ARPING shows TWO mac on 1 ip ? Is it normal ? - i guess not ?

[tytanick]

Hi,
I have strange problem when two nodes sometimes i see error below, and /etc/pve is blocked due to problem in comunication
corosync [TOTEM ] Retransmit List: 3b8 3b9 3ba 3bb 3bc 3bd 3be 3bf

Also after 6 hours i saw that arping is acting strange.
Tell me is it normal or i have something wrong in my network settings ?

In my network i have 2 mikrotiks bonded together, anso in node1 there is bond connected to boths switches, and second node also bond connected via rj45 to those two mikrotiks.
Meaby i have some kind of loop here ? and thats why i have Totem Retransmission problem ?

Tell me is it normal or not ?

root@node1:~# ifconfig | grep 44:a8:42:20:45
bond0 Link encap:Ethernet HWaddr 44:a8:42:20:45:a7
bond1 Link encap:Ethernet HWaddr 44:a8:42:20:45:a8
eth1 Link encap:Ethernet HWaddr 44:a8:42:20:45:a7
eth2 Link encap:Ethernet HWaddr 44:a8:42:20:45:a8
eth3 Link encap:Ethernet HWaddr 44:a8:42:20:45:a8
vmbr0 Link encap:Ethernet HWaddr 44:a8:42:20:45:a7



root@router:~# arping -I eth0 10.55.55.16
ARPING 10.55.55.16
60 bytes from 44:a8:42:20:45:a8 (10.55.55.16): index=0 time=516.740 usec
60 bytes from 44:a8:42:20:45:a7 (10.55.55.16): index=1 time=622.262 usec
60 bytes from 44:a8:42:20:45:a8 (10.55.55.16): index=2 time=678.168 usec
^C
--- 10.55.55.16 statistics ---
1 packets transmitted, 3 packets received, 0% unanswered (2 extra)


======================
root@node2:~# ifconfig | grep 44:a8:42:20:60
bond0 Link encap:Ethernet HWaddr 44:a8:42:20:60:95
bond1 Link encap:Ethernet HWaddr 44:a8:42:20:60:96
eth1 Link encap:Ethernet HWaddr 44:a8:42:20:60:95
eth2 Link encap:Ethernet HWaddr 44:a8:42:20:60:96
eth3 Link encap:Ethernet HWaddr 44:a8:42:20:60:96
vmbr0 Link encap:Ethernet HWaddr 44:a8:42:20:60:95



root@router:~# arping -I eth0 10.55.55.17
ARPING 10.55.55.17
60 bytes from 44:a8:42:20:60:95 (10.55.55.17): index=0 time=316.762 usec
60 bytes from 44:a8:42:20:60:96 (10.55.55.17): index=1 time=422.822 usec
60 bytes from 44:a8:42:20:60:96 (10.55.55.17): index=2 time=477.580 usec


Node2 ethernet config:
auto bond0
iface bond0 inet manual
<------>slaves eth1
<------>bond_miimon 100
<------>bond_mode balance-rr


auto bond1
iface bond1 inet static
<------>address 10.55.55.17
<------>netmask 255.255.255.0
<------>gateway 10.55.55.1
<------>slaves eth2 eth3
<------>bond_miimon 100
<------>bond_mode balance-rr

auto vmbr0
iface vmbr0 inet static
<------>address 192.168.0.12
<------>netmask 255.255.255.0
<------>bridge_ports bond0
<------>bridge_stp off
<------>bridge_fd 0

 

[tytanick]

i think that i fixed the problem


both bond0 and bond1 was connected to the same switch (the same vlan)
When arpinging , host answered from both bond0 and bond1 with different mac addresses so that it could be a problem in connection sometinmes.

I disabled this feature in linux and now i will test if it helped.

sysctl -w net.ipv4.conf.all.arp_filter=1