New setup - Network redundancy

[camelstrike]

Hello all,

I will be setting up a 3 node cluster at work and I have some questions regarding network redundancy.

Hardware:

• 2 x Netgear S3300 switches
• 3 x Nodes, each has:
  • 2 x 1 Gb NICs - onboard
  • 4 x 1 Gb NICs - card
  • Will be running ZFS:
    • vmpool: 4 x 4 TB SAS HDD - Raid10
    • ZIL/SLOG: 2 x 240 GB SSD - Raid1
• Network:
  • cluster:
    • ring0: 10.10.0.0/24
    • ring1: 10.10.1.0/24

So Im left with 4 ports (eno3-6) on each node to configure the following networks:

• Migration
• Management
• VMs-CT

And I dont know how to go about that, I was thinking of:

• Migration: LACP bond0 slaves eno3 eno4 --> vmbr0 10.1.2.0/24
• Management/VMs-CT: LACP bond1 slaves eno5 eno6 --> vmbr1 192.168.1.0/24

With this setup will I be able to power off one switch and things will continue to work?

Do you have any other recommendations?

Thanks for taking the time to read, best regards.

 

[wolfgang]

Hi,

With this setup will I be able to power off one switch and things will continue to work?

This dependence on your switch and if stacking works properly.

Do you have any other recommendations?

Ative-Backup will always work but use just one line at a time.