Russell L. Carter wrote: > Bob Proulx wrote: > > Russell L. Carter wrote: > >> Ok, I need to do some experimenting here. I've broken out a long patch > >> cable to bypass the switch > > > > I would really be surprised if the switch has broken down. Not impossible > > of course. But what are the odds? I think it very unlikely. If I were to > ... > to look at this again. I reconfigured my net to get the DHCP > conversation on the least used net, patched the laptop directly to the > server, and on the server ran > ... > And it worked!
Wow! It *was* the switch. I put that in the list because it was possible in the divide and conquer debugging but truly I expected it to be something else. > So then the problem was to figure what was awry with the switch, a > D-Link DGS-1224T. I thought I had sometime in the distant past > lobotomized it but it turns out no, I didn't get it done completely. > In particular, IGMP snooping was enabled. I downloaded the switch's > manual (thanks D-Link!) and here is what it said: > > "With Internet Group Management Protocol (IGMP) snooping, the > Web-Smart Switch can make intelligent multicast forwarding decisions > by examining the contents of each frame’s Layer 2 MAC header. IGMP > snooping can help reduce cluttered traffic on the LAN. With IGMP > snooping enabled globally, the Web-Smart Switch will forward > multicast traffic only to connections that have group members > attached." > > But there were no "group members" defined. Oops. Disabled, applied, > power off/on, voilà. Ah... So not broken but simply not configured appropriately. Good deal! And good job at working that through to root cause. Always nice when problems are found and fixed. :-) Bob
signature.asc
Description: Digital signature