EX4200 Virtual Chassis NIC teaming.

  • Hi All,

    This is my first post here.  This forum has been a great source of help for me.  I am stumped on a particular configuration.  We have setup 2 virtual chassis of EX4200 switches connected with LAG.  We are having some issues with connecting the teamed NICs from HP servers. After the server folks have configured the teaming on the server, we connected one NIC to one VC and the other NIC to the other VC.

    We don’t need to have the link aggregation, we just need Active/Passive setup.  Right now, when we test failover (pull cable on the active NIC).  The server drops connectivity.  We had the similar setup with Nortel that worked for years.  Do we need to configure anything special in the Juniper world?

    Any help is greatly appreciated.

    Thanks in advance.

  • Global Moderator

    Thanks for a good post kshah! You’re right to clearify this. Should have included this in my answer.

  • if you use LACP on the server side that means you team two nic cards and you will need LACP configuration on the switch side.  This is Active active NIC teaming. Here you can use 100% uplink bandwidth.

    In case of Active standby NIC teaming you do not need any LAG configuration on the switch side ( not even static lag) The failover on the server side should be taken care by NIC teaming software on the server. So lets say if one connection goes down it would be NIC card teaming / driver’s responsibility to bring another link up. Switch will show one interface up and another down and vice versa depending on which NIC card interface is up, when fail over happnes the MAC address learning will also occur. With this configuration you can use only 50% of available bandwidth.

  • Global Moderator

    LAG interfaces only work from 1 device (or VC in this case). If you want redundancy to two VC’s you have to use SPT (apnning tree) I’m afraid.