Has anyone successfully created a bonded interface (bond0) with say two nics AND also then aliased that interface to get bond0 and bond0:1 (for an alias ip)
I just want to know if it's possible and reliable
Thanks
On Sun, Sep 5, 2010 at 10:33 AM, Robert Arkiletian robark@gmail.com wrote:
Has anyone successfully created a bonded interface (bond0) with say two nics AND also then aliased that interface to get bond0 and bond0:1 (for an alias ip)
I just want to know if it's possible and reliable
Thanks
Oh I forgot to mention I would be using link aggregation mode 4, 802.3ad bonding.
On Sep 5, 2010, at 1:35 PM, Robert Arkiletian robark@gmail.com wrote:
On Sun, Sep 5, 2010 at 10:33 AM, Robert Arkiletian robark@gmail.com wrote:
Has anyone successfully created a bonded interface (bond0) with say two nics AND also then aliased that interface to get bond0 and bond0:1 (for an alias ip)
I just want to know if it's possible and reliable
Thanks
Oh I forgot to mention I would be using link aggregation mode 4, 802.3ad bonding.
It works fine just create the ifcfg-bondX:X files for your aliases and they'll be added once the bond comes up.
-Ross
On 5 Sep 2010, at 18:35, Robert Arkiletian robark@gmail.com wrote:
Has anyone successfully created a bonded interface (bond0) with say two nics AND also then aliased that interface to get bond0 and bond0:1 (for an alias ip)
I just want to know if it's possible and reliable
Oh I forgot to mention I would be using link aggregation mode 4, 802.3ad bonding.
Possible and very reliable, at least with Intel quad NICs (can't remember model) and Procurve 2900 switches in my last job.
Ben
This reminded me of a recent issue doing just what you are asking about. I had an issue with sending traffic out an aliased interface on a bonded interface a few weeks ago. I needed certain traffic to look like it was from the aliased interfaces IP so the ACL on the receiving side would allow posts. But ALL traffic was still looking like it came from the real interface's IP address. and this caused the ACL to deny it. I solved it using iptables.
-rick
On Sun, Sep 5, 2010 at 12:08 PM, Benjamin Donnachie benjamin@py-soft.co.ukwrote:
On 5 Sep 2010, at 18:35, Robert Arkiletian robark@gmail.com wrote:
Has anyone successfully created a bonded interface (bond0) with say two nics AND also then aliased that interface to get bond0 and bond0:1 (for an alias ip)
I just want to know if it's possible and reliable
Oh I forgot to mention I would be using link aggregation mode 4, 802.3ad bonding.
Possible and very reliable, at least with Intel quad NICs (can't remember model) and Procurve 2900 switches in my last job.
Ben _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On 05/09/2010 18:33, Robert Arkiletian wrote:
Has anyone successfully created a bonded interface (bond0) with say two nics AND also then aliased that interface to get bond0 and bond0:1 (for an alias ip)
I just want to know if it's possible and reliable
Possible and seems to be quite stable. Heartbeat does exactly this for the shared IP and we haven't had any problems whatsoever since we set it up.
Thanks
On Mon, Sep 6, 2010 at 1:14 AM, Gabriel Tabares gabriel.tabares@roboreus.com wrote:
On 05/09/2010 18:33, Robert Arkiletian wrote:
Has anyone successfully created a bonded interface (bond0) with say two nics AND also then aliased that interface to get bond0 and bond0:1 (for an alias ip)
I just want to know if it's possible and reliable
Possible and seems to be quite stable. Heartbeat does exactly this for the shared IP and we haven't had any problems whatsoever since we set it up.
Thanks
Thanks for all the replies guys.
On Sun, 5 Sep 2010, Robert Arkiletian wrote:
Has anyone successfully created a bonded interface (bond0) with say two nics AND also then aliased that interface to get bond0 and bond0:1 (for an alias ip)
I just want to know if it's possible and reliable
Yes to both. I have been doing this for about four years now on various versions of CentOS. The one thing you need to watch out for is the same as when not using aliases - that eth0 (etc) always comes up as eth0, and doesn't get renumbered. Solve this with HWADDR lines in ifcfg-eth0 (etc). Otherwise it is solid.
Steve
-
Benjamin Donnachie -
Gabriel Tabares -
Rick Philbrick -
Robert Arkiletian -
Ross Walker -
Steve Thompson