· Connecting my NUC running ROCK to my network via ethernet wreaks havok on the network, nothing that using the internet works. I see the connection lights blink normally on the switch for each ethernet port when everything is running smoothly. I plug in the NUC and the switch lights blink way faster and the internet in the entire home stops working. I'm running the latest version on ROCK. I thought it was the NUC so I bought a Nucleus and the same thing happend so I sent it back. This setup worked for years without problems.
Tell us about your home network
· Araknis router (AN-310-RT-4L2W) and switch (AN-110-SW-R-24)
I also presume this is a duplicate IP-address.
Have you used a static IP-address for you NUC or for something else?
Is you DHCP-server running correctly?
Thanks for writing in and for sharing your report! I believe what others here have asked are the correct next areas to troubleshoot:
First, there seems to be multiple active subnets within your local network - temporarily disable all but a single subnet, reboot your devices to ensure they’re all on the same subnet, and see if you experience the same issues.
From there, setting a static IP for your server would be an excellent troubleshooting step.
ROCK NUC triggers a network-wide broadcast storm the instant it connects — everything else ruled out. Layer 2 loop, or something else?
Follow-up to an earlier thread of mine (linked below) that auto-closed. I’ve since done a lot of elimination and want a sanity check before my integrator and I spend money on managed switches.
Setup: Intel NUC running ROCK (RoonOS), single M.2 for OS/database, separate 4TB SSD for ~1,100 albums. Network is Araknis router + unmanaged Araknis switches (Core 24-port in main rack, plus a Home Theater switch), with UniFi APs managed by a CloudKey. Single subnet (192.168.2.0/24), one VLAN, one DHCP pool.
Symptom: The moment the NUC joins the wired network, the switch LEDs start flailing and the entire network goes down — Control4, Wi-Fi, internet, everything. Unplugging the NUC’s Ethernet stops it instantly, every time. Classic broadcast-storm behavior.
What I’ve ruled out:
Roon hardware — happens with two different cores (old NUC and a Nucleus I since returned)
Ethernet cable — tested multiple, including a brand-new known-good cable today. Still storms.
Switch port — storms on the Core switch and on other ports/switches too
Multiple subnets — confirmed single subnet / single VLAN / single DHCP scope
Duplicate/static IP conflict — with the NUC off, pinging its reserved IP (192.168.2.144) returns “host is down,” so nothing else holds that address
Roon software/database — fully recovered and healthy; library and playlists all intact
The one constant: the NUC’s traffic alone triggers the storm regardless of cable, port, or switch. My working theory is a latent Layer 2 loop somewhere in the wired infrastructure that only ignites when a device pushes heavy multicast/broadcast (RAAT/mDNS/SSDP discovery on Roon startup). The Araknis switches are unmanaged (no STP, no storm control), so nothing contains it.
My questions for the community:
Does this symptom set definitively indicate a broadcast loop, or could a single failing switch / duplex fault / multicast-flooding issue present the same way?
Before buying managed switches (UniFi w/ STP + storm control) to replace the unmanaged Araknis ones, is there a way to confirm it’s a loop first — e.g., a specific test, or temporarily inserting one managed switch to read port stats / RSTP events?
Anyone solved an identical “ROCK storms the whole LAN on connect” case — and what was the actual root cause?
Want to make sure we fix the real thing rather than throw hardware at a guess. Thanks in advance.
Thanks for the thorough write-up - you’ve done a lot of good elimination work here.
Before investing in managed switches, there are a couple of simpler ways to confirm whether this is actually a Layer 2 loop:
Physical isolation test - with everything else still running normally, unplug every Ethernet cable in the house except the one going directly from the NUC to a single switch port, and power the NUC on. If the storm still happens with just that one cable connected, the loop is not in your wiring - something about the NUC’s traffic itself is triggering it. If it doesn’t storm, start reconnecting cables one at a time until you find the segment that triggers it.
Packet capture - if you have a laptop you can plug into the same switch and run Wireshark for 30-60 seconds right as the NUC connects. A broadcast storm will be immediately obvious - you’ll see thousands of frames per second, and the source MAC will tell you which device or port is the origin.
Regarding managed switches - worth noting that adding STP/storm control can actually interfere with Roon’s multicast and IGMP traffic if not configured carefully, so it’s not always a straightforward fix.
What does your wiring topology look like between the Core 24-port and the Home Theater switch - are they connected by a single uplink cable, or multiple?
Just checking in on this. Were you able to try the isolation test we mentioned, with every Ethernet cable unplugged except the direct run from the NUC to one switch port, and then reconnecting cables one at a time to see when the storm returns? If you had a chance, a short Wireshark capture on a laptop during the NUC boot would also help us confirm whether we are seeing a broadcast storm and which MAC address is at the center of it, and we are still wondering whether the Core 24-port and Home Theater switch are linked by a single uplink or multiple cables. Let us know what you found, thanks!