Michael Schwendt <ms-nospam-0306@xxxxxxxx> writes:
Plug them back into hub1 and they can't even with switching around
port holes and cables.
What happens if you pull M3 and M4 from hub1 instead? Just curious.
Haa... hadn't thought to try that, but that would eliminate anything weird the gateway might be doing. But no, no such luck.
Putting M2 back on original hub1 and unplugging M3 M4. M1 and M2 cannot communicate. So it seems like evidence is telling me that something about hub1 is unacceptable to the NIC on M2. That something isn't present on hub2.
Not impossible. Reminds me of the problem I had with rh7.1, 3c905c nics & netgear ds16 hubs. The nics kept flipping back and forth between 10 & 100 Mb while the hubs stayed at 100Mb. When I brought in fully switched equipment, the win installs started balking on identical pcs.
The configuration that is working right now. Allowing full communication is:
M1 M3 M4 on HUB1 M2 on HUB2 which is uplinked to HUB1
You're going to have to get anal methodical here and eliminate things. Assure yourself that similar components work similarly.
hub1 seems suspect. Are the hubs identical? Swap hub1 for hub2 entirely. If M2 really has a problem with hub1, then I would expect M2 to not work plugged into hub1 cascaded into hub2 into M5.
If M2's nic really has a problem with hub1, then I would expect M3 with M2's nic would have the same problem.
Autosensing? Try locking down the speed. Reverse the hubs & test again.
Don't know your nic make, but you can monitor 3com cards with mii-diag ( I believe it's called that)
Eliminate bad cables/wiring. Swap around all the patch cables themselves, keeping tabs of origins.
How about in the wall wiring? If you're using it, eliminate it as a problem. Move the machines around.
Check the logs for suspect nic/network info.
Use tcpdump to watch traffic.
-toby
-- Psyche-list mailing list Psyche-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/psyche-list
