networking

Fadia Almarei — Sun, 25 Jan 2004 07:26:49 GMT

hi all,

i have four network cards on one server, i know that tru64 supports load balancing on a group of network cards (NetLag), and also supports fail-over for balanced network cards (netrain)....

how could i make:
-load balancing with the same IP ?
-fail-over with the same IP also ?

thanX in advance

Re: networking

Balaji N — Sun, 25 Jan 2004 12:27:06 GMT

this is one link which i got from google.
http://www.utterberg.com/howtos/teaming_howto.php

google for network teaming or network bonding.
also i got this from google.

http://sourceforge.net/projects/bonding/

hth
-balaji

Re: networking

Balaji N — Sun, 25 Jan 2004 12:28:31 GMT

more to add. here is a mail from one of the mailing list i subscribe to called lost. (check it out for Linux One Stanza Tips. http://lost.sourceforge.net)

> Try it out if you have 2 NIC. It's easy to do. rh7.x works, later kernel seems to handle fail over
better. I tried on 2.4.9 default 7.2 kernel, that does not seems to handle well when I unplug 1 NIC,
the bw drops. 2.4.20 handles very well when is disconnect on of the NIC. need to do more testing.
>
> [kent@kent kent]$ cat /etc/sysconfig/network-scripts/ifcfg-bond0
> DEVICE=bond0
> IPADDR=192.168.2.99
> NETMASK=255.255.255.0
> NETWORK=192.168.2.0
> BROADCAST=192.168.2.255
> GATEWAY=192.168.2.1
> ONBOOT=yes
> BOOTPROTO=none
> USERCTL=no
>
> [kent@kent kent]$ cat /etc/sysconfig/network-scripts/ifcfg-eth0
> DEVICE=eth0
> ONBOOT=yes
> BOOTPROTO=none
> MASTER=bond0
> SLAVE=yes
>
> [kent@kent kent]$ cat /etc/sysconfig/network-scripts/ifcfg-eth1
> DEVICE=eth1
> ONBOOT=yes
> BOOTPROTO=none
> MASTER=bond0
> SLAVE=yes
>
> [root@kent network-scripts]# cat ifcfg-bond0:0
> DEVICE=bond0:0
> IPADDR=192.168.2.196
> NETMASK=255.255.255.0
> NETWORK=192.168.2.0
>
> Restart Network...
>
> [root@kent network-scripts]# /sbin/ifconfig -a
> bond0 Link encap:Ethernet HWaddr 00:90:27:2C:7F:EA
> inet addr:192.168.2.99 Bcast:192.168.2.255 Mask:255.255.255.0
> UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
> RX packets:12278869 errors:7 dropped:0 overruns:0 frame:7
> TX packets:8124496 errors:18 dropped:0 overruns:0 carrier:18
> collisions:6 txqueuelen:0
> RX bytes:464315659 (442.8 Mb) TX bytes:3228444502 (3078.8 Mb)
>
> bond0:0 Link encap:Ethernet HWaddr 00:90:27:2C:7F:EA
> inet addr:192.168.2.196 Bcast:192.168.2.255 Mask:255.255.255.0
> UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
> RX packets:12278869 errors:7 dropped:0 overruns:0 frame:7
> TX packets:8124496 errors:18 dropped:0 overruns:0 carrier:18
> collisions:6 txqueuelen:0
> RX bytes:464315659 (442.8 Mb) TX bytes:3228444502 (3078.8 Mb)
>
> eth0 Link encap:Ethernet HWaddr 00:90:27:2C:7F:EA
> inet addr:192.168.2.99 Bcast:192.168.2.255 Mask:255.255.255.0
> UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
> RX packets:11731921 errors:4 dropped:0 overruns:0 frame:4
> TX packets:7182275 errors:4 dropped:0 overruns:0 carrier:4
> collisions:0 txqueuelen:100
> RX bytes:326305683 (311.1 Mb) TX bytes:1996540149 (1904.0 Mb)
> Interrupt:11 Base address:0x6c00 Memory:e6600000-e6600038
>
> eth1 Link encap:Ethernet HWaddr 00:90:27:2C:7F:EA
> inet addr:192.168.2.99 Bcast:192.168.2.255 Mask:255.255.255.0
> UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
> RX packets:546948 errors:3 dropped:0 overruns:0 frame:3
> TX packets:942221 errors:14 dropped:0 overruns:0 carrier:14
> collisions:6 txqueuelen:100
> RX bytes:138009976 (131.6 Mb) TX bytes:1231904353 (1174.8 Mb)
> Interrupt:10 Base address:0x7000 Memory:e6601000-e6601038

and some more info.

more:

> oops forgot you need the bonding module, exec this as root to enable the mod on the fly.
>
> insmod /lib/modules/2.4.20-19.8.SGI_XFS_1.2.0.kent/kernel/drivers/net/bonding.o
>
> and add this in /etc/modules.conf so it picks up during bootup.
>
> alias bond0 bonding

Re: networking

Steven E. Protter — Sun, 25 Jan 2004 14:45:14 GMT

As always my forums colleageue's advice looks solid.

I would not try this on a production server without having tried it and run for 30-60 days on a test box.

Seems workable though.

The nice part is unlike HP-UX, Linux will tolerate two physical devioces on the same subnet.

SEP

Re: networking

Fadia Almarei — Mon, 26 Jan 2004 01:59:33 GMT

hi

is this methods can be done for any linux systems ,that i have a linux 9 system

fadia

Re: networking

Alexander Chuzhoy — Mon, 26 Jan 2004 03:00:50 GMT

I don't know if it can be done for any Linux systems , but for yours I'm pretty sure it can be done..

You have also full manual here, although the guys already have provided you with more than enough:

http://www.redhat.com/docs/manuals/enterprise/RHEL-3-Manual/ref-guide/s1-networkscripts-interfaces.html#S2-NETWORKSCRIPTS-INTERFACES-CHAN

topic Re: networking in Operating System - Linux

networking

Re: networking

Re: networking

Re: networking

Re: networking

Re: networking