Dale Shaw wrote:
Hi,
2009/2/12 Antonio Soares <amsoa...@netcabo.pt>:
This seems buggy to me :(
Bad documentation, at least. The IOS 12.4 *mainline* command reference
and config guide waffle on and on about 'bgp transport' yet the
command (BGP router config mode or per-neighbour) doesn't even exist
in that train.
I suspect when it does exist (allegedly in 12.2(33)SRA, 12.2(31)SB,
and 12.2(33)SXH and 12.4(20)T), the behaviour is as described -- BGP
uses the BGP-specific TCP MSS settings. Without it, it uses the global
default (PMTUD off), and you need the 'ip tcp path-mtu-discovery'
global command to make it work.
I've tested it on IOS 12.4 mainline and even for directly connected
neighbours, MSS is 516 bytes. Once 'ip tcp path-mtu-discovery' is
switched on (on both peers), MSS is 1440 bytes.
Just for fun I looked at my BGP mesh and found the following:
* Directly connected neighbors MTU=516
* Transit to upstreams MTU=1440 or 1460
This is under 12.4(16) and I've never touched 'ip tcp
path-mtu-discovery' or per-neighbor MTU settings.
~Seth
_______________________________________________
cisco-nsp mailing list cisco-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/