I have a single Dell 2650 with Multiple onboard NICs. I have successfully bonded eth0 and eth1 for failover (mode=1) and have verified that it works great. I do show that eth1 has the NOARP flag set.
For testing purposes, the machine is hooked up to an old hub with half duplex 10 base T. Both NICs are plugged into the hub. (When completely configured it will move to a nice colo environment with a managed HP Procurve switch.) I add this info because it might be the cause of the problem.
When I ping hosts, I always get a DUP! message. In some ways, it makes sense that the data is received on each interface. However, I would have thought that one of the interfaces would "ignore" the received data until it became active. As it stands, it appears the data traverses up the network stack and always gets processed twice. At some point, IP/ICMP/TCP/UDP detect the duplicate.
Is this the norm? Do all bonding drivers work this way (i.e. require twice the processing power)? Will this go away if I plug both interfaces into a single switch?
Thanks in advance,
- Steve
P.S. Couldn't find any mention of this in deja.com so I figure I'm doing something pretty dain bramaged.
I have encountered the same problem.
It is all because of your old hub half duplex 10 base T hub ,but it is not a problem,the communication is fine.
If you change the hub to switch,you will not get DUP anymore.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi all,
I have a single Dell 2650 with Multiple onboard NICs. I have successfully bonded eth0 and eth1 for failover (mode=1) and have verified that it works great. I do show that eth1 has the NOARP flag set.
For testing purposes, the machine is hooked up to an old hub with half duplex 10 base T. Both NICs are plugged into the hub. (When completely configured it will move to a nice colo environment with a managed HP Procurve switch.) I add this info because it might be the cause of the problem.
When I ping hosts, I always get a DUP! message. In some ways, it makes sense that the data is received on each interface. However, I would have thought that one of the interfaces would "ignore" the received data until it became active. As it stands, it appears the data traverses up the network stack and always gets processed twice. At some point, IP/ICMP/TCP/UDP detect the duplicate.
Is this the norm? Do all bonding drivers work this way (i.e. require twice the processing power)? Will this go away if I plug both interfaces into a single switch?
Thanks in advance,
- Steve
P.S. Couldn't find any mention of this in deja.com so I figure I'm doing something pretty dain bramaged.
P.P.S. To give full info:
bond0 Link encap:Ethernet HWaddr 00:06:5B:F6:C1:89
inet addr:192.168.3.56 Bcast:192.168.3.255 Mask:255.255.255.0
UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
RX packets:845904 errors:0 dropped:0 overruns:0 frame:0
TX packets:388040 errors:0 dropped:0 overruns:0 carrier:0
collisions:2250 txqueuelen:0
RX bytes:157181258 (149.8 Mb) TX bytes:106440192 (101.5 Mb)
eth0 Link encap:Ethernet HWaddr 00:06:5B:F6:C1:89
inet addr:192.168.3.56 Bcast:192.168.3.255 Mask:255.255.255.0
UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
RX packets:234632 errors:0 dropped:0 overruns:0 frame:0
TX packets:388040 errors:0 dropped:0 overruns:0 carrier:0
collisions:2250 txqueuelen:100
RX bytes:26093785 (24.8 Mb) TX bytes:106440192 (101.5 Mb)
Interrupt:28
eth1 Link encap:Ethernet HWaddr 00:06:5B:F6:C1:89
inet addr:192.168.3.56 Bcast:192.168.3.255 Mask:255.255.255.0
UP BROADCAST RUNNING NOARP SLAVE MULTICAST MTU:1500 Metric:1
RX packets:611272 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
RX bytes:131087473 (125.0 Mb) TX bytes:0 (0.0 b)
Interrupt:29
I have encountered the same problem.
It is all because of your old hub half duplex 10 base T hub ,but it is not a problem,the communication is fine.
If you change the hub to switch,you will not get DUP anymore.