Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

long-time reader first-time poster.

I managed a network of 70 odd PCs connected to 3x different network switches (3 buildings all next door to each other).

Looks like this:

BUILDING 1 SWITCH -> BUILDING 2 SWITCH -> BUILDING 3 SWITCH

BUILDING 3 is where the router and main server are.

So, yes, a PC in building one has to go through building 2's switch to get to the server or internet gateway in building 3.

Now, weirdness, we have a device in building 2 (a PABX) that can be accessed and pinged from building's 1 & 2 no worries. In building 3, however, it will respond to 1x ping request and then nothing. Destination host unreachable the whole way.

Everything else seems find and this is the only device we cannot access (or know that we cannot access) in the whole network.

My suspicion is that we could have a device IP conflict on the switch in building 3 - so traffic in building 1 and 2 is fine as it finds the correct device before it makes it to building 3. I am going to spend a little time unplugging and replugging cables in building 3 and testing that way, but other than an IP conflict I'll admit I'm a little lost as to what else could cause this? MAC conflict perhaps?

Just wondering if anyone out there has experienced anything like this? Ideas?

Thanks! -Scott

share|improve this question
    
Can't give much definite, so just a comment, but check arp tables for the PABX's mac address. –  Joel Coel Aug 12 '12 at 2:37
    
If it was me, I would try to mirror the switch trunk and then fire up a sniffer and see what is actually happening. Or run a wireshark on the client, or server or something. –  Zoredache Aug 12 '12 at 4:53
    
Can you confirm that "BUILDING 3 is where the router and main server are" implies that the switches are forming a flat layer-2 network (i.e.: same subnet). If so, then this does suggest an ARP issue. Also, when you get your first ping, and then destination unreachable, what needs to be done in order to get another successful ping? e.g.: wait some time, or reboot client? –  Simon Catlin Aug 12 '12 at 19:13
    
Hi guys, thanks for the hints. Simon, it just requires about 5-10 minutes of waiting before another ping will get one response, then stop. –  ScottNZ Aug 17 '12 at 3:22
    
Smells like some kind of ARP caching problem somewhere. Do you have access to or know someone who has access to the CLI on the switches (assuming they are managed devices)? Just thinking you could ask them to check the cache on there, plus the MAC-to-port mapping (can't remember the name of this). –  Simon Catlin Aug 21 '12 at 21:30

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.