Forum Discussion

Gambler_168259's avatar
Gambler_168259
Icon for Nimbostratus rankNimbostratus
Nov 14, 2014

Does for redundant LTM devices need to be configured with HA with gateway and vlan fail-safe mode to avoid service disruption?

We have deployed two LTM 5000 Load Balancers to load balance SIEBEL web traffic. Both are deployed in redundant mode and connected with upstream Cisco Nexus 5000 Switches in full mesh mode. When we test redundancy by power-off the Nexus-1 switch, the downstream active LTM-01 & LTM-02 backup causing service disruption on servers, causing an application to be non-responsive for some minutes and ping response is timed out. We had already configured config-syn and fail-over as well. To avoid this issue Should we need to configure HA group or System ›› High Availability : Fail-safe : Gateway or VLAN Fail-safe ???

 

Need urgent support and help

 

application delivery
availability
deployment
performance
TMOS
TMSH

2 Replies

Sort By
  • nathe's avatar
    nathe
    Icon for Cirrocumulus rankCirrocumulus
    Nov 14, 2014

    Sounds like you're using aggregated links / trunks from your description. If so then I'd recommend HA Group to monitor the health of the interfaces within this trunk. i.e. if an interface goes down then possibly failover.

     

    I believe this is recommended over vlan fail-safe as it's quicker to failover (it's called Fast Failover). But for extra redundancy then vlan fail-safe is an option. Basically it checks that traffic is traversing this vlan (on a per vlan basis) and if not then it will initiate traffic before deeming that there is an issue (interface down) and failing over.

     

    Hope this helps,

     

    N

     

  • Gambler_168259's avatar
    Gambler_168259
    Icon for Nimbostratus rankNimbostratus
    Nov 14, 2014

    Yes Nathan,

     

    We have two upstream Cisco Nexus 5000 L3 Switches and both our F5 LTM Load Balancers have connected to them in full mesh i.e. LTM-01 is connected to Nexus-01 & Nexus-02, similarly LTM-02 is connected to Nexus-01 and Nexus-02 as well + Both LTM are connected to each other on HA VLAN trunk and with heartbeat cable as well.

     

    Two VLANs are configured on it External VLAN & Internal VLAN. Both VLANs have their floating IPs.

     

    What we had do is that we had disable layer-3 connectivity of nexus-01 with client upstream core router and the primary path definitely went unavailable for traffic and in the meanwhile it switches to the secondary nexus-02 which is having a layer-3 link with client upstream secondary router. We then observes a drop with request time-out in the continuous ping + a service disruption as well and the active LTM-01 appears as Backup and the backup LTM-02 appears as primary without even a manual intervention or without even using any manual forceful standby option in LTM.

     

    Looking for response.