03-24-2013 03:29 PM
hi everybody! I have an issue with vrrpe implementation, actually I have 2 icx 6610 working as active-active, and issue goes like this:
if icx-a is the master for vlan 124 and icx-b is the master for vlan 123 when users on vlan 124 want to go to vlan 123 results that that switch has the backup uplink so the comunnictation is never getting back to vlan 124 , so as far as I know mct is not supported on icx so I was wondering that I need to pass that traffic to the other redundant master box for vlan 124, but I tried to create a trunk between the icx's trying to pass the straffic to the responsible swtich for that vlan, here my big doubt is .... Is there anything alike mct on I cx's ? the weird issue with this is that I can reach some hosts on certain vlan and others I cant, any thoughts about this? Thanks!
03-25-2013 07:11 AM
Hi Michael, thanks For your response, the customer wants something like cisco uses with their hsrp solutions, I mean must be a trunk or etherchannel between the switches (not mct) this fotr pass the traffic thst belongs to every master equipment Oin each vlan. I will try to do this with spanning tree and vrrpe mixed... Anyway if is there something that might help me I will reaaly appreciate. It! Thanks
03-26-2013 05:16 PM
The traffic will not go to the backup router in the first place as the client should have a DG that is the floating IP address of the master VRRP-E instance - if it fails over then the IP address moves to the backup and it becomes master. But you can and should a link between the routers for heartbeats etc and this can and I would recommend be LACP(trunk) link.
Suggest you have a look at the config guide for 7.4.00 page 1694 (PDF page no 1752) called VRRP-E example
03-27-2013 05:24 PM
Hi mschipp! thank your for your response! actually is what I loooking for, connect a trunk or lacp trunk between switches to pass the control or state of the vrrp-e when the uplink is backup in one of the switches, so in this way the traffic can flow to the master box between this trunk.
I reviewed all the vrrp-e information that you posted thank you! in fact I have already configured the equipments in production but at this time I had to shutdown one of them because the loss of connectivity. I also was reading about spanning tree on the vrrp-e topology and maybe this could help me creating a L2 balancing making root on one equipment for certain vlans and root the other equipment for the rest.
I will do a test next monday, I will post the result...
Thanks!!! If there is another idea is welcome!!!
05-22-2013 12:11 PM
Hi Michael. I'm actually trying to do what you suggest in building a stack with the two ICX units running layer3. Although when I fail the master, the standby reboots (takes several minutes) and then takes over as the master. Is it possible to have an immediate handover to the standby switch for layer 3 functionality....or do I just run in a more traditional mode of two separate boxes and run VRRP between the two switches/routers.
05-28-2013 01:15 AM
Sorry for the late reply - Yes you but hitless failover is not enabled by default.
Enabling hitless stacking failover
To enable hitless stacking failover, enter the following command at the global CONFIG level of the CLI:
The command takes effect immediately. Hitless switchover is allowed, and in the event of a failover, the standby controller will take over the active role without reloading the stack.
Syntax: hitless-failover enable
Use the no form of the command to disable hitless stacking once it has been enabled.