I have a RedHat 6.2 system that has been running for several months with no problems. I have a block of 5 static IP addresses thru a DSL ISP.

I have 2 NICs, eth0 and eth1. eth0 connects to the DSL modem and maps to all 5 IP addresses. eth1 maps to a class C network (intranet) that I use IPChains with in order to provide internet access to my other PCs - it goes to an 8 port hub.

This system has been working beautifully. Then a friend suggested I try to upgrade sshd, since it wasn't working on my system. I did this (using RH RPMs), without considering that my network monitoring program would freak out about one of its monitored programs being "hacked". It dutifully shutdown all the ports on eth0.

Well, this was all easy enough to recover from, except that when I fixed everything (restarted so it would generate new checksums), I had to "manually" restart the network (running the /etc/rc.d/init.d/inet script), and for some reason 2 of my addresses, 32 and 34, would not respond to pings from external machines.

I've checked over all my config and I can't find anything strange. The setup for eth0:0 thru eth0:4 are identical. It doesn't make sense for 3 out of 5 IPs to work perfectly and the other 2 to be completely dead. It seems mighty coincidental, but I don't see how this problem could be anything on my end - it has to be the ISP, right? I have another friend who insists that the problem is some config on my server, but he's an NT guy and hasn't been able to suggest what might be wrong or what I should look for.

The ISP logged onto the router and they say that there are no MAC addresses showing up for those 2 IPs. I switched out the NICs and that didn't help.

The specific IPs that don't work are 4.18.169.32 and 4.18.169.34. The ones that do work are 4.18.169.30, 31 and 34.

Anyone have any suggestions?