What's the mtu and encap on the ce facing interfaces? If the default 1514 frame/1500 data, then a PE ping with size 1472, when combined with IP and ICMP (28 bytes), yields the largest packet supported w/o frag.
You mention that you have enabled rsvp PMTU signaling and frag. Its not clear what you mean by "solves problem but causes L2 packet drops". Are these drops for the CE's described below, or for some other L2 service that touches these PEs? Thinking out loud, as RSVP is not enabled on the PE-CE facing interfaces, its possible that RSVP PMTU signaling it not discovering the 1500 byte MTU on the CE links, and this prevents ingress PR frag, which then causes loss. Not clear how enabling could/would break an L2 service. HTHs > -----Original Message----- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf Of Salik Mobin > Sent: Monday, March 05, 2007 2:31 AM > To: [email protected] > Subject: [j-nsp] L3VPN: Pinging CE from PE with larger packet size > > > Dear All, > > We have configured an l3vpn b/w two sites in the following topology: > > [CE-A]----FE----[PE-A(M10i)]------E3------[PE-B(<10i)]----FE----[CE-B] > > Although everything b/w CEs works well, we are unable to ping > CE-B from PE-A above packet size of 1472 Bytes. It is the > case with CE-A from PE-B also. We are not using PMTUD (rsvp > mtu-signalling) and even enabling it does not make change in > problem. Enabling "allow-fragmentation" under "path-mtu" > though solve this problem but it causes packet drops for l2circuits. > > Do we need to re-confgure MTU some where? Default MTU for E3 > interface is 4474. > > Please suggest. > > Regards, > Mobin > > > _______________________________________________ > juniper-nsp mailing list [email protected] > https://puck.nether.net/mailman/listinfo/juniper-nsp > _______________________________________________ juniper-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/juniper-nsp
