Hi @ryan_stratton,
Thanks for the logs! After analyzing them in detail, we have a clear picture of what’s happening.
The good news: this is not a Roon bug. The logs point strongly to a network-level issue in your home setup.
What the logs show
1. All endpoints drop simultaneously - this is the smoking gun
When slowness occurs, the logs show all 7 audio devices losing RAAT connectivity at the exact same moment:
06/08 15:53:49 Warn: [AURALiC ARIES_G1 @ 192.168.68.54] RAAT Session timed out during initialization
06/08 15:53:49 Warn: [Rose RS130 @ 192.168.68.65] RAAT Session timed out during initialization
06/08 15:53:49 Warn: [Naim Audio Uniti Nova @ 192.168.68.69] RAAT Session timed out during initialization
06/08 15:53:49 Warn: [Naim Audio Mu-so @ 192.168.68.55] RAAT Session timed out during initialization
06/08 15:53:52 Warn: [Bowers & Wilkins Formation Flex @ 192.168.68.56] RAAT Session timed out during initialization
06/08 15:53:52 Warn: [ARCAM SA45 @ 192.168.68.66] RAAT Session timed out during initialization
All six devices fail within a 3-second window. This pattern rules out any individual device as the cause - when every device on the network drops at the same time, the network itself is the problem.
2. Devices are changing IP addresses - the root cause
This is the most significant finding. Across the logs every single audio endpoint appears at two completely different IP addresses, and the changes happen simultaneously across all devices:
06/06 15:24:15 - all devices get new IPs at the exact same second (coincides with the first mass timeout burst in the logs):
| Device |
IP before |
IP after |
| AURALiC ARIES_G1 |
- |
192.168.68.62 |
| ARCAM SA45 |
- |
192.168.68.58 |
| Rose RS130 |
- |
192.168.68.52 |
| Bluesound POWERNODE 2 |
- |
192.168.68.65 |
| B&W Formation Flex |
- |
192.168.68.54 |
| Naim Audio Mu-so |
- |
192.168.68.66 |
06/08 - another full rotation, all devices on different IPs again:
| Device |
Old IP |
New IP |
| AURALiC ARIES_G1 |
.62 |
192.168.68.54 |
| ARCAM SA45 |
.58 |
192.168.68.66 |
| Rose RS130 |
.52 |
192.168.68.65 |
| Bluesound POWERNODE 2 |
.65 |
192.168.68.51 |
| B&W Formation Flex |
.54 |
192.168.68.56 |
| Naim Audio Uniti Nova |
.68 |
192.168.68.69 |
| Naim Audio Mu-so |
.66 |
192.168.68.55 |
Notice that IPs are being recycled between devices - Bluesound previously held .65, then Rose RS130 got it. AURALiC moved from .62 to .54, which B&W had held before. The TP-Link Deco DHCP pool is reassigning addresses without MAC-based reservations.
Here is what happens from Roon’s perspective when a device changes IP:
- Roon Server has the old IP cached (e.g.
.62 for AURALiC)
- Device roams to a different mesh node, disconnects briefly, gets a new IP (
.54)
- Roon tries to reconnect to
.62 - 10-second timeout, address no longer exists
- Roon waits for mDNS re-discovery to learn the new IP
- mDNS re-discovery runs on its own schedule, not immediately
- Only after re-discovery does Roon re-establish the RAAT session
Multiply this by 7 devices all roaming simultaneously, and you have your 5-minute delay.
3. The ARCAM SA45 is the worst offender
Timeout count across all logs:
| Device |
Total timeouts |
Notes |
| ARCAM SA45 |
49 |
35 init + 14 drop |
| Naim Audio Uniti Nova |
26 |
|
| Naim Audio Mu-so |
19 |
|
| B&W Formation Flex |
13 |
|
| Bluesound POWERNODE 2 |
12 |
|
| Rose RS130 |
10 |
|
| AURALiC ARIES_G1 |
6 |
most stable |
The AURALiC (only 6 events, consistently starts playback in ~4 seconds) vs. the ARCAM (49 events) on the same network and same Roon Server strongly suggests the ARCAM is physically located in a weaker coverage area and roams more frequently between mesh nodes.
4. Play → audio delay comes from retry cascades after IP changes
When you press Play during a bad network moment, this is what Roon goes through:
[raat] [AURALiC ARIES_G1] client connection failed. Retrying in 500ms
[raat] [AURALiC ARIES_G1] client connection failed. Retrying in 750ms
[raat] [AURALiC ARIES_G1] client connection failed. Retrying in 1125ms
...
Each retry cycle is 10 seconds of waiting before the failure is declared, then a backoff pause before the next attempt. With 7 devices all failing simultaneously after an IP change, total delays of several minutes are entirely expected.
On a good day when IPs are stable, the same logs show clean startups:
06/06 16:21:18 [zone RS - 130] Playing 11 Items ← Play pressed
06/06 16:21:26 [RS - 130] [zoneplayer] Starting playback ← audio starts
8 seconds total, no errors. The system works perfectly when the network is stable.
5. Why Spotify and AirPlay are unaffected
Spotify Connect and AirPlay stream directly from the cloud to the device — they don’t require a local-network RAAT session between the Nucleus and each endpoint. An IP change on the local network doesn’t disrupt them. This confirms the issue is specifically with local LAN stability, not your internet connection or the devices themselves.
Root cause summary
Your TP-Link Deco mesh network is reassigning DHCP leases when devices roam between mesh nodes. Devices that sit on the border of two nodes’ coverage areas are the most affected - they roam back and forth, drop their DHCP lease, and come back on a new IP. The ARCAM SA45 appears to be in the worst physical location by this metric. When multiple devices roam simultaneously (which happens when the mesh backbone link renegotiates), Roon loses all endpoints at once and must rediscover all of them before playback can resume.
Fix: one specific action required
Set DHCP reservations (static IPs by MAC address) for all 7 Roon endpoints in the Deco app. This ensures each device always gets the same IP regardless of which mesh node it connects to, so Roon never has to re-discover them after a roam event.
In the Deco app: More → Advanced → DHCP Server → Add Reserved IP - add an entry for each device using its MAC address.
You’ll find the MAC addresses in the Deco app under the client list, or in each device’s own settings menu.
Additionally:
- Update all Deco nodes to the latest firmware - several versions had roaming stability bugs
- Check physical placement of the ARCAM SA45 - with 49 timeout events it’s likely sitting in a coverage gap; moving it or adding a Deco node nearby would help
- Wire the Nucleus+ if at all possible - its IP
192.168.68.58 suggests it may be on Wi-Fi; a wired core removes one variable from the critical path
After setting the static IP reservations, reboot all Deco nodes and your Roon endpoints. The slowness should resolve immediately.