Unsolved
This post is more than 5 years old
1 Message
0
6436
June 26th, 2011 22:00
ESXi specific redundant network scenario - 2 x 4 nic R710, 2 x 5224
Hoping for some general pointers on best practice on this one. I have 2 x 4 gigabit NIC R710 server, 2 x PC5224 switches and a VMware vSphere Essentials Plus, which includes a HA license that I will be using for VM failover. Other point of note is that the servers are broken up across two VLANs (ids 10 and 20). Obviously this is for a small business (about 8 servers, and a very small number of users).
Would like to set it up as a fairly resilient system, but not have HA shutting down and transitioning machines if one of the switches dies, or if I happen to glare at the networking the wrong way!
In respect of connecting the swithes to each other, because the 5224's are not stackable, my thoughts were that I would set up a 2 port LAG with LACP using ports 23+24 on each switch to enable a 2Gbps channel between them. I would VLAN tag the LAG as 10, 20 General, and set the underlying ports as general but otherwise as default. Is this a valid approach or is the some glaring bits I am missing here on the first bit.
On the ESXi side of things though, there seems to be a number of possibilities, and not sure which way to go. Because this is such a small scenario, should I just
* amalgamate all 4 ports for management, vmotion and vm networks at the vsphere level
* Make all 4 nics active and turn on ip hash for the single vswitch (and on repeat on vmanagment and vmnetworks as well???)
* create an identical lag group on each of the5224 switches containing the ports that ESX servers plug into (I think configured as L3?)
vlan tag the lag as 10, 20 general
Am I kind of close?
Would appreciate someone setting me on the straight and narrow on how to best achieve this level of redundancy and failover in this scenario!
Cheers
Andrew McKenzie
Would like to set it up as a fairly resilient system, but not have HA shutting down and transitioning machines if one of the switches dies, or if I happen to glare at the networking the wrong way!
In respect of connecting the swithes to each other, because the 5224's are not stackable, my thoughts were that I would set up a 2 port LAG with LACP using ports 23+24 on each switch to enable a 2Gbps channel between them. I would VLAN tag the LAG as 10, 20 General, and set the underlying ports as general but otherwise as default. Is this a valid approach or is the some glaring bits I am missing here on the first bit.
On the ESXi side of things though, there seems to be a number of possibilities, and not sure which way to go. Because this is such a small scenario, should I just
* amalgamate all 4 ports for management, vmotion and vm networks at the vsphere level
* Make all 4 nics active and turn on ip hash for the single vswitch (and on repeat on vmanagment and vmnetworks as well???)
* create an identical lag group on each of the5224 switches containing the ports that ESX servers plug into (I think configured as L3?)
vlan tag the lag as 10, 20 general
Am I kind of close?
Would appreciate someone setting me on the straight and narrow on how to best achieve this level of redundancy and failover in this scenario!
Cheers
Andrew McKenzie
No Events found!