Subject: | |
From: | |
Reply To: | |
Date: | Mon, 2 Jun 2014 07:40:53 -0400 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
On Mon, Jun 2, 2014 at 2:08 AM, Johan Guldmyr <[log in to unmask]> wrote:
> Hi!
>
> On a SL6.5 machine we now have two connections to the same switch/vlan. One is 1GbE (eth0) and the other is 10GbE (eth5).
> Could I just do a "brctl addif br0 eth5" and then "brctl delif br0 eth0" and it switch to use the 10GbE link?
> After this we could change the network-scripts to make it work after reboot.
> Will this prevent existing connections from being cut?
You're going to hurt yourself trying to do handstands to avoid
downtime. I suggest you accept a very short service interruption as
one module is brought down, and the other brought up. It's easy enough
to configure them to the same IP address and set the new one to be up
and the old one down, then do "/sbin/service network restart" Or you
can bring up the second node on a second IP address for testing, then
swap the IP addresses.
Unless the upstream switches support pair bonding operations, it's
just not going to work well. See my old notes at
https://wikis.uit.tufts.edu/confluence/display/TUSKpub/Configure+Pair+Bonding,+VLANs,+and+Bridges+for+KVM+Hypervisor
if you really have to do this sort of thing.
> # brctl show
> bridge name bridge id STP enabled interfaces
> br0 8000.18a9054c4d34 no eth0
> vnet1
> br4 8000.00265556f190 no eth4
> vnet0
> virbr0 8000.5254009398c7 yes virbr0-nic
>
>
> // Johan Guldmyr
> Systems Specialist
> CSC - IT Center for Science
> http://www.csc.fi
|
|
|