[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: e1000 in rawhide kernel



> Peter Robinson <pbrobinson gmail com> wrote:
>
> >The higher end ones take ages to negotiate because they try and neg
> >Spanning Tree Protocol first so the port goes through a blocking,
> >discovery and then is finally into the forwarding stage.
>
> "Higher end" switches (such as el cheapo Dells) can take up to 50 seconds to
> start forwarding data while the Spanning Tree Protocol runs through its
> BLOCKING, LISTENING, LEARNING, FORWARDING states; but that is assuming you are,
> for some reason, running STP on an Access port.
>
> Normally you'll have configured these with the equivalent of (Cisco) IOS'
> "spanning-tree portfast" to shortcut directly from BLOCKING to FORWARDING.
> Semi-recent IOS can also do this for trunk ports if you need to pass tagged
> VLANs out to your server for some reason.
>
> IOW, switches that take ages to start passing traffic on Access ports — where
> all relevant FC boxen will live — are either mis-configured or broken.

Or not configured. The number of sites that I've been to that have
switches that run the default config is amazing and its a question
that needs to be asked.

Pete


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]