Roon Core Machine
AMD Ryzen 3950X, 64G DDR4, ASUS ROG VIII HERO, Samsung NVMe SSDs, NVIDIA 1070+2070s
Latest Roon 1148.
Most of library is on a custom 30T ZFS NAS with NVMe caching, on the local network.
Networking Gear & Setup Details
Router: Mikrotik RB5009
Switches: Netgear/TP-Link smart managed switches, various
All devices hardwired on Ethernet.
All Ethernet cables and routes certified for 1000BASE-T with a Fluke Versiv DSX8000. Patch cords certified to 5e or 6A as appropriate.
All switches monitored. All endpoint devices monitored, with smokeping, for latency and frame/packet loss, continuously, along with a variety of internet endpoints, including TIDAL API and media servers used by Roon.
Internet is a 1Gbps FTTH Ethernet circuit. No packet loss, and speed tests via various means all exceed 800-900Mbps even in peak time.
Connected Audio Devices
Core has ASIO RME Fireface UCX attached directly by USB, driving one stereo output.
Two other grouped outputs, both Roopiee RPi devices, both Raspberry Pi 3 Model B Rev 1.2. Roon bridge 1.8, build 1125.
Bathroom: Using a HiFiBerry AMP hat.
Living Room: Using a Focusrite 2i2 Scarlett USB.
Have tried with default settings or (as logged) forcing all sample rates and bitrates to interface-native. Turned off any and all MQA/DSD related stuff. Keeping it simple.
Number of Tracks in Library
Circa 20,000 tracks.
Description of Issue
So, been having this problem for literal years now and never had Roon support responses. Last try before I give up and replace everything with Sonos instead of buying a lifetime license.
Logs directory here: TalkUnafraid Nextcloud
At random when listening to music in a queue, Roon stops somewhere between 3 and 4 seconds after starting a new track. No error is given, no toasts pop up, no explanation - just stops. Pressing play as soon as it happens resumes the playback without issue and it will continue until the next time it decides to stop. This seems to mostly happen with streaming services; I default to and prefer TIDAL but have Qobuz set up also.
Network would be the obvious contender except that this setup is about as good as it gets - I’m a network engineer who specialises in test and measurement and have verified everything. I also work for my ISP, and have verified all is well on that end all the way through to the carrier/transit peering which reaches TIDAL.
All the local network devices have consistently <1ms latency, there is a local stratum 2 NTP server provided on DHCP and used by all network devices, and everything appears fine in terms of hardware. I reimaged the Roopiee boxes recently to no avail.
Logs have this which at least appears to correlate time-wise:
11/25 08:32:50 Trace: [snd_rpi_hifiberry_amp] [raatclient] GOT [59] {"status":"Dropout","samples":15145}
11/25 08:32:50 Trace: [Scarlett 2i2 USB] [raatclient] GOT [16] {"status":"Dropout","samples":2438}
...
11/25 08:32:51 Trace: [ASIO Fireface USB] [raatclient] GOT [17] {"samples":11209,"status":"Dropout"}
...
11/25 08:32:53 Warn: [James Genelecs + Bathroom + Downstairs HiFi] [zoneplayer/raat] Too many dropouts (>3s dropped out in the last 30s). Killing stream
But there’s no reason why this should be occurring, and no further detail.
How do I troubleshoot further?