I'm having a strange problem with a PM2eR/10Port running ComOS 3.3.1 - the
problem also existed in 3.1.4. Since then I've totally reset the box
[flash] and reconfigured it (no problem, just felt like starting from
scratch before I put it in the other city). I had an identical problem a
while back (same PM) but after a few days it "went away." The basic
problem is that the ether0 (10baseT and BNC ports) can recieve and sort of
send [but can't even get a packet across]. Here are some details:
...
Local ethernet with Local ethernet with
my servers on it. currently one PC.
| |
+-----+ +-----+
| pm1 | ---[ Line to another city ]--- | pm3 |
+-----+ +-----+
|
Connection to
the Internet
...
I currently have a FreeBSD on pm3's ethernet for debugging purposes
(The Win95 PC isn't very helpful). While testing I ptrace a filter which
permits icmp packets on the portmasters and run tcpdump -vv icmp on the
FreeBSD box.
* When pinging from pm3 to the FreeBSD box tcpdump shows nothing but
I can see activity on the ethernet. The arp entires on pm3 and FreeBSD
both list each other so that can't be the problem.
* When pinging from FreeBSD to pm3 I can see activity on the ethernet and
see the icmp echo requests going out [according to tcpdump]. And pm3 sees
the echo requests coming in.
* When pinging from FreeBSD to a server on my pm1 ethernet both PM's show
the echo request and then the echo reply coming back, up to pm3. tcpdump
never sees the echo reply.
And so on.. I've tried this with different IP addresses, tried it with
BNC instead of 10baseT, tried playing with the routes, using different
cables, using the FreeBSD box instead of the Win95 one.. all to no avail.
I can provide extra details, show the arp tables, routes, etc upon
request. Thanks for reading this far - sorry it's such a long message :0
Anyone have any ideas? Can Livingston support come in and check the
the ether0 hardware if I called? Thanks in advance..
- Steve