I investigated a little further; here are the results: 1) When connecting two of the machines directly via a cross-over cable, all multicast traffic seems to reach the other machine, so obviously our Cisco is "eating" most of the traffic.
2) snooping on the ce1 interface and looking at the multicast packets with our (admittedly quite old) ethereal, all the dropped packets have the same UDP checksum value of 0x511b which ethereal reports as incorrect. The few packets that reach the other machines are reported to have a valid checksum. 3) When snooping using the setup described in (1), most of the packets still have an invalid checksum but the value is 0x511a in this case. So it looks like if the checksum is either miscalculated or gets corrupted somewhere. I'll try to reproduce this behavior with Sol9/SRSS 3.1 later today to see if it's an Solaris 10 or SunRay issue... This message posted from opensolaris.org _______________________________________________ networking-discuss mailing list [email protected]
