Wired

last person joined: 7 hours ago 

Bring performance and reliability to your network with the Aruba Core, Aggregation, and Access layer switches. Discuss the latest features and functionality of the ArubaOS-Switch and ArubaOS-CX devices, and find ways to improve security across your network to bring together a mobile first solution.
Expand all | Collapse all

VSF

  • 1.  VSF

    Posted Jan 11, 2019 09:12 AM

     

    Hi

    I would like to confirm this scenario and see if there is any alternative to this

    Current Scenario:

    2x 5400 chassis running VSF
    MAD LACP enabled

     

    Member 1 priority 255 - Master
    Member 2 priority 100 - Stanby


    Member 1 has fibre links coming in from the access switches as well as various copper links for servers, Firewall etc.

    Member 2 has only the redundant fibre links coming in from the Access switches

     

    Future Scenario:

    Member 1 goes down Member 2 takes over as Master

    Member 1 comes up stays as Standby - no preempt option available

    network continue to run

    Then..

    Fibre break between the two 5400 Cores MAD LACP now
    shutsdown Member 1 as it is the standby member

    This will now be an issue as the servers and firewalls connected to Member 1
    will be down and the company will not be able to operate

    So I would like to know

    Am I correct with the above?

    I have read elsewhere that MAD will not shutdown all interfaces
    but only the interfaces for any device with redundant links to
    both members in which case the servers and firewalls will stay up

    However if all interfaces in member 1 will be shut is there any workaround
    to this other than reloading member 2so it comes back as standby?

     

    thank you



  • 2.  RE: VSF

    Posted Jan 11, 2019 10:49 AM

    @tars_B wrote: I have read elsewhere that MAD will not shutdown all interfaces but only the interfaces for any device with redundant links to both members in which case the servers and firewalls will stay up

    AFAIK if a VSF fabric operating as a chain is going to experience (a) a failure of all of its VSF Links (so the whole VSF Link goes down) or (b) a failure of just one of its VSF Fabric Members, it will end up in a Split Stack condition...in case of (a) VSF Fabric members remain operational (VSF Link is down) and you end up [*] with a Standby member with all ports down (except those configured as member of VSF Link).

     

    [*] that's because the Standby member is supposed to be part of the VSF Fabric fragment that will become inactive, the Commander will be part of the the active fragment.

     

    IMHO if you are going to use (or are just using already) VSF with a MAD mechanism you should then dual homing eveything against VSF (so Switch and Hosts) and you should do it by extensively using LACP link aggregations.