[root@brik1 ~]# arp -i eth0 -n | grep c4 192.168.1.15 ether 00:11:32:1d:c4:31 C eth0 192.168.1.16 ether 00:11:32:1d:c4:32 C eth0 192.168.1.17 ether 00:11:32:1d:c4:32 C eth0 192.168.1.18 ether 00:11:32:1d:c4:32 C eth0 [root@brik2 ~]# arp -i eth0 -n | grep c4 192.168.1.15 ether 00:11:32:1d:c4:32 C eth0 192.168.1.16 ether 00:11:32:1d:c4:32 C eth0 192.168.1.17 ether 00:11:32:1d:c4:32 C eth0 192.168.1.18 ether 00:11:32:1d:c4:32 C eth0On the Synology itself (ifconfig -a):
eth0 Link encap:Ethernet HWaddr 00:11:32:1D:C4:31 inet addr:192.168.1.15 Bcast:192.168.1.255 Mask:255.255.255.0 eth1 Link encap:Ethernet HWaddr 00:11:32:1D:C4:32 inet addr:192.168.1.16 Bcast:192.168.1.255 Mask:255.255.255.0 eth2 Link encap:Ethernet HWaddr 00:11:32:1D:C4:33 inet addr:192.168.1.17 Bcast:192.168.1.255 Mask:255.255.255.0 eth3 Link encap:Ethernet HWaddr 00:11:32:1D:C4:34 inet addr:192.168.1.18 Bcast:192.168.1.255 Mask:255.255.255.0On the Cisco SG300:
101 00:11:32:1d:c4:31 gi7 dynamic 101 00:11:32:1d:c4:32 gi9 dynamic 101 00:11:32:1d:c4:33 gi8 dynamic 101 00:11:32:1d:c4:34 gi10 dynamicWith manual ARP entries I was able to get 247 MB/s dd write performance. Without the manual ARP entries this was only 142 MB/s. This clearly shows that there is a performance impact due to this ARP problem. So I asked Synology tech support to comment on this setup. They came up with a solution, there is a special option in the Control Panel / Network section, for this setup:
[x] Reply to ARP requests if the target IP address is a local addres configured on the incoming interfaceOn the Synology, only one option of the linux kernel is changed (to 1) when you select this:
echo -n 1 > /proc/sys/net/ipv4/conf/all/arp_ignore arp_ignore - INTEGER Define different modes for sending replies in response to received ARP requests that resolve local target IP addresses: 0 - (default): reply for any local target IP address, configured on any interface 1 - reply only if the target IP address is local address configured on the incoming interfaceWith another change in kernel parameters, results are better:
for i in /proc/sys/net/ipv4/conf/*/rp_filter ; do echo -n 0 > $i ; done rp_filter - INTEGER 0 - No source validation. 1 - Strict mode as defined in RFC3704 Strict Reverse Path Each incoming packet is tested against the FIB and if the interface is not the best reverse path the packet check will fail. By default failed packets are discarded.