Please help stop the mDNS flood. I have enabled MulticastDNS in one of my network files (see below). When I start the systemd-resolved service it consumes >90% CPU. Other machines on the network are being flooded by mDNS packets and their CPU also gets
loaded. Capturing a trace with tcpdump shows the server is sending a flood of mDNS packets. If I disable MulticastDNS in the network file or stop systemd-resolved service, the mDNS flood stops. Here’s a small portion of tcpdump showing the flood from one server. $ tcpdump -i eno0 port 5353 and src srv 09:59:13.864876 IP srv.mdns > mdns.mcast.net.mdns: 0 [b2&3=0x400] [1au] TXT (QM)? PX3-1173-N1Q1V2K002 My PDU._snmp._udp.local. (114) 09:59:13.865023 IP srv.mdns > mdns.mcast.net.mdns: 0 [b2&3=0x400] [1au] SRV (QM)? PX3-1173-N1Q1V2K002 My PDU._raritan_rpcs._tcp.local. (125) 09:59:13.865208 IP srv.mdns > mdns.mcast.net.mdns: 0 [b2&3=0x400] [1au] TXT (QM)? PX3-1173-N1Q1V2K002 My PDU._raritan_rpcs._tcp.local. (132) 09:59:13.891815 IP srv.mdns > mdns.mcast.net.mdns: 0 [b2&3=0x400] [1au] TXT (QM)? PX3-1173-N1Q1V2K002 My PDU._raritan_rpcs._tcp.local. (132) 09:59:13.891973 IP srv.mdns > mdns.mcast.net.mdns: 0 [b2&3=0x400] [1au] SRV (QM)? PX3-1173-N1Q1V2K002 My PDU._raritan_rpcs._tcp.local. (125) 09:59:13.892108 IP srv.mdns > mdns.mcast.net.mdns: 0 [b2&3=0x400] [1au] TXT (QM)? PX3-1173-N1Q1V2K002 My PDU._https._tcp.local. (115) 09:59:13.892222 IP srv.mdns > mdns.mcast.net.mdns: 0 [b2&3=0x400] [1au] TXT (QM)? PX3-1173-N1Q1V2K002 My PDU._snmp._udp.local. (114) Why is systemd-resolved issuing these queries? How do I configure it to stop the mDNS flood? $ networkctl cat 10-eno0.network # /etc/systemd/network/10-eno0.network [Match] KernelCommandLine=!nfsroot Name=eno0 [DHCP] ClientIdentifier=mac RouteMetric=10 UseDomains=yes UseHostname=yes UseMTU=yes [IPv6AcceptRA] UseRedirect=no [Link] RequiredForOnline=no [Network] DHCP=yes LinkLocalAddressing=yes MulticastDNS=yes Thanks, Matt. |