Thread Info
  • State Verified Answer
  • Locked Locked
  • Replies 5 replies
  • Subscribers 5 subscribers
  • Views 1805 views
  • Users 0 members are here
Options
Suggested
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

SG310 in HA to redundant switches

TCF

I Have 2 SG310s in HA, and 2 Cisco SX350X switches connected with 2 DAC cables in a LAG. There is a 4 node Nutanix Block. Coming out of each node in the block  are 2 DAC cables going to each SX350X switch for redundancy. I am trying to connect these 2 switches to my SG310s with redundant links and use the SG310 as my Gateway to the rest of the Active Interfaces/Network. Spanning tree is disabled on the switches due to causing issues with the Nutanix Block, so I am using LACP and Protected Ports instead. I was thinking of doing an Aggregation Group on the UTM, so 4 connections in total with the HA, but wasn't sure if that would work coming from 2 different switches. The link below is a diagram of what I envision to make this easier to follow. If a switch goes down the idea is for the other switch to take over and the UTM Interface to use the interface that is still up.

What is the best way to accomplish this. LAG ? Multipath Rule?

Any help is appreciated.



This thread was automatically locked due to age.
Parents
  • +1

    Hi TCF,

    my recommendation:

     2 LAG Groups - at the cisco switches. Every LAG-Group may/should be distributed over the 2 Switches within the stack.

    One UTM to one Cisco-LAG-Group (so you have only one LAG configured at SG) -

    Both LAG-Groups has the same configuration at the cisco switches.

    Compare: community.sophos.com/.../sophos-utm-ha-failover-and-stacked-cisco-switches-cisco-loses-connection-to-sophos-after-a-failover


    Dirk

    Systema Gesellschaft für angewandte Datentechnik mbH  // Sophos Platinum Partner
    Sophos Solution Partner since 2003
    If a post solves your question, click the 'Verify Answer' link at this post.

  • 0 in reply to dirkkotte

    Dirkkotte, Thank you for the recommendation. If I lag across the stack will the 10GB ports still operate at 10GB for each of the 4 node ports? My concern with stacking is a performance loss, as each of the Nutanix nodes requires a full 10GB connection. I have read that ports in a stack suffer a bandwidth loss. Does lagging over a stack load balance or will LACP shut off one of the ports so all of the data is going through the switch with Stack ID = 1?

  • 0 in reply to TCF

    Never heard from performance loss with cisco stacking ... possible a bit overhead ...
    ... but as only one sophos is active at a time, you have only 2 active 10GBit port. (also with active/active)

    ... and you use only one LACP-link per connection, so you never got 20 GBit between 2 peers.
    With multiple connections, you got the 20GBit


    Dirk

    Systema Gesellschaft für angewandte Datentechnik mbH  // Sophos Platinum Partner
    Sophos Solution Partner since 2003
    If a post solves your question, click the 'Verify Answer' link at this post.

  • 0 in reply to dirkkotte

    sorry ... not realized a detail from your picture until now .... for distributed LAG you need stacking or some similar virtual function...


    Dirk

    Systema Gesellschaft für angewandte Datentechnik mbH  // Sophos Platinum Partner
    Sophos Solution Partner since 2003
    If a post solves your question, click the 'Verify Answer' link at this post.

Reply
  • 0 in reply to dirkkotte

    sorry ... not realized a detail from your picture until now .... for distributed LAG you need stacking or some similar virtual function...


    Dirk

    Systema Gesellschaft für angewandte Datentechnik mbH  // Sophos Platinum Partner
    Sophos Solution Partner since 2003
    If a post solves your question, click the 'Verify Answer' link at this post.

Children
No Data
Unfiltered HTML