How to configure AFF200 nodes for 10Gb VRRP, HSRP, ESRP switch configuration?

[cgeck0000]

Hi there.

Currently setting up AFF200 nodes with 10Gb connectivity to 2 Extreme Black Diamond 8810 switches running ESRP.

ESRP is Extreme Networks propriatery standard comparable to Cisco HSRP and the open standard VRRP.

I found the following stating VRRP and HSRP is supported and "no special configuration is needed on the storage."

Does NetApp support VRRP and HSRP?

https://kb.netapp.com/support/s/article/ka21A0000000jxLQAQ/does-netapp-support-vrrp-and-hsrp?language=en_US

The thing is I do not know what that means.

Can someone point me in the direction of configuring NetApp for this? Thanks.

[andris]

Hi back,

The idea behind all of these "router redundancy" protocols is you can have multiple physical routers cooperating to provide a reliable "virtual router" gateway IP address. The protocol itself runs between the routers, only.  All the hosts (like the AFF A200) need to know is what the IP address of the virtual router/gateway is - just a configuration item when setting up the network LIFs. 

[cgeck0000]

Painting the picture...

netapp1 is connected to switch1

netapp2 is connected to switch2

ESRP\HSRP\VRRP (all the same) is configured between switch1 and switch2.

The virtual gateway IP address is sitting on switch2.

How does netapp1 connected to switch1 get to that gateway?

And if the virtual gateway IP fails over to switch1 then how does netapp2 reach that now?

Hope this helps more with my question.

[andris · Accepted solution]

What is your goal?  L2 redundancy, L3 (router) redundancy or both?

To summarize from a L3 redundancy perspective... with these protocols, you have a virtual router IP address AND a virtual ethernet MAC address for this router. When one fails, the other takes over advertising the virtual IP and using gratuitous ARP requests to "teach" everyone the new location of the virtual MAC address.

As you currently describe it, the NetApp nodes will not have L2 redundancy.  That's OK - but if the single link to the switch fails, or the switch itself completely fails (as opposed to the router interface not working or downstream paths not working), the connectivity will fail.

To configure L2 redundancy you'd dual-home the NetApp nodes to each switch, but you'd need a multi-switch link aggregation feature on the switch side.  Extreme's MLAG feature, Cisco's VPC.  The usual requirement on the edge/host/server/ONTAP side would be that the 2 ports are configured as a LACP (802.3ad) port-channel.