Using rb333 and rc9. The network layout is like this: EDIT: can't get the bridged sectors and power monitor to show up correctly on here. bridged sectors go above switch and power monitor above HP managed switch.
bridged sectors power monitor
| |
local network ---nstreme2---rb333dn---switch---rb333up--- nstreme2---rb333dn2----HP managed switch----edgerouter----internet
The rb333up has a secondary public IP for the customers on the bridged sectors.
From rb333dn2 to the edgerouter it is all ethernet. There are not any bridges or secondary addresses from that point out to the internet.
The rb333dn - switch - rb333up is a common configuration on all towers.
Problem:
If I try to ping or traceroute the edgerouter from rb333up it times out, last returned address in the traceroute is rb333dn2.
If I try to ping or traceroute the edgerouter from rb333dn2 it works.
If I try to ping or tracroute the edgerouter from rb333dn it makes it all the way with some timeouts.
This tells me the routing tables are correct (static routes) since I can back up a step and it makes it. Also I have looked at all of them to make sure they are correct.
If I back up into the local network to the next tower back, the behavior is the same from that tower, up times out, dn gets through with timeouts.
As the load increases on this path, retries increase rapidly.
If I back up further into the network where the boards are 532a and stable software, the traceroute makes it.
I cannot think of any reason for this behavior other than perhaps a software issue, maybe in the nstreme2 protocol.
Maybe someone else has other ideas or is there a problem with the software?
I have not sent a supout, didn't think it would show anything in this case.
tk