OK, I ping OLT from MX like you Saku wrote and i get information: With JunOS, the size you specify on the ping command is actually the size of the ICMP payload. 1472 of payload + 8 bytes of ICMP header + 20 bytes of IP header results in a 1500 bytes IP packet, which is the largest packet you can send without fragmentation.
When i ping OLT with 1500 I don't get respons. L2circuit/epipe is up, isis adj is up. No every website work eg. ripe.net Where is the problem? 2013/8/7 Saku Ytti <s...@ytti.fi> > On (2013-08-07 17:43 +0200), Mark Tinka wrote: > > > True, but the OP suggested that some of the web sites he's > > browsing don't load correctly, which could allude to an MTU > > issue in the data plane. > > Mea culpa, luckily point remains it does not matter what the eline MTU is. > > Fully agreed on your suggestion that just guarantee core is large enough, > no point doing 'exactly the right size' but large enough. > > For sake of argument 'exactly the right size' in untagged ethernet core > > 6B DMAC (not calculated by IOS) > 6B SMAC (not calculated by IOS) > 2B type (not calculated by IOS) > 4B IGP label > 4B VPN label > 4B control word (maybe, maybe not) > 6B inner DMAC > 6B inner SMAC > 2B inner type > xB payload > 4B FCS (not calculated by IOS or JunOS) > > > It might be prudent to figure out customer experienced MTU by doing over > the l2circuit: > > a) ping with 1500B DF set > b) ping with 1501B DF not set > > -- > ++ytti > _______________________________________________ > juniper-nsp mailing list juniper-nsp@puck.nether.net > https://puck.nether.net/mailman/listinfo/juniper-nsp > _______________________________________________ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp