Your safest approach is: > Your transport providers need to guarantee you a specific MTU and you > need to configure your router L3 interfaces with that number.
If you only set your router tcp mss, that will work fine for BGP sessions but will not fragment transport data packets properly. You have a hard requirement here to set the interface MTU to whatever the minimum is that the provider can guarantee. Nick CiscoNSP List wrote: > Thanks for all the replies - Given our circumstances, I think the > "safest/sanest" approach would be to disable PMTUD, and enable "ip tcp mss > xxx" (Maybe 1300 or 1400). > > Cheers > > ________________________________________ > From: Nick Hilliard <[email protected]> > Sent: Thursday, 25 February 2016 6:35 AM > To: CiscoNSP List > Cc: [email protected] > Subject: Re: [c-nsp] transport path-mtu-discovery - ME3600....too > unpredictable to use? > > CiscoNSP List wrote: >> Quick synopsis of our network, multiple pops, all connected via >> various 3rd party carriers, who all use differing MTUs, that can also >> "change" unexpectedly(Unavoidable unfortunately!) > > MTU means _maximum_ transport unit, so if your underlying transport > network unexpectedly changes this so that the max transport packet size > decreases below the configured MTU on the layer 3 side, then you are > going to drop packets, end of story. > > Your transport providers need to guarantee you a specific MTU and you > need to configure your router L3 interfaces with that number. > > Nick > > _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
