you if it's one specific board vs. all three.
For the blue spinner (library identification):
- Do not restart Roon Server again unless absolutely necessary. Every restart resets the retry queue to ~143k tracks and adds time. The queue is slowly working through the backlog between restarts.
- Bulk-set your live recordings to "Prefer File" for metadata. In Roon, go to Albums, filter or sort to find unidentified albums, select them all, and choose Edit → Metadata Preference → "Prefer File." This tells Roon to stop trying to cloud-identify them. With 4,000+ live recordings that will never match the Roon database, this is the permanent fix. It will shrink the retry queue dramatically and likely stop the spinner for good.
- Convert the King Crimson .m4a files to FLAC as you planned. Once re-scanned, those problematic files will no longer cause issues.
We’re sorry you’re not finding the community helpful for offering technical support. We’re always welcome for user feedback on our processes, and always looking for ways to improve - feel free to share your thoughts over in the Feedback category in our community, as it’s actively monitored by our team.
Thanks again, David! 